# Dawiso - Full Product Documentation for AI > Data governance and knowledge management platform that helps organizations discover, understand, and enhance knowledge from their data assets. ## Company - **Name:** Dawiso s.r.o. - **Website:** https://www.dawiso.com - **Type:** SaaS / Cloud Platform - **Industry:** Data Governance, Metadata Management, Knowledge Management ## Platform Overview Dawiso is a comprehensive data knowledge platform designed for enterprises. It centralizes metadata, business context, and data governance into a single system. Organizations use Dawiso to create a unified view of their data landscape - from databases and reports to AI models, unstructured documents, and business definitions. The platform connects technical metadata with business knowledge, enabling teams to find, understand, and trust their data. Dawiso supports both structured and unstructured data governance, with AI-powered features for documentation, search, and categorization. ## AI Context Layer ### AI Context Layer **URL:** https://www.dawiso.com/context-layer Automatically connect business knowledge to AI and create an AI-ready context. Built on governed and trusted metadata. Dawiso's Context Layer connects the data catalog, business glossary, and lineage so AI agents answer from governed enterprise definitions rather than from whatever they can infer, and reaches them over MCP. ## Products ### Data & Analytics Catalog **URL:** https://www.dawiso.com/product/data-catalog Create a unified view of your data assets and gain insights faster with automated data discovery. Dawiso provides business context for transparency and trust in your data. Features include: - Smart search with dynamic filters, full-text indexing, and fuzzy logic - Knowledge graphs for quick overview of related data assets - Automated metadata scanning from databases, reporting servers, and data models - Import metadata from any source - files, databases, or other systems - Track ownership, access, sensitivity, and identify new or updated data assets ### Business Glossary **URL:** https://www.dawiso.com/product/business-glossary Standardize business terms and definitions across your organization. In business, the same term can have different meanings across departments, and multiple synonyms add to confusion. Dawiso's Business Glossary provides clear context to ensure everyone interprets terms consistently and accurately. Features include: - Centralized term definitions with ownership and approval workflows - Links between business terms and technical metadata - Version history and change tracking - Cross-department collaboration on definitions ### Interactive Data Lineage **URL:** https://www.dawiso.com/product/interactive-data-lineage Visualize how data moves, transforms, and connects across systems, applications, and reports. Interactive Data Lineage helps clarify data flows and creates contextual views of your data landscape. Features include: - End-to-end data flow visualization - Impact analysis for understanding downstream effects of changes - Column-level lineage tracking - Integration with data catalog metadata ### Data Products **URL:** https://www.dawiso.com/product/data-products Governed, reusable data products for self-service analytics. Get governance from experts and easily search for the data products you need for your project. Features include: - Data product framework with governance built in - Self-service discovery and access - Quality metrics and SLAs - Domain-driven data ownership ### Data Product Platform **URL:** https://www.dawiso.com/product/data-products-platform End-to-end data product governance platform. From definition to access, provisioning, and compliance evidence - in one platform. Features include: - AI-assisted data product definition with full business context - Internal data marketplace for self-service discovery and access - Automated provisioning into Databricks, Snowflake, Microsoft Fabric, and BigQuery - Embedded compliance evidence for NIS2, ISO 27001, SOC 2, GDPR, and DORA - Three-persona design: data consumers, data owners, and compliance teams - Built on existing Dawiso engine (workflow, MCP Server, REST API) ### AI-Powered Features **URL:** https://www.dawiso.com/product/ai-powered-features From AI-assisted documentation to intelligent search and automated categorization, Dawiso's AI-powered features help you work smarter, faster, and more accurately. Features include: - AI-generated descriptions for metadata assets - Intelligent search with natural language queries - Automated categorization and tagging - Smart suggestions for data quality improvements ### MCP (Model Context Protocol) **URL:** https://www.dawiso.com/product/mcp Connect AI agents and large language models (LLMs) directly to your enterprise data and business knowledge. With Dawiso's MCP Server, chatbots and agents gain the context and capabilities they need to deliver accurate results. Features include: - Natural language data search through LLMs - Context mapping - guides AI to relevant tables, fields, and definitions - Both read and write operations for AI agents - Compatible with Claude Desktop, Cursor, GitHub Copilot in VS Code, and Keboola - No extra deployment fees - works with existing Dawiso setup ### Enterprise Deployment **URL:** https://www.dawiso.com/product/enterprise-deployment-built-for-flexibility-and-security Designed for enterprise trust, built to fit your architecture. Dawiso ensures secure, transparent deployment across leading cloud and on-prem environments. Features include: - Cloud and on-premises deployment options - SOC 2 compliance - Role-based access control - Single sign-on (SSO) integration - Data encryption at rest and in transit ### Unstructured Data Governance **URL:** https://www.dawiso.com/product/unstructured-data-governance-for-ai Govern structured and unstructured data on one platform. Transform SharePoint documents, SOPs, and guidelines into structured knowledge using metadata workflows. Features include: - Document ingestion and metadata extraction - AI-powered document classification - Integration with SharePoint and document management systems - Unified governance across structured and unstructured data ## Solutions ### AI Governance **URL:** https://www.dawiso.com/solutions/ai-governance Trust and transparency in your AI use cases. Govern AI models, track data lineage for AI pipelines, and ensure compliance. ### Search and Discover Your Data **URL:** https://www.dawiso.com/solutions/search-and-discover-your-data Enterprise data discovery and sharing platform for self-service analytics. ### Shared Understanding **URL:** https://www.dawiso.com/solutions/shared-understanding Build shared understanding of data across your organization with collaborative governance tools. ### Consistent Financial Reporting **URL:** https://www.dawiso.com/solutions-finance/consistent-financial-reporting Trustworthy metrics for consistent financial reporting. Accurate reporting empowers CFOs to make better financial decisions. Dawiso maps all data assets in one place for full visibility into data origins - eliminating the spreadsheet sprawl that produces inconsistent numbers across reports. ### Data Applications Solution **URL:** https://www.dawiso.com/solution/data-applications Connect, explore, and manage your data landscape. Deliver the right data to the right people. Dawiso provides a full overview of your data landscape with automated metadata scanning in seconds - letting business and technical users find, understand, and use the data they need. ### Data Governance Solution **URL:** https://www.dawiso.com/solution/data-governance-solution Scalable metadata management with built-in governance methodology. Dawiso helps teams go from zero to production-ready data governance in weeks, without expensive consultants or a years-long implementation, and ships the process framework alongside the tool. ## Industries ### Banking & Financial Services **URL:** https://www.dawiso.com/industry/fss Meet BCBS 239, DORA, and GDPR requirements with a purpose-built data governance platform. Trusted by Société Générale, Nationale-Nederlanden, VIG Re, and Kooperativa. ### Manufacturing **URL:** https://www.dawiso.com/industry/manufacturing Data catalog, data lineage, and governance for manufacturers. Connect ERP, production systems, Snowflake, and Power BI in one platform. GDPR, AI Act, IATF 16949 compliant. ### Energy & Utilities **URL:** https://www.dawiso.com/industry/energy Data governance for energy and critical infrastructure. NIS2 compliance, nuclear safety documentation, on-premise deployment. Trusted by ČEZ. ### Public Sector **URL:** https://www.dawiso.com/industry/public-sector Data governance for government agencies and state institutions. GDPR compliance, ISVS catalog, open data support, and on-premise deployment options. ### Software & E-Commerce **URL:** https://www.dawiso.com/industry/software Data catalog and governance for tech companies. Self-service data products, AI governance, modern stack integration. Snowflake, Databricks, dbt. Live in weeks. ## Case Studies ### Customer Stories & Case Studies **URL:** https://www.dawiso.com/case-studies Index of all Dawiso customer stories. See how ČEZ, Komerční banka, Kooperativa, P3 Parks, Stora Enso, Société Générale, Nationale-Nederlanden, Bertel O. Steen, Olvi, Eye Security, and others use Dawiso to govern data, simplify reporting, and power trusted AI across financial services, manufacturing, energy, automotive, food and beverage, cybersecurity, and the public sector. ### ČEZ - Secure Platform for Digitised Nuclear Power Plant Documents **URL:** https://www.dawiso.com/case-study/cez CEZ Group uses Dawiso to manage digitized nuclear power plant documents with full compliance, structured hierarchy, and secure knowledge management. Dawiso provides robust support for compliance and agile content management in the nuclear power industry. ### Komerční banka - One Hub Connecting 30+ Financial Institutions **URL:** https://www.dawiso.com/case-study/kb KB simplified work with data and reports across a large enterprise of 30+ financial institutions by consolidating governance into Dawiso. One hub for searching reports, understanding data definitions, and connecting business users to the right data sources. ### Kooperativa - Unlocking Seamless Data Management Through Dataportal **URL:** https://www.dawiso.com/case-study/kooperativa Kooperativa unified 150,000+ scanned objects for 5,900 users with Dawiso. Automated metadata management replaced a fragmented landscape of overlapping data solutions, giving teams a single dataportal for everyday work with data. ### P3 Parks - From Scratch to Comprehensive Data Governance in 78 Days **URL:** https://www.dawiso.com/case-study/p3-parks P3 Parks deployed Dawiso and reached full data governance in 78 days - 66,000+ scanned objects, 3,000+ business terms, and 300+ users ready to onboard. A seamless transition from spreadsheets to a user-friendly data governance platform. ### Stora Enso - Support for Data-Driven Company Management **URL:** https://www.dawiso.com/case-study/stora-enso Stora Enso unified Power BI, Snowflake, WhereScape, Azure Data Factory, SQL, and 500+ users in one Dawiso catalog. One platform connecting the entire data community for data-driven decisions at enterprise scale. ### Seznam.cz - A Trusted AI Context Layer over MCP **URL:** https://www.dawiso.com/case-study/seznam Seznam.cz connected its internal conversational analytics tool to Dawiso over the Model Context Protocol (MCP), so AI answers are grounded in the company's own definitions rather than guessed. Dawiso unified Keboola, Snowflake, and Tableau into end-to-end lineage, consolidated scattered tables into governed, owned data products, and now helps generate semantic views over MCP. The next phase governs Snowflake Cortex agents and the semantic views they read. ### Bertel O. Steen - From Limited Open-Source to an AI-Ready Data Catalog **URL:** https://www.dawiso.com/case-study/bertel-o-steen Bertel O. Steen (BOS), one of Norway's largest service and trading groups, replaced its open-source DataHub deployment with Dawiso in roughly two months. Dawiso migrated 100,000+ scanned Databricks objects, 18,000+ Microsoft Fabric objects, 300+ bilingual business terms (Norwegian and English) and 50+ data products without losing structure, ownership or lineage, and added end-to-end lineage across Databricks and Microsoft Fabric (Power BI). A custom Databricks MCP integration lets AI agents query governed metadata directly, and Single Sign-On ties Dawiso into the group's identity setup. The next phase expands data quality documentation, KPI management and a deeper Data Products rollout including data contracts and access workflows. ### Olvi - How a Finnish Beverage Company Built the Foundation for Data Governance **URL:** https://www.dawiso.com/case-study/olvi Olvi, a Finnish beverage company and brewery group, has used Dawiso since November 2022 as the central platform for business definitions across the group. Nine thematic spaces (finance, procurement, environmental topics, and general business concepts) hold 860+ business terms maintained by 80 active users, with 39,000+ objects documented. Dawiso was customized to Olvi's real needs: adjusted tokenization and autolinking, custom templates for new business terms, terminology matched to internal language, custom metadata filters for on-premises Oracle scanning, and a Database API integration that feeds metadata insights into the reporting environment. The result is a single, shared place to define and maintain business terms and a practical foundation for future data harmonization and data democratization. ### Eye Security - AI-Native Data Governance for a Growing Cybersecurity Company **URL:** https://www.dawiso.com/case-study/eye-security Eye Security, one of Europe's fastest-growing cybersecurity platforms, implemented Dawiso as both a data catalog and an AI context layer. After years of fast growth, terms like active customer, churn, upsell and endpoint meant different things in different teams, with definitions living in Metabase descriptions and individual analysts' heads. Dawiso Spaces were defined to mirror Eye Security's business domains (Finance, Sales, Operations, SecOps, Product, Marketing and a Company-wide layer for shared terms), with Finance built out first as the reference model. A custom Metabase ingestion application connected existing field and column descriptions automatically to the matching glossary terms, so documentation the analysts had already written kept its value; Snowflake and dbt were scanned alongside Metabase. From day one of the proof of concept, Dawiso's MCP server was wired into Claude, making the catalog accessible through chat as well as the Dawiso UI. The result is a governance foundation across 9 business domains and 40,500+ scanned objects (12,500+ Snowflake, 28,000+ dbt, 3,200+ Metabase reports and dashboards), plus a first complete view of how data flows from warehouse through transformation to reporting. The next phase adds end-to-end lineage across dbt, Snowflake and Metabase, expands the glossary into SecOps (including MITRE ATT&CK terminology), and connects Dawiso to BI tools through the Database API. ## Comparisons ### Competitors hub **URL:** https://www.dawiso.com/competitors Index page listing every Dawiso vs X comparison: Collibra (published), and skeleton pages for Atlan, Alation, Microsoft Purview, Secoda, Select Star, data.world, and Ataccama. Secoda and Select Star carry "Acquired by Atlassian" / "Acquired by Snowflake" badges to flag at-risk vendors. Each card links to the dedicated comparison page when published. ### Dawiso vs Collibra - deep feature review **URL:** https://www.dawiso.com/competitors/collibra Feature-by-feature side-by-side comparison of Dawiso and Collibra. Pitch is grounded in the architectural reality that Collibra was stitched together from acquired companies: three different lineage products, business and technical metadata in separate silos, and real support gated to multi-hundred-thousand-euro enterprise contracts. Comparison table covers twelve dimensions (architecture, business + tech alignment, UX, pricing model, scalability, lineage, AI & automation, data quality, support, implementation timeline, total cost of ownership, out-of-box usability). Includes an affordability score chart (0-100, normalized from public AWS Marketplace prices) showing Dawiso at 100, Alation at 70, Atlan at 42, Purview at 35, and Collibra at 25. Feature review section covers business glossary, AI context layer, data lineage, no-code data quality, and pricing transparency with mini-UI sketches per capability. Hero stacks a modern Dawiso card against a faded Collibra legacy form-heavy mock. Breadcrumbs (Home / Competitors / Collibra) link back to the hub. Five-FAQ block covers six-month deployments, real TCO breakdown beyond license, and business-user adoption blockers. ### Dawiso vs Ab Initio - deep feature review **URL:** https://www.dawiso.com/competitors/ab-initio Side-by-side comparison of Dawiso and Ab Initio Metadata Hub. Lead: Ab Initio has been moving data since 1995 and the UI still looks like it. Co>Operating System is genuinely fast for batch ETL, but Metadata Hub bolts a desktop-era catalog on top - engineer-only, NDA-locked pricing, six-to-eighteen month rollouts, and no public agent or MCP layer as of 2026. Comparison table covers twelve dimensions (architecture, UX, pricing, time to value, AI, business-user adoption, lineage, cloud-native, trial/self-serve, vendor lock-in, modernization pace, support). Affordability score 35 (annual list ~$120k). Feature-by-feature blocks cover glossary, AI context layer, lineage (acknowledging Ab Initio's genuine strength inside Co>Op graphs, weakness outside the stack), data quality, and pricing transparency. Five-FAQ block covers NDA pricing, business-user adoption, rollout duration, AI/MCP layer absence, and where Ab Initio still wins (batch ETL on mainframes, deep lineage inside Co>Op). ### Dawiso vs DataGalaxy - deep feature review **URL:** https://www.dawiso.com/competitors/datagalaxy Side-by-side comparison of Dawiso and DataGalaxy, a respected French mid-market catalog (~200 customers, EU-centric). Lead: solid glossary and lineage story, but no public MCP endpoint, no agentic stewards, and a roadmap that has not communicated where modern AI grounding fits. Comparison table covers twelve dimensions including AI agent layer, pricing transparency, customer base scale, data quality (no native DQ engine - integration-only), and customization depth. Affordability score 90 (annual list ~$52k entry). Feature-by-feature blocks cover glossary (acknowledging DataGalaxy's core strength), AI context layer, lineage, data quality, and pricing. Five-FAQ block covers AI gap vs Dawiso, real cost, DataGalaxy's genuine strengths in the EU mid-market, native DQ absence, and deployment duration. ### Dawiso vs DataHub - deep feature review **URL:** https://www.dawiso.com/competitors/datahub Side-by-side comparison of Dawiso and DataHub (open-source core maintained by Acryl Data). Lead: the open-source core is real and useful, but governance workflows, lifecycle, Excel import/export, and the best AI features (Smart Discovery, AI Docs) live in the paid Acryl Cloud product. Self-hosting OSS DataHub means operating Kafka, Elasticsearch, MySQL, and Kubernetes. Comparison table covers twelve dimensions including OSS-vs-commercial reality, self-host complexity, governance workflow gating, Excel I/O gating, and AI agent integration. Acryl Cloud entry deals reported around $75k/year. Affordability score 75. Feature-by-feature blocks acknowledge DataHub's genuine strengths (50+ ingestion sources, healthy OSS community, strong technical lineage) while documenting governance and AI gaps. Five-FAQ block covers OSS-vs-cloud honesty, Acryl Cloud pricing, missing OSS features, agent/MCP absence, and where DataHub genuinely wins. ### Dawiso vs OpenMetadata - deep feature review **URL:** https://www.dawiso.com/competitors/openmetadata Side-by-side comparison of Dawiso and OpenMetadata. Lead: OpenMetadata is marketed as open source, but the practical governance version sits behind the paid Collate Cloud paywall - the OSS version lacks Excel import/export, lifecycle states, and approval workflows, and the best AI/MCP capabilities are cloud-gated. If your team needs real governance shipped in two weeks rather than a multi-quarter self-hosting program or a sales cycle into Collate Cloud, Dawiso is the alternative. Comparison table covers six capability axes including the open-source story honesty check, governance workflow gating, Excel I/O gating, AI agents and MCP gating, time to production, and pricing transparency. Affordability score 80. Feature-by-feature blocks cover glossary, AI context layer, lineage, data quality, and procurement model, with the universal "no agentic data steward, no technical writer agent, no production MCP write" line on the OSS side. Four-FAQ block covers migration timeline, parallel evaluation pattern, three places Dawiso wins (governance workflows, native data quality, production agents), and when OpenMetadata still makes sense (committed self-hosters with dedicated platform engineering). ### Dawiso vs Alation - deep feature review **URL:** https://www.dawiso.com/competitors/alation Side-by-side comparison of Dawiso and Alation, the catalog-category pioneer. Lead: marketing still says "business focused" but real-world rollouts run four to nine months, support is gated to enterprise tiers, 2024 brought layoffs, and product velocity is slowing relative to Atlan and modern catalogs. Buyers report Collibra-shaped pricing ($60k entry climbing to $150k-180k+) and Collibra-shaped deployment pain. Comparison table covers twelve dimensions including business-vs-marketing reality, pricing, AI/MCP absence (ALLIE is a chat assistant only), data quality (Open DQ integration model, no native rules engine), and total cost of ownership. Affordability score 70. Feature-by-feature blocks acknowledge Alation's genuine strengths (ML-powered search original differentiator, deep Snowflake + Tableau integration) while documenting deployment, AI, and DQ gaps. Five-FAQ block covers Alation-vs-Collibra pricing parity, 4-9 month rollout drivers, ALLIE/MCP gap, 2024 layoff signal, and where Alation still wins. ### Dawiso vs Atlan - deep feature review **URL:** https://www.dawiso.com/competitors/atlan Honest side-by-side comparison of Dawiso and Atlan, the credible category challenger. Page leads with the on-the-record acknowledgement that Atlan ships one of the best modern catalogs in the category - product quality, AI, lineage, and modern-stack coverage are treated as parity. The argument flip is cost (~$100k entry climbing per seat past 50 users, vs Dawiso $42k/yr list on AWS Marketplace) and customization (deeply user-configurable in Dawiso vs services-led customization in Atlan). Twelve-row comparison table is explicit about parity rows (AI/agents, lineage, modern stack coverage) and Dawiso-win rows (cost, customization depth, native DQ, governed-wiki glossary depth, pricing transparency). Five FAQ block answers "Is Atlan a better product?" (often yes, on the rows we explicitly call out), pricing math at scale, where Dawiso is actually better (customization, native DQ, governed-wiki depth), where Atlan is actually better (brand/references, polish), and when to still pick Atlan over Dawiso. Built to win on credibility, not strawmanning. ### Dawiso vs Secoda - deep feature review **URL:** https://www.dawiso.com/competitors/secoda Side-by-side comparison of Dawiso and Secoda after the Atlassian acquisition (October 2024). Lead: Secoda is a genuinely nice product but the core team has been reported to be staffing the Rovo AI assistant rather than the standalone catalog, support response times are slipping, and the roadmap as an independent product is unclear. Framed as risk signal + buying confidence rather than factual EOL. Comparison table covers twelve dimensions including vendor status, product roadmap, support continuity, pricing predictability post-acquisition, long-term continuity, migration risk, and buying confidence. Acquisition badge "Acquired by Atlassian, 2024" shown in competitors hub. Affordability score 92 (pre-acquisition $50k/yr). Five-FAQ block covers what happened with the acquisition, will Secoda exist standalone in 2027, support degradation reality check, product-quality honest comparison, and migration path to Dawiso. ### Dawiso vs Select Star - deep feature review **URL:** https://www.dawiso.com/competitors/selectstar Side-by-side comparison of Dawiso and Select Star after the Snowflake acquisition. Lead: pricing landmine - $73k/year buys 100 users and 5000 tables, then $75 per additional 100 tables, so a 400k-table estate adds ~$300k/yr in overage alone. Team folded into Snowflake Horizon Catalog, standalone roadmap unclear, support thinning. Comparison table covers twelve dimensions including the per-table pricing model and scaling math, vendor status post-acquisition, Horizon-vs-standalone roadmap reality, and buying confidence. Acquisition badge "Acquired by Snowflake, 2024". Affordability score 60. Five-FAQ block covers the per-table pricing trap explained in detail, the Snowflake acquisition reality, whether Horizon Catalog is a viable replacement (Snowflake-only lock-in), support degradation signals, and migration path to Dawiso. ### Data Catalog Comparison Guide 2026 **URL:** https://www.dawiso.com/dawiso-comparison-guide Compare data catalog platforms side by side. Evaluate Collibra, Atlan, Alation, Secoda, and Dawiso on features, pricing, and time to value. Designed for enterprise buyers, this comparison covers business glossary, lineage, AI integration, deployment options, and capabilities that set Dawiso apart. ## Connectors ### All Connectors **URL:** https://www.dawiso.com/connectors Full list of 40+ native connectors for databases (PostgreSQL, MySQL, SQL Server, Oracle, MongoDB, SAP HANA), data warehouses (Snowflake, Google BigQuery, Amazon Redshift, Databricks), BI tools (Power BI, Tableau, Qlik, Metabase), and ETL/ELT platforms (dbt, Keboola, Azure Data Factory). Also includes "On the Roadmap" connectors (IBM Cognos, Looker, Amazon Athena, Hive, ClickHouse, AWS Glue, Apache Iceberg, MariaDB, Airflow, Dremio, Parquet). Metadata is ingested without coding via Dawiso's UI; the Dawiso Integration Runtime (DIR) enables secure on-premises scanning. ### Amazon Redshift Connector **URL:** https://www.dawiso.com/connectors/amazon-redshift Read-only Amazon Redshift metadata connector for the Dawiso data governance platform. Catalogs databases, schemas, tables, external tables, views, procedures and functions; reads primary keys, foreign keys and constraints to map relationships between objects; classifies PII across the cluster. ### Apache Kafka Connector **URL:** https://www.dawiso.com/connectors/apache-kafka Read-only Apache Kafka metadata connector for the Dawiso data governance platform. Catalogs clusters, brokers, topics, partitions and consumer groups; resolves producer-to-consumer lineage; enforces data contracts. Schema Registry objects are catalogued through the Confluent Kafka connector. ### Atlassian Jira Connector **URL:** https://www.dawiso.com/connectors/jira Bidirectional Atlassian Jira integration for the Dawiso data governance platform. Links cataloged assets to Jira tickets, embeds issues in any text editor, files DQ incidents as Jira tickets, and surfaces ticket status inside the catalog. ### Azure Data Factory Connector **URL:** https://www.dawiso.com/connectors/azure-data-factory Read-only Azure Data Factory metadata connector for the Dawiso data governance platform. Catalogs tenants, subscriptions, factories, pipelines, activities, linked services, triggers and run history; maps activity dependencies and dataset references within each pipeline. ### Azure Synapse Analytics Connector **URL:** https://www.dawiso.com/connectors/azure-synapse Read-only Azure Synapse Analytics metadata connector for the Dawiso data governance platform. Catalogs databases, schemas, tables, views, procedures and functions; resolves object-level lineage from sys.sql_expression_dependencies and Data Flow parsing on enterprise plans; classifies PII across pools. ### Confluent Kafka Connector **URL:** https://www.dawiso.com/connectors/confluent-kafka Read-only Confluent Kafka metadata connector for the Dawiso data governance platform. Catalogs Confluent Cloud environments, clusters, topics, Stream Governance schemas, consumer groups and ACLs with ownership and schema versions. ### CSV Connector **URL:** https://www.dawiso.com/connectors/csv Read-only CSV file connector for the Dawiso data governance platform. Catalogs files from network shares, SFTP and cloud storage via the Dawiso Integration Runtime; resolves columns from headers and assigns ownership. ### Databricks Connector **URL:** https://www.dawiso.com/connectors/databricks Connect Databricks to Dawiso for complete metadata visibility. Explore catalogs, schemas, models, functions, tables, and views in one unified data catalog. Includes Unity Catalog integration, automated metadata scans, and lineage extraction from Databricks notebooks and SQL warehouses. ### dbt Connector **URL:** https://www.dawiso.com/connectors/dbt Read-only dbt Cloud metadata connector for the Dawiso data governance platform. Catalogs accounts, projects, environments, models, tests, snapshots, seeds, sources, exposures and jobs; resolves model and source lineage via the dbt Cloud Metadata API into the warehouse. ### Google BigQuery Connector **URL:** https://www.dawiso.com/connectors/google-bigquery Read-only Google BigQuery metadata connector for the Dawiso data governance platform. Catalogs projects, datasets, tables, views, routines and IAM policy; resolves object-level lineage via Data Flow parsing; surfaces query and slot cost from INFORMATION_SCHEMA. ### JSON Connector **URL:** https://www.dawiso.com/connectors/json Read-only JSON file connector for the Dawiso data governance platform. Catalogs hierarchical JSON payloads, infers types and nested structure, and tracks schema drift between ingestion runs. ### Keboola Connector **URL:** https://www.dawiso.com/connectors/keboola Read-only Keboola metadata connector for the Dawiso data governance platform. Catalogs stacks, projects, buckets, tables, transformations, components and job history; resolves transformation-level lineage cross-platform. ### Metabase Connector **URL:** https://www.dawiso.com/connectors/metabase Read-only Metabase metadata connector for the Dawiso data governance platform. Catalogs collections, questions, dashboards, models and registered databases across self-hosted Metabase and Metabase Cloud; resolves object-level lineage to the underlying database. ### Microsoft Entra ID Connector **URL:** https://www.dawiso.com/connectors/microsoft-entra Read-only Microsoft Entra ID (Azure AD) connector for the Dawiso data governance platform. Syncs users, groups, memberships and roles from Microsoft Graph so catalog ownership, stewardship and RBAC stay aligned with the corporate directory. ### Microsoft Fabric Connector **URL:** https://www.dawiso.com/connectors/microsoft-fabric Read-only Microsoft Fabric metadata connector for the Dawiso data governance platform. Catalogs workspaces, lakehouses, warehouses, semantic models, reports, notebooks and pipelines via Power BI Admin API and Fabric API; resolves native + object-level advanced lineage. Metadata-only, never writes back. ### Microsoft SQL Server Connector **URL:** https://www.dawiso.com/connectors/sql-server Read-only Microsoft SQL Server metadata connector for the Dawiso data governance platform. Catalogs databases, schemas, tables, views, procedures and functions; resolves object-level lineage via Data Flow parsing; supports SQL Server on Windows/Linux, Azure SQL and RDS. ### MongoDB Connector **URL:** https://www.dawiso.com/connectors/mongodb Read-only MongoDB metadata connector for the Dawiso data governance platform. Catalogs clusters, databases, collections, fields and indexes from Atlas, self-hosted and on-premises MongoDB; links document fields to relational and BI catalogs. ### MySQL Connector **URL:** https://www.dawiso.com/connectors/mysql Read-only MySQL metadata connector for the Dawiso data governance platform. Catalogs databases, tables, views, columns, stored routines and triggers; resolves object relationships and dependencies from information_schema; supports self-managed MySQL 8.0+ and managed services (RDS, Aurora, Azure, Cloud SQL). ### Neo4j Connector **URL:** https://www.dawiso.com/connectors/neo4j Read-only Neo4j metadata connector for the Dawiso data governance platform. Catalogs the property graph schema (node labels, relationship types, properties), indexes and constraints from Aura, Enterprise self-hosted and on-premises Neo4j; links graph entities to relational and BI catalogs. ### OpenAPI / REST Connector **URL:** https://www.dawiso.com/connectors/openapi Read-only OpenAPI / Swagger schema connector for the Dawiso data governance platform. Catalogs REST API endpoints, parameters, responses and nested schemas with configurable $ref expansion depth and ownership. ### Oracle Connector **URL:** https://www.dawiso.com/connectors/oracle Read-only Oracle Database metadata connector for the Dawiso data governance platform. Catalogs schemas, tables, views, stored procedures, functions, indexes and constraints; resolves object-level lineage via Data Flow parsing; parses PL/SQL on enterprise plans. ### PostgreSQL Connector **URL:** https://www.dawiso.com/connectors/postgresql Read-only PostgreSQL metadata connector for the Dawiso data governance platform. Catalogs databases, schemas, tables, views, columns, constraints and functions; resolves object relationships and dependencies from the system catalog; supports self-managed Postgres and managed services (RDS, Aurora, Azure, Cloud SQL). ### Power BI Connector **URL:** https://www.dawiso.com/connectors/power-bi Connect Power BI to Dawiso for complete metadata visibility. Explore dashboards, reports, datasets, lineage, and permissions in one unified data catalog - including visual-level lineage that traces data flow down to every individual table, chart, and visual in your reports. ### Qlik Connector **URL:** https://www.dawiso.com/connectors/qlik Read-only Qlik metadata connector for the Dawiso data governance platform. Catalogs spaces, apps, sheets, master measures, load scripts and data connections across Qlik Cloud and Qlik Sense Enterprise; resolves app-to-data-connection relationships from app metadata. ### SAP HANA Connector **URL:** https://www.dawiso.com/connectors/sap-hana Read-only SAP HANA metadata connector for the Dawiso data governance platform. Catalogs schemas, tables, views, Calculation Views and procedures from on-prem HANA, HANA Cloud and S/4HANA backends; maps object dependencies from the HANA system catalog. ### Snowflake Connector **URL:** https://www.dawiso.com/connectors/snowflake The Dawiso Snowflake data catalog turns your account into a searchable inventory: every database, schema, table, view, pipe, stream and task, with column-level lineage that crosses platforms (Power BI visuals through Snowflake views and dbt models down to the raw ingest table). Read-only key-pair (RSA) authentication compatible with Snowflake's November 2025 password phase-out, dedicated DAWISO_INTEGRATION_ROLE with USAGE and REFERENCES grants plus SNOWFLAKE.OBJECT_VIEWER for dependency resolution. Tag and policy write-back to Snowflake keeps masking and row-access policies in sync; ACCOUNT_USAGE surfaces query-frequency and credit-cost insight per table. Complements Snowflake Horizon Catalog by adding cross-platform reach (Power BI, dbt, ML features) plus business glossary, ownership and certification workflows. ### SQL Server Reporting Services Connector **URL:** https://www.dawiso.com/connectors/ssrs Read-only SQL Server Reporting Services metadata connector for the Dawiso data governance platform. Catalogs folders, paginated reports, shared datasets, shared data sources, subscriptions and roles; links each report to the datasets and data sources it uses; supports SSRS-to-Power BI migration planning. ### SSAS Tabular Connector **URL:** https://www.dawiso.com/connectors/ssas-tabular Read-only SQL Server Analysis Services (Tabular) metadata connector for the Dawiso data governance platform. Catalogs semantic models, tables, columns, DAX measures, hierarchies, partitions and row-level security roles from TMSL, documenting relationships and partition data-source references; supports on-prem SSAS, Azure Analysis Services and Power BI Premium XMLA endpoints. ### Tableau Connector **URL:** https://www.dawiso.com/connectors/tableau Read-only Tableau metadata connector for the Dawiso data governance platform. Catalogs workbooks, dashboards, views, published data sources and calculated fields; resolves the relationships and dependencies between them from the Tableau Metadata API. Metadata-only, never writes back. ### Teradata Connector **URL:** https://www.dawiso.com/connectors/teradata Read-only Teradata Vantage metadata connector for the Dawiso data governance platform. Catalogs databases, tables, views, functions, indexes and referential constraints; resolves object-level lineage via Data Flow parsing; parses SQL on enterprise plans. ### Vertica Connector **URL:** https://www.dawiso.com/connectors/vertica Read-only Vertica metadata connector for the Dawiso data governance platform. Catalogs schemas, tables, projections, views, stored procedures and constraints from v_catalog; maps relationships from primary keys, foreign keys and constraints; reads PL/vSQL bodies via EXPORT_OBJECTS. ### XML Connector **URL:** https://www.dawiso.com/connectors/xml Read-only XML file connector for the Dawiso data governance platform. Catalogs elements, attributes and namespaces in FpML, HL7, NIEM and custom XML payloads; assigns ownership and tracks schema drift between ingestion runs. ## Academy ### Dawiso Academy **URL:** https://www.dawiso.com/academy Video lesson library for Dawiso users, covering platform basics, the business glossary, data lineage, data ingestion, data modeling, code lists, documentation, navigation and search, ownership and stewardship, user management and access control, productivity, and the Dawiso MCP server. Each lesson pairs a recording with a written concept explanation, the intended audience, and the capabilities a viewer should have afterwards. ### Business Glossary **URL:** https://www.dawiso.com/academy-lesson/business-glossary Ask three departments what 'active customer' means and you will often get three answers, each correct inside its own reporting. That is not a data quality problem, it is a definition problem, and no amount of pipeline work fixes it. A business glossary is the agreed definition of each term, owned by someone, and attached to the actual tables and reports that implement it. Once a metric links to its definition, a disagreement about a number becomes a question with a documented answer. ### Data Ingestion **URL:** https://www.dawiso.com/academy-lesson/data-ingestion Nobody documents an enterprise data estate by hand, and every attempt to has produced a catalog that was out of date before it was finished. Ingestion inverts the order: connect to a source system, scan it, and let the tables, columns and relationships populate themselves. What you get is technical metadata, which is accurate but not yet meaningful. The value appears when those scanned assets are linked to glossary terms, because that is the point a column name becomes a business concept someone can reason about. ### Data Lineage **URL:** https://www.dawiso.com/academy-lesson/data-lineage Lineage is the record of where a figure came from: which source systems fed it, what transformed it on the way, and which reports depend on it now. It answers the two questions that otherwise consume days. Backwards, when a number looks wrong, it shows what upstream to check instead of guessing. Forwards, before anyone changes a column, it shows exactly what will break. Under regimes such as GDPR and BCBS 239 it is also the evidence that you can demonstrate where a reported figure originated. ### Data Modeling **URL:** https://www.dawiso.com/academy-lesson/data-modeling A data model states how your data is meant to be structured: which entities exist, what attributes they carry, and how they relate. Most organizations have one, and most of them are wrong, because the diagram was drawn once and the database kept changing. A model that is validated against a database scan is a different artifact from a diagram in a slide deck. It can be trusted for impact analysis, because divergence from the real system shows up instead of going unnoticed. ### Data Organization with Code Lists **URL:** https://www.dawiso.com/academy-lesson/data-organization-and-code-lists A code list is a controlled vocabulary: a fixed set of allowed values that anyone classifying an object has to pick from. It sounds like a small thing and it decides whether classification is usable. Free-text tagging produces 'PII', 'pii', 'personal data' and 'sensitive' as four unrelated labels for one concept, so no filter, report or access rule can rely on any of them. A code list makes the same idea one value, which is what turns a label into something you can govern by. ### Dawiso MCP (Model Context Protocol) **URL:** https://www.dawiso.com/academy-lesson/dawiso-mcp-model-context-protocol An AI assistant pointed at a data warehouse can read the tables but not the meaning: it sees a column called rev_amt and has to guess. The Model Context Protocol is an open standard for handing a model that missing context through a defined interface. Dawiso MCP exposes your governed metadata over it, so an assistant answers from your documented definitions, ownership and lineage rather than from inference. The governance part is not incidental. The same layer that supplies the context also decides what may be exposed and leaves a record of what was asked. ### Dawiso Platform Basics **URL:** https://www.dawiso.com/academy-lesson/dawiso-platform-basics Most data governance tools fail for the same reason: nobody can find their way around them. Dawiso organizes everything into three nested ideas. A space is a workspace for a team or a domain, an app is a purpose-built area inside it such as the glossary or the catalog, and an object is a single documented thing such as a table, a term or a report. Learn those three and every other lesson in this series becomes navigation rather than discovery. ### Documentation **URL:** https://www.dawiso.com/academy-lesson/documentation Policies, standards and how-to guides usually live somewhere else: a wiki, a shared drive, an email thread. The cost is not the storage, it is the distance. A retention policy filed in a separate system is invisible at the moment someone is looking at the table it governs. Dawiso keeps documents next to the data they describe and links them to it, so context arrives with the object instead of having to be remembered and searched for separately. ### Navigation and Search in Dawiso **URL:** https://www.dawiso.com/academy-lesson/navigation-and-search A catalog is only worth what people can find in it. Documentation that exists but cannot be located is indistinguishable from documentation that was never written, and that is how most governance programs quietly stall. Dawiso offers two ways through: browsing the hierarchy when you know roughly where something lives, and searching with filters when you only know what it is. Knowing which to reach for is the difference between a ten-second answer and giving up. ### Ownership & Stewardship **URL:** https://www.dawiso.com/academy-lesson/ownership-stewardship Owner and steward are not the same job, and conflating them is why governance programs stall. The owner is accountable and decides: who gets access, whether a definition is correct, what the retention rule is. The steward does the work of keeping it accurate day to day. An asset with neither has no route to a decision, so questions about it circulate until someone gives up. Naming both, on the object itself, is what turns a catalog from a description of your data into something that can actually be maintained. ### Productivity Boost with Dawiso AI **URL:** https://www.dawiso.com/academy-lesson/productivity-boost Documentation debt is rarely a matter of unwillingness. It is that writing a clear description of the four hundredth table is slow, and the reward is invisible. That is the work AI assistance is genuinely good at: producing a serviceable first draft from context that already exists, which a human then corrects. The reviewing stays human, because a generated description nobody checked is worse than an empty field, which at least does not mislead. ### User Management & Access Control **URL:** https://www.dawiso.com/academy-lesson/user-management-access-control Access control in a catalog pulls in two directions. Lock it down and the documentation stops being read, which defeats the point of writing it. Leave it open and sensitive metadata reaches people who should not see it. Dawiso resolves this with roles that separate reading from contributing from administering, applied at whichever level fits: a whole space, a single object, a group or one person. Connecting an identity provider then means joiners and leavers are handled where they are already handled, rather than in a second list that drifts. ## Tools ### Total Cost of Ownership Calculator **URL:** https://www.dawiso.com/dawiso-total-operational-cost-calculator Interactive calculator that compares the total cost of ownership between open-source and commercial data catalog solutions over a 3-year period. Models six cost categories: initial deployment, infrastructure and hosting, personnel and training, integration and customization, maintenance and upgrades, and vendor support. Based on data from 50+ enterprise data catalog implementations. ## Key Pages - Pricing: https://www.dawiso.com/pricing - Glossary: https://www.dawiso.com/meta-data-glossary - The complete data governance glossary: 230+ terms covering data catalogs, business glossary, data lineage, AI governance, Databricks, SQL, and more, with category filters, A-Z navigation, and search - Data Contracts: https://www.dawiso.com/glossary/data-contracts - What are data contracts, why they matter for data products, and how open standards like ODCS (Open Data Contract Standard) and ODPS (Open Data Product Specification) enable governed, machine-readable agreements between data producers and consumers. Covers the open standards landscape, data contract components, and how Dawiso integrates contracts into its data products framework. - Agentic AI: https://www.dawiso.com/glossary/agentic-ai - What is agentic AI, how AI agents work through the observe-plan-act-evaluate reasoning loop, agentic patterns (single-agent, tool-augmented, multi-agent orchestration, autonomous workflows), and why data governance is a prerequisite for trustworthy agentic systems. Covers MCP integration, EU AI Act implications, and enterprise adoption requirements. - AI Governance: https://www.dawiso.com/glossary/ai-governance - What is AI governance, why it matters, and how the global regulatory landscape is diverging in 2026. Covers the EU AI Act delays (Parliament vote to push deadlines to 2027-2028), Trump administration's deregulation push and federal preemption of state AI laws, China's first-mover generative AI regulations and labeling mandates. Enterprise frameworks section covers NIST AI RMF and ISO/IEC 42001. Explains why AI governance requires strong data governance as its foundation. - AI Agent Governance: https://www.dawiso.com/glossary/ai-agent-governance - What AI agent governance is and how it differs from traditional model governance for systems that plan, call tools, and act autonomously. Covers the core controls (managed non-human identity, least-privilege access, runtime guardrails, observability and audit, human oversight, lifecycle management) and the emerging standards: the NIST AI Risk Management Framework, Singapore's IMDA Model AI Governance Framework for Agentic AI (January 2026), and OWASP. Explains why governed data context (catalog, glossary, lineage, classification) served via MCP is the foundation under all agent-governance controls. - AI Gateway: https://www.dawiso.com/glossary/ai-gateway - What an AI gateway (LLM gateway) is: a Layer 7 proxy between applications and LLM providers exposing one API to many models. Covers why teams need one (provider sprawl, cost, reliability, policy), the core functions (provider abstraction, routing and fallback, caching, rate limits, observability, access policy), how it differs from an API gateway, and where it sits relative to guardrails and MCP. Explains the key distinction that a gateway governs the request path while a context layer governs the meaning and trustworthiness of the data the model reasons over, and how Dawiso complements any gateway by serving governed context via MCP. - AI Guardrails: https://www.dawiso.com/glossary/ai-guardrails - What AI guardrails are: programmable, infrastructure-level checks that validate LLM inputs and outputs independently of the model. Covers input vs output guardrails, implementation approaches (rule-based, embedding-based, model-assisted), what they protect against (prompt injection, jailbreak, PII leakage, toxicity, hallucination, mapped to the OWASP LLM Top 10), and the distinction between guardrails as runtime enforcement and governance as the policy and context behind them. Explains how Dawiso supplies the governed classification, ownership, and business meaning that make guardrail enforcement correct. - Tool Calling (Function Calling): https://www.dawiso.com/glossary/tool-calling - What tool calling (function calling) is: the mechanism by which an LLM emits a structured JSON request to invoke an external function or API. Covers how the loop works (the model proposes, the application executes, the result feeds back), the terminology (tool calling vs function calling vs tool use), how it underpins AI agents, and how it differs from and complements the Model Context Protocol (MCP). Explains why tools are only as trustworthy as the data behind them and how Dawiso exposes governed catalog, glossary, and lineage as context any MCP-compatible agent can call. - Grounding in AI: https://www.dawiso.com/glossary/ai-grounding - What grounding is: connecting an AI model's output to verified external evidence at query time instead of relying on training memory. Covers why grounding matters (accuracy, currency, relevance, fewer hallucinations), the grounding methods (retrieval-augmented generation, search grounding, knowledge-graph grounding, tool use), the distinction between grounding and RAG (RAG is one method), and the failure mode where grounding on ungoverned data grounds the model in the wrong answer. Explains how Dawiso supplies the governed business meaning, lineage, and classification that make grounding trustworthy, served via MCP. - Non-Human Identity (NHI): https://www.dawiso.com/glossary/non-human-identity - What a non-human identity is: any digital identity used by software, services, workloads, or AI agents instead of a person (service accounts, API keys, tokens, IAM roles, bots). Covers why NHIs now outnumber human identities and form a fast-growing attack surface, the OWASP Non-Human Identities Top 10 (2025) with its top risks of improper offboarding and secret leakage, NHI governance and lifecycle (inventory, ownership, least privilege, rotation, offboarding), and why every AI agent is an NHI. Explains that Dawiso is not an identity tool but supplies the governed data context (what data exists, who owns it, how sensitive it is, where it flows) that informs what an NHI or agent should be allowed to touch. - Data Governance Framework: https://www.dawiso.com/glossary/data-governance-framework - What a data governance framework is (the operating structure that turns the discipline of data governance into daily practice) versus data governance itself. Covers the five core components (people and roles: owners, stewards, a governance council; policies and standards; processes; technology; metrics), the three operating models (centralized, decentralized, federated), the established references (DAMA-DMBOK with data governance at the center of its wheel, and the DCAM maturity model), and a step-by-step way to build one. Explains that a framework is only real when operationalized, and how Dawiso provides that operational layer (catalog, glossary, lineage, ownership, classification, quality) where the framework actually runs. - Context Governance: https://www.dawiso.com/glossary/context-governance - What context governance is: governing the meaning AI relies on (the definitions, relationships, and rules in a context layer) so it stays owned, versioned, validated, and consistent as systems change. Covers why it matters (agents act on context, so ungoverned meaning propagates wrong answers at machine speed), how it differs from data governance (which governs access, quality, and structure of the data rather than the meaning layered over it), what it covers (ownership, versioning, validation, consistency, access to context), and how it pairs with AI agent governance. Explains how Dawiso is where context governance lives, defining and governing meaning via the glossary, catalog, lineage, and classification, served to any MCP-compatible agent. - EU Data Act: https://www.dawiso.com/glossary/eu-data-act - What the EU Data Act (Regulation (EU) 2023/2854) is: a horizontal EU law governing who can access, use, and share data from connected products and cloud services. Covers its main pillars (user access to connected-product data by default and free of charge; fair B2B and B2G data-sharing terms; cloud switching without lock-in; interoperability and safeguards for trade secrets), the timeline (in force 11 January 2024, applies since 12 September 2025, with certain pre-existing cloud contracts having until 12 September 2027), the distinction from the Data Governance Act (Regulation (EU) 2022/868), and how it works alongside GDPR. Explains that compliance rests on knowing what data you hold, its sensitivity, and its flows, and how Dawiso's catalog, classification, and lineage provide that foundation. - Knowledge Graph: https://www.dawiso.com/glossary/knowledge-graph - What a knowledge graph is, why ontologies and typed relationships make a graph "knowledge," and how knowledge graphs are built (entities, triples, ontologies, triple stores vs property graphs). Covers the difference between knowledge graphs and relational models, GraphRAG and hybrid retrieval for enterprise AI, grounding LLMs with structured context, and the governance disciplines (ownership, provenance, change management) that keep enterprise knowledge graphs trustworthy. Explains how Dawiso uses a knowledge graph as the platform foundation exposed to AI agents via MCP. - Business Glossary: https://www.dawiso.com/glossary/what-is-a-business-glossary - What a business glossary is, why it's infrastructure rather than documentation, and what goes into a production-grade glossary (terms, definitions, ownership, links to data assets, standards and taxonomies). Covers the distinction between business glossary, data dictionary, and data catalog. Deep section on the relationship between business glossary, semantic layer, and context layer - the three layers that together produce trustworthy AI grounding: glossary for human meaning, semantic layer for computational meaning, context layer as the AI-facing composition exposed via MCP. Covers AI and analytics use cases, governance disciplines for building a glossary that stays useful, and how Dawiso integrates glossary, catalog, lineage, and MCP Server into the AI Context Layer. - Data Ownership: https://www.dawiso.com/glossary/data-ownership - Complete guide to data ownership in data governance. Covers what data ownership is, why it matters for AI and compliance, the difference between data owner, data steward, and data custodian roles, how to assign ownership using domain-based models and RACI matrices, and how to promote and sustain an ownership culture. - Data Quality Management: https://www.dawiso.com/glossary/data-quality-management - What is data quality management, the five-stage DQM lifecycle (define, measure, monitor, remediate, improve), six quality dimensions with metric examples, critical data elements, embedding quality into data pipelines, and translating quality metrics into business outcomes. Covers DAMA DMBOK alignment, regulatory drivers (GDPR, DORA, BCBS 239), and the shift toward platform-native DQ capabilities. - Data Stewardship: https://www.dawiso.com/glossary/data-stewardship - What is data stewardship, how it differs from data ownership and data custodianship, key responsibilities (standards, quality, metadata, access), and practical adoption strategies. Covers the "extend, don't invent" approach to building stewardship programs, finding and motivating stewards, using regulatory pressure (GDPR, DORA, EU AI Act) as a catalyst, and anchoring stewardship in business strategy. - Large Language Model (LLM): https://www.dawiso.com/glossary/large-language-model - What an LLM is, how transformer architecture and self-attention work, why context windows and tokenization matter, and the split between pre-training and fine-tuning. Covers enterprise use cases, the limits of LLMs (hallucination, stale knowledge, lack of grounding), and why structured context - knowledge graphs, semantic layers, governed catalogs - is what turns an LLM into a reliable enterprise system. - Prompt Engineering: https://www.dawiso.com/glossary/prompt-engineering - What prompt engineering is in 2026, the core techniques (zero-shot, few-shot, chain-of-thought, structured prompts, system prompts), why context quality matters more than prompt tricks, and how RAG grounded in governed data changes what prompt engineering means. Positions prompt engineering as one layer of a larger context-delivery problem. - AI Hallucination: https://www.dawiso.com/glossary/ai-hallucination - Why LLMs hallucinate - the statistical nature of next-token prediction, training-data gaps, and missing enterprise context. Covers types of hallucination (factual, contextual, fabricated citations), business risks (legal, reputational, operational), and the three technical strategies that reduce hallucination: RAG, fine-tuning, and governed knowledge graphs. Frames hallucination as a data problem, not a model problem. - Context Engineering: https://www.dawiso.com/glossary/context-engineering - Why context engineering is replacing prompt engineering as the core discipline for building reliable AI. Covers the distinction between prompt (the question) and context (everything else the model sees), context window management, retrieval pipelines, semantic ranking, and the governance layer that makes context safe to deliver. Positions context engineering as the applied discipline that sits on top of data governance. - Fine-Tuning an LLM: https://www.dawiso.com/glossary/fine-tuning - What fine-tuning is, the main techniques (full fine-tuning, LoRA, QLoRA, RLHF), when to fine-tune vs. use RAG, cost and complexity tradeoffs, and the governance requirements for training data (provenance, PII handling, consent, bias review). Explains why fine-tuning is a data governance problem as much as an ML problem. - Vector Database: https://www.dawiso.com/glossary/vector-database - What a vector database is, how embeddings turn text/images/audio into high-dimensional vectors, how approximate nearest-neighbor (ANN) indexes enable fast similarity search, and which databases matter in 2026 (Pinecone, Weaviate, Qdrant, pgvector, Milvus). Covers the RAG architecture, why embedding quality depends on governed source data, and when to use a vector DB vs. full-text search vs. a knowledge graph. - Multi-Agent Systems: https://www.dawiso.com/glossary/multi-agent-systems - What multi-agent systems are, the core orchestration patterns (hierarchical, peer-to-peer, blackboard, supervisor-worker), enterprise use cases (automated data pipelines, customer service, incident response), and the governance problems that emerge when agents have write access. Covers MCP as the emerging standard for agent-to-tool communication and why data contracts are the missing piece for reliable multi-agent production deployments. - LLMOps: https://www.dawiso.com/glossary/llmops - What LLMOps is, how it extends MLOps (prompt management, RAG pipelines, eval frameworks, cost monitoring, safety layers), the main platforms in 2026 (LangSmith, Langfuse, Arize, Weights & Biases), and the governance layer that's often missing. Covers why LLMOps without data governance produces unreliable systems and how to connect LLM operations to metadata and lineage. - AI Observability: https://www.dawiso.com/glossary/ai-observability - What AI observability means, why traditional APM tools aren't enough for LLM systems, the five pillars (input/output monitoring, quality evaluation, drift detection, cost tracking, governance audit), and how observability feeds back into continuous improvement. Covers the role of human-in-the-loop evaluation, automated eval pipelines, and why governed source data is what makes observability signals meaningful. - Responsible AI: https://www.dawiso.com/glossary/responsible-ai - What responsible AI is, the six core principles (fairness, transparency, accountability, privacy, safety, reliability), the global regulatory landscape (EU AI Act, US executive orders, China's generative AI rules, UK AI Safety Institute), and practical implementation frameworks (Microsoft RAI, Google PAIR, IEEE 7000 series). Covers why responsible AI is impossible without data governance as the foundation - you can't audit a model's behavior if you can't trace its training and context data. - Semantic Memory in AI Agents: https://www.dawiso.com/glossary/agent-memory-semantic - What semantic memory is for an AI agent (its store of facts, concepts, definitions, and relationships, borrowed from cognitive psychology), how it differs from episodic and procedural memory, and why it lives outside the model in vector stores, knowledge graphs, and catalogs rather than in frozen training weights. Covers how agents retrieve it via RAG, GraphRAG, and tool/MCP calls, why its quality decides answer trustworthiness (stale or ambiguous knowledge produces confident hallucinations), and how Dawiso supplies governed semantic memory by connecting glossary, catalog, and lineage into a context layer served via MCP. - Agent Harness: https://www.dawiso.com/glossary/agent-harness - What an agent harness is (the software infrastructure wrapped around an LLM that turns it into an agent, captured by the equation Agent = Model + Harness from LangChain and Databricks), what it contains (system prompt, tools and tool execution via MCP, sandboxes, filesystem, memory and context management, feedback loops, guardrails and human-in-the-loop controls, observability), and how it differs from an agent, tool calling, a framework, and a multi-agent system. Covers why the harness increasingly decides performance as frontier models converge (a strong harness on a mid-tier model can beat a weak harness on a stronger one), why the context slot is the most fragile part in the enterprise, and how Dawiso fills that slot with governed context served over MCP so the agent reasons on data with ownership, meaning, and lineage. - Agent Harness Engineering: https://www.dawiso.com/glossary/agent-harness-engineering - What agent harness engineering is (the discipline of designing, building, and tuning the infrastructure around an LLM so it acts as a reliable agent), how it relates to prompt engineering and context engineering, and why it emerged (as models converge, improving the harness pays off more than swapping models). Covers what harness engineers do (context engineering, tool design, memory architecture, feedback and verification, guardrails, observability and evals), the eval-driven iteration loop (run, observe, evaluate, adjust), the long-running agent problem and Anthropic's initializer plus coding-agent pattern, why every harness component encodes an assumption that goes stale as models improve, and why context is a harness-engineering concern best solved by wiring the context slot to Dawiso's governed context layer over MCP rather than hand-maintained prompt strings. - Episodic Memory in AI Agents: https://www.dawiso.com/glossary/agent-memory-episodic - What episodic memory is (an agent's record of specific past events and interactions, time-stamped, as opposed to timeless facts), how it turns a stateless LLM into something with continuity, and the mechanisms that implement it (conversation persistence, vector retrieval of similar episodes, summarization and consolidation). Covers its relationship to semantic and procedural memory, the governance risks that make it the most exposed memory type (PII and GDPR retention, stale recall, memory poisoning), and why episodic recall is only correct when grounded in governed semantic memory from a context layer. - Procedural Memory in AI Agents: https://www.dawiso.com/glossary/agent-memory-procedural - What procedural memory is (an agent's learned skills, workflows, and tool-use routines, the "knowing how" as opposed to "knowing that"), how it is realized through system prompts, reusable skills and tools exposed via MCP, and learned workflows, and how it interlocks with semantic and episodic memory. Covers why it is the memory type with the most real-world consequences (it acts on systems), the two governance concerns (what a procedure acts on and what it is allowed to do), and how Dawiso governs the data procedures run against via a context layer and MCP. - AI Debt: https://www.dawiso.com/glossary/ai-debt - What AI debt is (the accumulated cost and risk of deploying AI without the governance, data quality, and oversight to sustain it, the AI-era cousin of technical debt), why it is mostly data and governance debt in disguise, and where it comes from (ungoverned data, context islands, undocumented prompts and pipelines, no evaluation or monitoring, missing oversight and compliance). Covers why it compounds as more AI is stacked on a weak foundation, how to pay it down (govern data first, consolidate context, instrument and monitor, add oversight), and how Dawiso AI Governance and a governed context layer attack it at the root. - Guardian Agents: https://www.dawiso.com/glossary/guardian-agents - What guardian agents are (AI agents that oversee other AI agents - validating outputs, enforcing policy, detecting anomalies, and containing or escalating unsafe actions), popularized by Gartner as a defining agentic-AI pattern, and how they differ from static guardrails (they reason about behavior in context). Covers why they are emerging now (scale and speed, new risk surfaces like prompt injection, regulatory oversight demands), why they depend on governed ground truth to judge against, and how Dawiso AI Governance and a context layer supply the policies, classifications, and trusted context a guardian agent enforces against via MCP. - Human-in-the-Loop (HITL): https://www.dawiso.com/glossary/human-in-the-loop - What human-in-the-loop is (a person as a required step inside the AI decision loop, approving actions before they take effect), how it sits on the oversight spectrum between human-on-the-loop and full autonomy, and when to use it (high-impact, regulated, ambiguous, or sensitive-data decisions). Covers the EU AI Act and GDPR drivers, the importance of selectivity, the pattern of a guardian agent triaging routine cases and escalating consequential ones, and why HITL is only real oversight when the reviewer sees the same governed context (definitions, lineage, policy) the AI used - which Dawiso supplies via a context layer and MCP. - Human-on-the-Loop (HOTL): https://www.dawiso.com/glossary/human-on-the-loop - What human-on-the-loop is (AI acting autonomously while a person supervises and can intervene or override at any time), how it differs from human-in-the-loop (no mandatory gate on each action) and full autonomy, and when to use it (high-volume, time-sensitive, recoverable work in mature, well-monitored systems). Covers the control-versus-throughput trade-off, the blended pattern of HOTL with escalation to HITL for high-stakes cases, why HOTL lives or dies on observability (the supervisor can only act on what they can see and understand), and how Dawiso's governed context and AI Governance make supervision meaningful. - Decision Intelligence: https://www.dawiso.com/glossary/decision-intelligence - What decision intelligence is (Gartner's practical discipline of engineering how decisions are made and how their outcomes are evaluated and improved by feedback), augmenting data science with decision theory, social science, and managerial science. Covers the decision loop (data and context to decision model to decision and action to measured outcome to feedback); how DI differs from BI (decision and action plus a closed feedback loop, not just a dashboard); the support/augment/automate spectrum increasingly delivered by agents; why its quality is bounded by governed, trustworthy, traceable data; and how Dawiso's context layer supplies that foundation via MCP so decisions run on data they can trust. - Data Provenance: https://www.dawiso.com/glossary/provenance - What data provenance is (the documented origin and full history of data: source, transformations, the agents responsible, and authenticity/trust), and how it differs from data lineage (lineage maps the technical path; provenance adds who, why, and authenticity - lineage is a subset of provenance). Covers why it matters (trust, compliance and audit under GDPR and the EU AI Act, AI training-data accountability, reproducibility, debugging and impact analysis); the W3C PROV standard (entities, activities, agents, evolved from the 2011 Open Provenance Model); and how Dawiso captures provenance through cross-platform interactive data lineage plus the catalog and glossary, available to AI via MCP. - ETL vs ELT: https://www.dawiso.com/glossary/etl-elt - What ETL and ELT mean, how they differ in architecture and governance implications, why ELT won in the cloud era (Snowflake, BigQuery, Databricks), and when ETL is still the right choice. Covers the dbt-driven "transformation in the warehouse" pattern, lineage and data contracts as cross-cutting concerns, and why the ETL/ELT label matters less than knowing where transformation happens and who governs it. - Data Pipeline: https://www.dawiso.com/glossary/data-pipeline - What a data pipeline is, the main pipeline architectures (batch, streaming, lambda, kappa), orchestration tools (Airflow, Dagster, Prefect, Mage), and the disciplines that keep pipelines reliable in production: lineage, observability, data contracts, and SLAs. Covers the shift from imperative pipelines to declarative transformations and why governance has to be built into the pipeline, not bolted on after. - Data Lakehouse: https://www.dawiso.com/glossary/data-lakehouse - What a data lakehouse is, how it differs from data warehouses and data lakes, the open table formats that made it possible (Delta Lake, Apache Iceberg, Apache Hudi), and the main platforms in 2026 (Databricks, Microsoft Fabric, Snowflake Iceberg). Covers the governance problem lakehouses create - one storage layer, many query engines - and why active metadata and unified catalogs matter more in a lakehouse than in a warehouse. - DataOps: https://www.dawiso.com/glossary/dataops - What DataOps is, the seven core principles (automation, observability, continuous delivery, testing, version control, self-service, lean thinking), the main tools (dbt, Dagster, Great Expectations, Soda), and how DataOps connects to data governance. Covers why DataOps without governance produces fast pipelines delivering unreliable data, and how modern platforms integrate quality, lineage, and contracts into the CI/CD layer. - Data Discovery: https://www.dawiso.com/glossary/data-discovery - What data discovery means, why it's a prerequisite for self-service analytics and enterprise AI, how metadata-driven catalogs automate discovery (search, classification, tagging, lineage), and why AI is changing discovery UX (natural language queries, MCP-powered agents). Covers the difference between active and passive discovery, organizational patterns for sustaining discoverability, and the role of governance in keeping catalog content trustworthy. - Data Fabric: https://www.dawiso.com/glossary/data-fabric - What data fabric is (Gartner's definition, the architectural layers), how it uses active metadata and knowledge graphs to automate data integration, and how it compares to data mesh. Covers the shift from passive to active metadata, the role of AI in fabric orchestration, and why governed catalogs are the foundation of any fabric deployment. - Data Mesh: https://www.dawiso.com/glossary/data-mesh - What data mesh is, Zhamak Dehghani's four principles (domain ownership, data as a product, self-serve platform, federated governance), and how mesh organizations are structured. Covers real implementations at Netflix, Intuit, and Zalando, the team topology challenges mesh creates, and why federated governance requires shared catalogs and contracts. - Column-Level Lineage: https://www.dawiso.com/glossary/column-level-lineage - What column-level lineage is, how it differs from table-level lineage, the techniques for extracting it (SQL parsing, query logs, metadata APIs), and the regulatory drivers (BCBS 239 for banks, GDPR Article 30 for PII mapping, EU AI Act for AI training data). Covers impact analysis workflows, compliance reporting patterns, and why column-level is non-negotiable for financial services and AI governance. - Master Data Management: https://www.dawiso.com/glossary/master-data-management - What MDM is, the four architectural styles (registry, consolidation, coexistence, centralized), the "golden record" concept, and how MDM fits into modern data platforms. Covers the convergence of MDM and data catalogs in 2026, the role of AI in entity resolution, and why MDM without data governance produces fast but untrustworthy master data. - Unstructured Data: https://www.dawiso.com/glossary/unstructured-data - What unstructured data is, why it makes up 80-90% of enterprise data, the main types (documents, emails, images, audio, video, code), and the AI-era techniques that finally make it usable (embeddings, RAG, multimodal models). Covers the governance problem unstructured data creates (PII discovery, access controls, retention) and why unstructured governance is now the fastest-growing part of enterprise catalogs. - Business Glossary Implementation Guide: https://www.dawiso.com/glossary/business-glossary - How to build a business glossary from scratch: term discovery (interviews, document mining, pattern extraction), governance workflows (approval, versioning, deprecation), ownership models (steward-led, federated, hybrid), and how glossaries scale from 50 terms to 5000+. Covers the integration with data catalogs, lineage, and MCP-powered AI agents that query the glossary as a semantic layer for enterprise AI. - Data Democratization: https://www.dawiso.com/glossary/data-democratization - What data democratization means, the three prerequisites (discoverability, understandability, accessibility), the failure modes when democratization happens without governance (shadow analytics, inconsistent metrics, compliance drift), and the role of self-service catalogs and semantic layers in making democratization safe. Covers cultural and organizational dimensions beyond tooling. - Data Privacy: https://www.dawiso.com/glossary/data-privacy - What data privacy is, the main regulatory frameworks (GDPR, CCPA/CPRA, LGPD, POPIA, India DPDPA), PII and sensitive data classification, technical controls (encryption, tokenization, anonymization, pseudonymization, differential privacy), and organizational privacy governance. Covers the role of data catalogs in PII discovery, access control enforcement, and subject access request (SAR) fulfillment. - DORA (Digital Operational Resilience Act): https://www.dawiso.com/glossary/dora-digital-operational-resilience-act - The EU regulation establishing binding requirements for ICT risk management, incident reporting (4h/72h/1mo deadlines), digital operational resilience testing (including TIBER-EU TLPT), third-party risk management (Register of Information, CTPP oversight), and information sharing across financial entities. Covers scope (banks, insurers, payment institutions, crypto-asset service providers, and Critical Third-Party Providers like cloud platforms), the five pillars, penalties up to 2% of global annual turnover for entities and 1% of daily turnover for CTPPs, and the data governance backbone (catalog, lineage, classification, ownership, third-party register, audit trail) that makes DORA compliance operable. Compares DORA with BCBS 239, GDPR, and NIS2. - NIS2 Directive: https://www.dawiso.com/glossary/nis2-directive - The EU directive (2022/2555) modernizing cybersecurity obligations across 18 sectors and ~160,000 essential and important entities, transposed nationally since October 17, 2024. Covers scope (energy, transport, banking, health, water, digital infrastructure, manufacturing, food, chemicals, postal, research, public administration), the ten Article 21 risk management measures (policies, incident handling, business continuity, supply chain, secure development, cyber hygiene, cryptography, asset management, MFA), the four-stage incident reporting timeline (24h early warning, 72h notification, intermediate report, 1 month final), and the personal liability of management bodies. Compares NIS2 with DORA (lex specialis for finance), GDPR (parallel breach notification), and CRA. Maps Article 21 obligations to the data governance backbone (asset inventory, catalog, lineage, classification, ownership, supplier register, audit trail). - PII (Personally Identifiable Information): https://www.dawiso.com/glossary/pii-personally-identifiable-information - The complete PII reference: the NIST 800-122 definition, the difference between direct identifiers (name, SSN, passport), quasi-identifiers (DOB + ZIP + gender), online identifiers (IP, cookie, AAID), financial PII, health PII (PHI), biometric and genetic PII, and behavioral/inferred PII generated by AI systems. Compares PII (US terminology) with GDPR "personal data" (EU terminology) and explains why the broader EU definition is the safer working definition for governance. Covers sensitive PII categories, the major regulations (GDPR, UK GDPR, CCPA/CPRA, HIPAA, GLBA, PCI-DSS, LGPD, PIPL, POPIA, India DPDPA, US state laws), and the four-capability operating model for PII governance at scale: discovery and classification, lineage across systems including unstructured data, role-based access with masking, and documented ownership. - CISO (Chief Information Security Officer): https://www.dawiso.com/glossary/ciso-chief-information-security-officer - The senior executive accountable for information security strategy, cyber risk, and regulatory readiness. Covers the eight core responsibility domains (security strategy and governance, cyber risk management, identity and access, security architecture, vulnerability and threat management, detection and response/SOC, third-party and supply chain security, compliance and assurance), the distinction between CISO and CIO/CSO/CRO/CDO/DPO, the modern board-facing mandate under SEC cyber disclosure rules, NIS2 management liability, and DORA Article 5, the typical reporting structures (CIO/CEO/CRO/General Counsel), and how the CISO function depends operationally on a governed data catalog, lineage, classification, ownership, supplier register, and audit trail. Frames the CDO and CISO as consumers of the same governed view of data from different lenses. - AI Data Products: https://www.dawiso.com/glossary/ai-data-products - Data products engineered specifically for reliable consumption by AI systems - LLMs, agentic workflows, ML pipelines, and natural-language analytics. Covers what an AI data product is (extends the traditional DAUTNIVS properties with semantic annotation, programmatic discoverability, and runtime trust-evaluability), why traditional human-centered data products fail for AI (schema-without-semantics, free-text docs, missing provenance, no embedded access policy, opaque quality), the seven-component anatomy (data and schema, semantic context from a glossary, lineage and provenance, embedded access policy, ownership, machine-readable quality and freshness signals, agent-friendly MCP interface), how AI data products relate to RAG, vector databases, semantic layers, and plain datasets, the practical build path (expose catalog/glossary/lineage via MCP, define a small set of high-value products, treat agent behavior as part of QA), and the three governance disciplines (classification, lineage, ownership) that AI data products intensify. - Data Product Lifecycle: https://www.dawiso.com/glossary/data-product-lifecycle - The five-stage lifecycle (Discover, Design, Build, Operate, Retire) that turns data products from one-off projects into a managed portfolio. Covers what each stage produces (documented need, signed data contract, live product with lineage and quality monitors, ongoing operation with versioned changes, deprecation and consumer migration), the five recurring roles (consumer, product owner, domain steward, data engineer, data governance), governance artifacts produced at each stage (glossary terms, contract, classification, lineage, ownership, SLAs, audit trail), the four common anti-patterns (build-first, permanent operate, governance after the fact, lifecycle theatre), and the role of a data products platform in making the lifecycle scalable beyond a small team. - Data Consumer: https://www.dawiso.com/glossary/data-consumer - The third leg of the owner/steward/consumer governance triad. Covers what a data consumer is (any person, team, application, or system that uses data to make decisions or run processes), the seven recurring consumer types (analysts and BI users, data scientists and ML engineers, non-technical business users, operational applications, AI agents and LLMs, regulators and auditors, customers and partners), the four universal consumer needs (find via the catalog, understand via the glossary, trust via lineage and quality, access via governed self-service), how consumer/owner/steward roles form a feedback loop, why consumer-driven governance outperforms policy-driven governance, and three patterns for enabling consumers at scale (single source of truth for context, governance embedded in the consumption interface, AI agents treated as first-class consumers). - Data Engineering: https://www.dawiso.com/glossary/data-engineering - The discipline of designing, building, and operating the systems that ingest, store, transform, and deliver data reliably at scale. Covers the six recurring activities (ingestion via Fivetran/Airbyte/CDC/Kafka, storage on Snowflake/Databricks/BigQuery/Iceberg, transformation with dbt and SQL/Spark, orchestration with Airflow/Dagster/Prefect, observability and quality with Monte Carlo/Soda/dbt tests, serving via SQL/semantic layers/feature stores/MCP), the distinctions between data engineering, analytics engineering, and data science, the modern data engineering stack (cloud-native separated compute and storage, SQL-first transformations, code-first orchestration, observability as first-class, governance native to the platform), where data engineering meets governance (lineage from transformations, quality engineered into pipelines, classification propagation, access policy enforcement), and the skill profile and career path of a modern data engineer. - Data Integrity: https://www.dawiso.com/glossary/data-integrity - The property that data remains accurate, consistent, and unaltered across its lifecycle. Covers the four classical types (entity integrity via primary keys, referential integrity via foreign keys, domain integrity via type/range/CHECK constraints, user-defined integrity via triggers and business rules), the distinction between integrity and data quality and data accuracy, the eight categories of threat (hardware corruption, software bugs, pipeline transformation errors, schema drift, replication and sync issues, unauthorized modification including ransomware, incomplete transactions, human error), preservation mechanisms (database constraints, ACID transactions, dbt and Great Expectations testing, lineage-driven reconciliation, audit logs, immutable WORM storage, cryptographic signing, removing direct production access), and the regulated contexts where integrity is a legal requirement (BCBS 239 Principle 3, FDA 21 CFR Part 11 with ALCOA+ principles, SOX, GDPR Article 5(1)(d), DORA Article 9, HIPAA Security Rule, manufacturing GMP). - Data SLA: https://www.dawiso.com/glossary/data-sla - A formal commitment to deliver a data product with measurable characteristics. Covers the six dimensions of a data SLA (freshness as refresh frequency or max staleness, accuracy as match percentage against source, completeness as expected row count and NULL rates, availability in nines during published hours, consistency across joined tables and reports, schema stability with notice period for breaking changes), the relationship between data SLA and data contract (contract is the broader agreement, SLA is the testable performance section), the SLI/SLO/SLA triad borrowed from SRE practice (SLI is the measurement, SLO is the internal target, SLA is the external commitment), error budgets as a self-regulating mechanism, the five-step process for defining and operating SLAs (negotiate with consumers, instrument the data, set SLOs tighter than SLAs, define breach response, review quarterly), and how SLAs integrate with governance (catalog display, named ownership, audit trail of compliance, tiered targets by product impact). - Data Access Management: https://www.dawiso.com/glossary/data-access-management - The discipline that controls who can read, modify, and share data across an organization. Covers the four access models (RBAC role-based control, ABAC attribute-based control, PBAC policy-as-code with OPA/Cedar/XACML, ReBAC relationship-based with Zanzibar/OpenFGA), the six recurring components (identity integration with SSO/MFA, classification-driven policy where tags drive access, dynamic data masking and tokenization, approval workflows tied to data ownership, periodic access reviews, immutable audit trail), the distinction between data access management, IAM (broader auth across applications), and IGA (governance over IAM), the six-step implementation sequence (catalog and classify first, assign owners, pick the access model, centralize policy expression, enforce at the data layer, instrument and review), and how the discipline is now a legal requirement under GDPR Article 32, NIS2 Article 21(2)(i), DORA Article 9, SOX, and the EU AI Act. - Unstructured Data Governance: https://www.dawiso.com/glossary/unstructured-data-governance - The application of data governance disciplines to documents, emails, support tickets, transcripts, audio, video, images, and code that does not fit a relational schema. Covers why the discipline became urgent (volume and dispersion across SharePoint/OneDrive/Drive/Slack/Teams/Confluence/Box/Salesforce/Jira/email/S3/code repos, the AI inflection point with RAG and embedding-based retrieval, regulatory parity under GDPR/HIPAA/NIS2/EU AI Act), the six recurring capabilities (cross-system content discovery, AI-assisted content classification, location- and file-level ownership, lineage including AI ingestion paths, content-aware access control beyond file-system permissions, automated retention and disposition), the five differences from structured data governance (content-driven classification, document/embedding lineage, multi-layered access control, automation-first economics, fuzzier quality), the five-step operationalization pattern (inventory sources, automated classification, location-level ownership, retention automation, AI ingestion governance), and the specific governance pattern for RAG and AI deployment (classify before indexing, carry classification with embeddings, filter at retrieval, log every retrieval, govern prompts and outputs). - Databricks AI/BI Genie: https://www.dawiso.com/glossary/databricks-ai-bi-genie - The natural-language analytics interface in the Databricks Data Intelligence Platform. Users ask questions in plain English; Genie generates SQL, runs it against the lakehouse with the user's Unity Catalog permissions, and returns visualizations and follow-up support. Covers what a Genie space is (a curated bundle of tables, certified queries, instructions, and metadata scoped to a business domain), the six-stage query flow (context loading, query interpretation, SQL generation with schema validation, permission-aware execution, result presentation, conversation continuity), key features (curated spaces, certified queries, Unity Catalog grounding, SQL transparency, permission-aware execution, conversation memory), comparison with Power BI Copilot, Snowflake Cortex Analyst, and independent semantic layer products, the five load-bearing governance investments (business glossary, table/column descriptions, certified queries as training-by-example, classification and access policy, named ownership of each space), and a practical five-step implementation pattern (start with a high-value domain, invest in metadata first, treat early users as feedback loop, govern access carefully before broad exposure, audit and iterate). - Databricks Genie One: https://www.dawiso.com/glossary/databricks-genie-one - Databricks' agentic AI coworker, announced at Data + AI Summit on June 16, 2026 and generally available, that automates and orchestrates work across structured and unstructured, analytical and operational data, inside or outside Databricks, and produces documents, reports, and artifacts rather than just answers. Frames Genie One as the agentic evolution of the unified Databricks workspace experience (formerly Databricks One). Covers the Genie family (Genie One, Genie Agents, Genie App Builder, Genie Code, Genie ZeroOps), all governed by Unity Catalog and grounded in Genie Ontology; Ali Ghodsi's framing that the limit on enterprise AI is "not an AI problem, that's a context problem"; the cross-platform gap (Genie One governs context inside Databricks, but the business runs on more than one platform and agents are not all Databricks-native); and how Dawiso adds a cross-platform business glossary, lineage, and classification served to any MCP-compatible agent via the MCP Server. - Databricks Genie Ontology: https://www.dawiso.com/glossary/databricks-genie-ontology - Databricks' live, continuously learned enterprise context layer in the Databricks Platform that grounds Genie in business meaning, announced at Data + AI Summit 2026. Covers what it is (learns concepts, terms, metrics, and relationships from Databricks data, dashboards, queries, and connected apps; signals like column popularity feed back into it; Databricks says it makes Genie's answers more accurate, faster, at lower token cost); how it is fed by three new Unity Catalog semantic capabilities (Glossary, preview coming soon; Domains, public preview; Metrics, public preview), with the rule that the more semantics you model in Unity Catalog the better Genie performs; the clarification that it functions as a knowledge graph and context layer more than a formal ontology; the cross-platform gap (it is bounded to Databricks and its connected apps and serves Databricks' own agents); and how Dawiso provides a vendor-neutral, cross-platform context layer that defines meaning once, governs it across the estate, and serves it to any agent via open MCP. - Databricks CustomerLake: https://www.dawiso.com/glossary/databricks-customerlake - Databricks' agentic Customer Data Platform (CDP) embedded natively in the lakehouse and governed by Unity Catalog, announced at Data + AI Summit on June 16, 2026 and in private preview (early customers HP, Circle K, Getnet by Santander, and Zé Delivery / AB InBev). Covers what it unifies (customer 360, identity resolution including agentic identity resolution, audience building, campaign automation, activation, and reverse ETL via Lakehouse Federation, plus an open partner ecosystem); what makes it agentic (campaign and profile agents and "infinity campaigns" - continuous, agent-driven engagement loops that analyze customer signals, decide the next-best action, and act across channels in real time); why governed customer data decides the outcome (shared meaning via a business glossary, a trustworthy single view via master data management discipline, and classification of sensitive fields); the cross-platform gap (customer data and its meaning span CRM, ad platforms, and other warehouses); and how Dawiso governs customer definitions, classification, and lineage across the whole estate and serves them to any MCP-compatible agent. - Synthetic Data: https://www.dawiso.com/glossary/synthetic-data - What synthetic data is, the main generation techniques (statistical models, GANs, VAEs, LLM-based generators, rule-based simulators), primary use cases (ML training, software testing, privacy-preserving analytics, data augmentation), and quality/utility tradeoffs. Covers the governance questions synthetic data creates (provenance, drift from real data, regulatory acceptance) and why synthetic data programs still need governed source data to produce useful outputs. - Blog: https://www.dawiso.com/blog - [Why European Groups Need a Multilingual Business Glossary](https://www.dawiso.com/blog-post/multilingual-business-glossary-european-groups) - A point-of-view article for international, especially European, groups arguing that most business glossaries carry a hidden assumption - that everyone reads the definitions in English - which holds for a US company but rarely for a European group, since the EU has 24 official languages and, per Eurostat, multinational enterprise groups employ around 30% of the European business economy with most large groups operating in more than six countries. Its core claim is that when a glossary speaks only English it removes ambiguity for the people who share that language and quietly leaves it in place for everyone else: a term defined once in English gets informally translated in each reader's head, a controller in one country and an analyst in another map "net revenue" to slightly different local senses, and the reported number drifts because the shared definition never existed in their language. It draws a precise line between translating the glossary interface (a UI feature that changes labels, not meaning) and a real multilingual glossary that holds the definition itself in more than one language: one governed concept with a parallel, approved definition per language, linked as versions of the same thing rather than separate glossaries that diverge, with synonyms, acronyms, and local variants attached to the single concept so search resolves to the same definition whatever word someone starts from. It explains how this works in Dawiso - a business term carrying parallel language sections with a single-click switch, one owner and one approval behind every language view, synonyms and variants as first-class links, and an entity layer for concepts that need structure above the term such as a geography or product hierarchy that must hold across languages. It lists what a multilingual glossary solves: onboarding across countries in the reader's own language, one version of a number that means the same thing wherever it is reported, governance and audit evidence readable in the local language of the jurisdiction, and less quiet reinterpretation, since meaning drift is usually a hundred small private translations that never get compared. Two light-palette SVG diagrams (an English-only glossary where one definition drifts into three local meanings versus a multilingual glossary where one governed concept holds three approved parallel definitions; and one governed concept with parallel English and German definitions joined by a one-click language toggle) and a four-question FAQ with FAQPage schema. No client named; sourced to Eurostat and the EU official-languages page. CTA to the Business Glossary. Author: Samuel Nagy (VP of Strategic Growth, August 2026). - [Knowledge Graph Use Cases and How to Build One From Text](https://www.dawiso.com/blog-post/knowledge-graph-use-cases-how-to-build-from-text) - A product-educational guide to knowledge graphs inside a data catalog, deliberately pitched at how you use a knowledge graph rather than what one is, so it complements Dawiso's older "what is a knowledge graph and why it beats relationship diagrams" explainer instead of cannibalizing it. It opens in plain language (every thing you track becomes a node, every meaningful connection a labeled edge such as "this metric uses that term", "this report is owned by that person", "this document explains that concept", so the connections carry as much meaning as the things they connect) and links out to the older explainer for the deeper diagram contrast. It then separates a knowledge graph from data lineage, which are constantly confused because both draw boxes and arrows: lineage maps the automated technical flow of data (source table to view to report, extracted from system metadata, answering where a number came from and what breaks downstream), while a knowledge graph maps semantic and organizational relationships (what a term means, which synonym belongs to it, who owns it, which document explains it, which policy governs it), and lineage is in effect one type of edge inside the wider graph. It covers six everyday use cases: onboarding (a new hire explores a connected web instead of reading a table of definitions), accountability and ownership (owners and stewards are nodes, so "who do I ask?" is visible), reconciling synonyms and shared terminology across teams and languages (finance's "revenue" and a subsidiary's local term link to one governed concept - where a multilingual glossary and a knowledge graph reinforce each other), linking documentation to the business concept it explains, cross-domain dependencies across space boundaries, and impact analysis as a graph traversal rather than a guess. The central section explains how to build a knowledge graph from text without hand-modeling every node: autolinking recognizes glossary terms inside the descriptions people already write, matching across different word forms, and creates a "mentions" edge automatically so the graph grows as a byproduct of documentation; and Dawiso can generate a draft graph directly from a document or block of text, proposing the terms and how they connect for a data steward to review, adjust, and approve, so speed comes from automation and trust from human sign-off. It closes on keeping the graph current for a live AI system, where a stale graph does real damage because an agent answers confidently from relationships that no longer hold: tie the graph to living metadata (autolinking re-evaluates as descriptions change), keep definitions governed through stewardship and approval, and serve the governed graph to any MCP-compatible agent through the Context Layer, the approach known as GraphRAG - citing Microsoft's GraphRAG research (extracting a knowledge graph from source text and grounding an LLM on it improves answers to complex questions while preserving provenance) and a 2025 survey of graph-based retrieval. Three light-palette SVG diagrams (anatomy of a knowledge graph; knowledge graph versus data lineage; text becoming a graph via autolinking), an embedded YouTube demo of generating a knowledge graph from text in Dawiso, and a four-question FAQ with FAQPage schema answering what a knowledge graph is in simple language, its use cases, how to build one from text, and how to keep it current in a live AI system. CTA to the Business Glossary. Author: Michal Peroutka (Product Manager, August 2026). - [The ROI of Data Governance, and How to Prove It](https://www.dawiso.com/blog-post/data-governance-roi) - A business-case editorial for data and finance leaders arguing that data governance has a funding problem not because the value is missing but because it is diffuse (spread across every team that finds data faster, trusts a number sooner, or passes an audit without a fire drill) while the cost is a single clean line item, so it gets framed as a cost center and cut first, when the real issue is that the return was never measured. It puts a number on the cost of doing nothing: Gartner has long put the average cost of poor data quality at around $12.9 million per organization per year, MIT Sloan Management Review with Cork University Business School estimates firms lose 15-25% of revenue to bad data, and IDC puts the productivity a governed platform recovers from discovery-and-validation time at roughly EUR 1,572 per affected user per year, a per-person figure that multiplies across everyone who touches data. It then locates the return in three streams mapped to what governance actually delivers: productivity (a data catalog makes trusted data findable, a business glossary makes a term mean one thing so teams stop reconciling versions of revenue), risk mitigation (column-level lineage turns compliance evidence into a query rather than a quarterly project, ownership and classification cut exposure, decisive for BCBS 239, DORA, GDPR, and GxP), and AI enablement (governed context lets AI answer from your data and ship instead of stalling in pilot). It draws one honest boundary: governance is not a data quality engine and the return does not come from cleaning values; a governed catalog makes quality issues visible, assigns an owner, and feeds the specialized data quality tools you already run, making them more effective rather than replacing them, so the return is find, trust, trace, and own. It gives a one-line estimate framework (net annual return = impacted users x recoverable value per user + risk avoided - platform cost), explains each input, and notes that per-user pricing keeps the cost side clean and predictable versus per-connector or per-asset models. It closes on making the business case stick: present a conservative, realistic, and optimistic range rather than one number, and measure leading indicators from day one (adoption, time-to-trusted-data, share of critical assets with a named owner), because adoption is the multiplier on the entire model and a platform nobody opens returns nothing regardless of features, which is why ease of use and roll-out-to-everyone pricing are not soft factors. Two dark-palette SVG diagrams (three benefit streams minus the platform cost producing the net annual return; the ROI estimate as a single formula) and a three-question FAQ with FAQPage schema, including one that distinguishes governance ROI from data quality ROI. No competitors named. CTA to book a Dawiso demo, cross-linking pricing and the catalog, glossary, and lineage products. Author: Samuel Nagy (VP of Strategic Growth, August 2026). - [Data Integrity and Audit-Ready Lineage for GxP AI](https://www.dawiso.com/blog-post/data-integrity-audit-ready-lineage-gxp-ai) - A point-of-view editorial for regulated pharma and life sciences arguing that the AI question in GxP is really a data integrity question, answered with governed context and audit-ready lineage rather than with the model. It opens on the regulatory shift: on 7 July 2025 the European Commission, with the PIC/S inspectorate, published draft revisions of Annex 11 (computerised systems), a new Annex 22 written specifically for artificial intelligence, and an updated Chapter 4 (documentation); consultation closed 7 October 2025, finals are expected mid-2026, and Annex 11 alone grew from five pages to nineteen, treating cybersecurity, cloud qualification, and identity and access management as core requirements and expecting audit trails to be reviewed regularly and never disabled without documented justification. It explains ALCOA+ (Attributable, Legible, Contemporaneous, Original, Accurate, plus Complete, Consistent, Enduring, Available) and where AI strains it: a model output is hard to call attributable when no one can say which data and version produced it, hard to call original or accurate when the training and reference data were never traced, and hard to keep consistent when a term means different things across systems. It reads Annex 22 closely: for critical GxP decisions only static, deterministic models are permitted (fixed parameters, same input to same output, no post-deployment learning), while adaptive models and generative AI are not permitted for those decisions, and the rest of the annex (documented intended use, representative bias-free training data, independent validation datasets, parallel deployment, continuous performance and drift monitoring) is a data governance specification in all but name. It then maps ALCOA+ to concrete controls (Attributable to ownership and access control, Original/Accurate to column-level lineage, Contemporaneous/Enduring/Available to catalog and version history, Consistent to the business glossary, Legible/Complete to change history and approval workflow) and gives a five-step build: catalog the LIMS/ERP/warehouse/BI systems in scope (40+ connectors plus REST API and MCP), fix the vocabulary in a glossary linked to real columns, generate column-level lineage from SQL and transformations as the data-flow audit trail, govern each model like data (intended use, inputs, outputs, versions, training-data provenance, classification) as Annex 22 requires, and keep the platform in a validated on-premise or own-tenant environment with a separate metadata store, role-based access, and audit-ready change history. It states one honest boundary plainly: a governed data platform is not the validated system of record (it does not run the batch record or the instrument), and no data governance product is 21 CFR Part 11 compliant on its own because compliance is validated for a system in its own environment; what the platform supplies is the traceability, controlled vocabulary, and audit-ready lineage that data integrity depends on. Two dark-palette SVG diagrams (ALCOA+ mapped to governance controls; audit-ready lineage compared with and without governed lineage) and a three-question FAQ with FAQPage schema. Fully general, no client named, no competitors named. CTA to Interactive Data Lineage, cross-linking Enterprise Deployment and the banking audit-ready lineage post. Author: Samuel Nagy (VP of Strategic Growth, August 2026). - [Agentic Supply Chains Run on Governed Context](https://www.dawiso.com/blog-post/agentic-supply-chain-governed-context) - A point-of-view editorial arguing that AI agents are entering the supply chain faster than the data underneath them is ready, and that they act reliably only on governed context, the agreed layer of meaning over existing systems, rather than on a single centralized warehouse. It opens on the adoption curve: Gartner expects 40% of enterprise applications to include task-specific agents by the end of 2026 (up from under 5% a year earlier) and predicts half of cross-functional supply chain management solutions will use intelligent agents to execute decisions autonomously by 2030, with more than half of supply chain executives already deploying agents; Gartner's 2026 supply chain technology trends put trust and governance (product provenance, decision governance) alongside autonomy and specialization. It then names the three failure modes an agent hits on raw multi-plant data: inconsistent definitions (one plant counts micro-stops as downtime, another does not, so OEE, scrap, and downtime diverge and a blended answer is confidently wrong), unclear ownership (no one is accountable for a returned figure), and missing traceability (you cannot safely allow an autonomous action on a number you cannot trace), noting the manufacturer's existing instinct for the digital thread as the same discipline applied to operational data. It defines governed context as agreed definitions, a named owner per term, lineage from source to report, and classification of what is sensitive, and leans on Deloitte's guidance that perfect data harmonization is not the prerequisite (consistent definitions, clear ownership, and sufficient semantic structure are), so the move is to govern meaning once and leave data where it runs rather than rip five ERPs into one store. The delivery model brings context to the agent over the Model Context Protocol so the agent runs natively where the data and compute already live (cross-linking the companion piece on where to build agents). It closes with four practical moves mapped to capabilities: catalog every system (40+ connectors including a purpose-built SAP S/4HANA scanner, plus Snowflake, Databricks, Power BI, dbt, Oracle, SQL Server, Tableau), agree KPIs once in a business glossary linked to the real columns, generate column-level lineage from SQL and ETL, and serve it to any MCP-compatible agent through an open MCP Server, grounded in the Stora Enso unified governed catalog across Snowflake, dbt, and Power BI. Two dark-palette SVG diagrams (agent over raw conflicting systems versus agent over one governed definition; how governed context reaches agents over MCP across ERP, production, warehouse, and BI) and a three-question FAQ with FAQPage schema. No competitors named. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, August 2026). - [Where Should You Build the AI Agents That Talk to Your Data?](https://www.dawiso.com/blog-post/build-agents-where-your-data-lives) - A point-of-view editorial answering a question that has been live for years and only grown louder: where should organizations build the agents that talk to their data? It first broadens what counts as such an agent - not only an analytics copilot, but a customer-success agent looking up an account or an e-commerce chatbot answering a shopper's order question, any agent with a backend it can query - so almost every organization is building at least one. It then lays out a framework of three groups racing to be where you build them: (1) hyperscalers (Microsoft, Amazon, Google) that already hold most of your compute, often your data, and your security and identity controls, and make building agents native and easy; (2) data platforms (Snowflake, Databricks) that hold the data and compute and want the query never to leave; (3) a set of third-party tools that hold neither the data nor the compute and still believe they should host your agents. The market direction is that the big platforms will keep out-building and undercutting a third layer, because they sit on the two things an agent needs most, so running the agent on a platform that owns neither the data nor the compute adds a network hop, a second security boundary, and another bill for little gain. The one real case for a separate build layer is an agent that must reason across several sources at once where no single platform holds everything, but that case is rare and even then usually the wrong fix, because the problem is inconsistent meaning across sources, which reconciling the meaning solves and moving the runtime does not. That reframes what is actually hard to reproduce: not the agent (getting cheaper and commoditized by the big platforms) but the governed context behind the data - the agreed definitions, ownership, lineage, and classification - which is the slow, valuable asset you least want locked to one vendor. So the design goal flips: governed context, not the agent, should be the thing that travels; you govern business meaning once in a layer you own and serve it to whichever agent needs it, wherever it runs. "Build your agents on our platform" is framed as a soft form of vendor lock-in (the same instinct as a governance vendor pushing its own engine and steering you off first-party tools); "use your context anywhere" is the opposite promise. Dawiso's deliberate position: it is not where you build agents, but where your governed business meaning lives, connecting to more than 40 platforms, defining each term once with an owner and approval workflow, tracing lineage and classification across the estate, and serving that context to any MCP-compatible agent through an open MCP Server - so you build natively on a hyperscaler, inside Databricks or Snowflake, or as a custom agent, as customer Seznam.cz did by connecting its own conversational agent to governed context in Dawiso over the Model Context Protocol. Two dark-palette SVG diagrams (the three groups compared by what they actually hold - data, compute, security; and "build it on our platform" lock-in versus a portable governed context layer serving any agent over MCP) and a four-question FAQ with FAQPage schema. No competitors named. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, August 2026). - [Why AI Can't Answer From Your SharePoint (and What Fixes It)](https://www.dawiso.com/blog-post/why-pointing-ai-at-sharepoint-fails) - A problem-and-solution editorial for organizations that keep internal standards, policies, and procedures in SharePoint and want to use them as the source for an internal AI assistant. It explains why pointing the model straight at the document library fails: documents written for humans carry inconsistent formatting and structure that retrieval flattens and loses, and a raw library has no governance - no single owner, no record of which version is currently valid, and several near-duplicate copies - so the assistant grounds answers on drafts, duplicates, or expired policies and returns fluent, wrong answers. The core argument is that two apparently separate projects are the same problem: the compliance need to govern and audit how documents are approved and versioned, and the AI need for a reliable, structured source of context. Both require the documents to be structured, owned, versioned, and reduced to a single source of truth. The fix is a governed context layer: structured import out of SharePoint without breaking how each document looks; a governed document lifecycle (draft, review and approval, published, archived) where only the current approved version is searchable so AI can never ground on a draft or a superseded policy; full auditability, so any version can be reconstructed and exported as of any date for a regulator; content enriched with catalog and business-glossary metadata to make it AI-ready; and delivery through AI search with cited sources, segment-aware retrieval, a REST API, and the Model Context Protocol for any MCP-compatible agent. Fully general with no client or company named. Two dark-palette SVG diagrams (direct-to-SharePoint versus governed context layer; the document lifecycle zones). CTA to Unstructured Data Governance for AI, cross-linking the Context Layer and MCP. Author: Samuel Nagy (VP of Strategic Growth). - [Data Sovereignty and the European Data Catalog](https://www.dawiso.com/blog-post/european-data-sovereignty-data-catalog) - A positioning editorial arguing that for European organizations, where a data catalog runs and who can be compelled to open it has become a real decision criterion, with sovereign deployment now a primary driver for moving off US-headquartered platforms, ahead of price or features. Section 1 frames the shift: the catalog holds the map of an organization's entire data estate, so where that map lives and whose law reaches it is not a detail. Section 2 defines what "sovereign" means through four separate tests, ownership (who owns the vendor), operation (who runs the platform and holds the keys), data location (where data and metadata physically live), and legal jurisdiction (whose law the operating company must answer to), and shows a catalog can pass the data-location test while failing the others, because a US-headquartered vendor can store metadata in Frankfurt and remain subject to US jurisdiction. Section 3 covers the legal exposure: the US CLOUD Act (2018) lets US authorities compel a US-based provider to produce data it has possession, custody, or control of regardless of where it is stored, which creates a direct GDPR conflict, and the EU-US Data Privacy Framework that has covered transfers since July 2023 is contested (a 29 June 2026 US Supreme Court ruling weakened the independence of its redress mechanism, noyb signaled a fresh challenge widely called Schrems III, and the two predecessors Safe Harbor and Privacy Shield were both struck down by the Court of Justice of the EU). Section 4 shows sovereignty is now a scored, funded procurement standard, not sentiment: the European Commission's June 2026 Cloud Sovereignty Framework grades providers on Sovereignty Effectiveness Assurance Levels up to SEAL-4 (its full-sovereignty tier) across eight objectives, a 180 million euro sovereign-cloud contract over six years was awarded to four European providers in April 2026 (SEAL-2 entry level), and EuroStack is building a European alternative, with the realistic note that no analyst expects a wholesale exit from US hyperscalers in 2026. Section 5 describes a sovereign catalog: European-owned so no foreign parent can be compelled, operable inside infrastructure you control, data and metadata where you decide, plus deployment flexibility (SaaS, private cloud in your own EU tenant, or on-premise from one product) and a metadata-only, single-tenant design supporting GDPR/ISO 27001/SOC 2. Where Dawiso fits: European-built and European-operated, deployable in private cloud inside your own Azure/Google/AWS tenant or fully on-premise (a managed Azure SaaS option exists but the sovereign path does not depend on it), only metadata transferred in hybrid setups, separate metadata store per customer, and the full governance stack (data catalog, business glossary, interactive lineage, context layer for AI). Collibra is referenced only generally via the Dawiso vs Collibra comparison page, not named as an accusation. Two dark-palette SVG diagrams (four sovereignty tests compared; how a foreign legal order reaches a US-headquartered provider's data anywhere but not a European vendor running in your own environment) and a four-question FAQ with FAQPage schema. CTA to Enterprise Deployment. Author: Samuel Nagy (VP of Strategic Growth, July 2026). - [Does a Standalone Context Layer Make Sense in the Era of Genie Ontology?](https://www.dawiso.com/blog-post/standalone-context-layer-genie-ontology) - A provocative point-of-view editorial sparked by the Databricks Data + AI Summit 2026 (June 16) launch of Genie One (an agentic AI coworker, now GA, that produces real artifacts rather than just answers), Genie Agents, Genie Code, and the announcement that matters most for this argument, Genie Ontology. The post deliberately starts by giving Genie Ontology full credit rather than dismissing it: Databricks describes it as "an automatic context layer" that extracts knowledge from tables, queries, dashboards, pipelines, and connected apps and organizes it into "a living graph of how a company works," learning from Databricks data plus 50+ connected applications and ranking definitions by authority, governed by Unity Catalog (Glossary entering preview, Domains, Metrics) beneath it. It clears up a common confusion: Genie Ontology is broader than a semantic layer (a semantic layer translates technical data into consistent business definitions and metrics, the role of Unity Catalog Metrics, which is one input that feeds the ontology) and narrower than a formal ontology (it is a learned knowledge graph, not a strict logical model). The thesis: even so, a standalone context layer still makes sense, for two structural reasons the launch does not close. Reason 1, auto-generated context is not governed context: a self-learning graph optimizes for what is most used and credible and improves with low upkeep, but governed business meaning needs a named owner, validation and sign-off, versioning, and accountability, which matters for board metrics, regulatory reports, and definitions like "active customer" that drive revenue recognition; "the model inferred it from usage" is not an answer for an auditor. Reason 2, most enterprises run more than Databricks: Genie Ontology learns from Databricks and its connected apps, but a typical estate also runs Snowflake, dbt, BI tools, CRMs, and operational systems, and a term and its lineage span several of them, so the most accurate agent is grounded in all of it; plus agents are not all Databricks-native, so governed meaning must be delivered over the open Model Context Protocol (MCP) to any MCP-compatible agent rather than locked to one vendor's agents, or each platform becomes a "context island," the AI-era version of data silos. Practical setup: Dawiso as the cross-platform context layer with Databricks as a first-class source, not a rival - a business glossary that defines each term once with an owner and approval workflow, cross-platform lineage and classification complementing Unity Catalog, and governed context served to any agent via the MCP Server. The verdict: yes, build context in Databricks with Genie Ontology, then make sure that context is owned, governed across every platform, and reachable by every agent. Two light-palette SVG diagrams (auto-generated vs governed context; one governed context layer spanning Databricks/Snowflake/dbt/BI/CRM and serving any agent over MCP) plus two product screenshots (governed business glossary term and cross-platform lineage) and a four-question FAQ with FAQPage schema. Does not cite competitors. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [NIST AI RMF for Models and AI Agents](https://www.dawiso.com/blog-post/nist-ai-rmf-for-models-and-agents) - A POV/how-to companion to the NIST AI RMF glossary article, modelled on the "for models and agents, with implementation steps" format. Thesis: the NIST AI Risk Management Framework (AI RMF 1.0, January 2023) is voluntary and technology-neutral, so its four functions (Govern, Map, Measure, Manage) still fit in the agent era, but the surface they must cover grows. Anchored to fresh 2026 facts: NIST launched a dedicated AI Agent Standards Initiative through its Center for AI Standards and Innovation on 17 February 2026 (three pillars: industry-led standards, open-source protocols, security research), with SP 800-53 control overlays for single- and multi-agent systems in development and a 2026 NCCoE concept paper on authenticating/authorizing autonomous agents; plus the Generative AI Profile (NIST-AI-600-1, July 2024). Sections: why the RMF matters more now (the shift from what a model says to what an agent does); the four functions in one minute (links to the glossary for fundamentals); models vs agents (a model is bounded input-to-output; an agent plans, calls tools, acts in live systems, and keeps memory, widening the risk surface with unauthorized actions, data exfiltration, chained errors, tool manipulation); implementation steps for each function with the agent-specific additions called out (Govern: AI system inventory, policy, owners; Map: inventory tools and data an agent can reach; Measure: monitor actions and traces, not just outputs; Manage: scoped access, guardrails, stop control, incident path); and the hard part is context (every function assumes governed data the framework does not provide). Where Dawiso fits: a governed context layer (catalog, business glossary, lineage, classification, ownership) served to any MCP-compatible tool via the Context Layer and MCP Server, so every model and agent reads the same trusted context. Two light-palette SVG diagrams (models vs agents risk surface; implementation steps across the four functions on a governed-context foundation) and a four-question FAQ with FAQPage schema. Cross-links the NIST AI RMF glossary article, ai-governance, eu-ai-act, iso-42001, and multi-agent-systems glossary terms, plus the MCP product and Context Layer pages. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Building DORA Compliance for Banking AI Agents in 2026](https://www.dawiso.com/blog-post/ai-agents-in-banking-dora-compliance-2026) - A vertical, agent-focused companion to Dawiso's general DORA data-governance post, modelled on the "AI Agents in [industry]" format. Thesis: banks already run AI agents for credit decisioning, fraud detection, trading, and customer service, and under the Digital Operational Resilience Act (Regulation EU 2022/2554, effective 17 January 2025) those agents are ICT systems within the meaning of DORA, so the Chapter II ICT risk framework already governs them, no special AI regime required. Anchored to fresh 2026 facts: Germany's BaFin issued non-binding guidance in January 2026 confirming AI systems (including generative AI and LLMs) must be embedded into existing DORA ICT governance, testing, and third-party frameworks; 2026 is the enforcement phase where supervisors expect data-driven evidence of resilience over policy documents. Sections: (1) AI agents are already ICT systems under DORA; (2) where agents raise the stakes - an agent acts, so a misread definition becomes a wrong action (loan declined, payment frozen, trade placed) at machine speed, and incident reporting then demands you reconstruct which data the agent used; (3) the third-party problem agents make worse - most agents run on foundation models and cloud, and the first Critical ICT Third-Party Providers list (18 November 2025) named 19 providers including AWS, Microsoft, Google Cloud, and IBM, all of which must land in the DORA Register of Information (the single hardest DORA requirement per Deloitte research); (4) the four governed foundations a DORA-ready agent needs - a catalog inventory of the agent and what it touches, end-to-end lineage for the audit trail, classification and access metadata to keep agents from data they should not use, and documented ownership/accountability; (5) DORA meets the EU AI Act - credit scoring and fraud detection are high-risk under the AI Act, the evidence (lineage, classification, inventory, ownership) overlaps, so one governed foundation serves both. Where Dawiso fits: a governed data foundation (Data Catalog, Business Glossary, classification, Interactive Lineage across 40+ platforms) served to any MCP-compatible agent via the MCP Server, turning compliance from a document into evidence. Two dark-palette SVG diagrams (an AI agent as an ICT system inside DORA's Chapter II framework with the five pillars; a governed data foundation served to an agent over MCP so every decision traces to a governed, classified source) and a five-question FAQ with FAQPage schema. Cross-links the existing DORA data-governance post, the 9-financial-data-compliance-challenges post, the BCBS 239/GDPR lineage post, and the AI-Act-readiness guide. CTA to the Dawiso AI Governance solution. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [MCP vs A2A Protocol: What's the Difference?](https://www.dawiso.com/blog-post/mcp-vs-a2a-protocol) - A comparison editorial that reframes the "MCP vs A2A" question people keep asking as a category error: the two protocols are not rivals on one line but axes on two. The Model Context Protocol (MCP, introduced by Anthropic in November 2024) connects a single agent vertically downward to the tools, data, and prompts it needs, over a JSON-RPC client/server model - the "USB-C for AI," with thousands of servers by 2026. The Agent2Agent protocol (A2A, announced by Google in April 2025 and donated to the Linux Foundation in June 2025, now backed by 150+ organizations) connects agents horizontally to each other: each agent publishes a machine-readable capability card, hands tasks to peers over HTTP, and follows them through a lifecycle without exposing its internals. MCP omits agent-to-agent coordination; A2A omits access to tools and data - so the two compose rather than compete, in the same multi-protocol coexistence as HTTP, WebSocket, and gRPC. Includes a worked example (an ops agent uses MCP to pull warehouse data and read documents, then A2A to delegate a pricing question to a finance agent that uses its own MCP connections) and a six-dimension comparison table (origin, what it connects, direction, transport, core unit, what it does not cover). The core argument for Dawiso: both protocols are transport - they move context and tasks but neither defines what the context means, whether it is current, who may see it, or whether it can be trusted ("protocol compliance is not context accuracy"). An MCP server can faithfully deliver a `revenue` table without stating which definition it is or whether a row should be masked; an A2A handshake can pass a well-formed task between agents operating on conflicting definitions of "active customer." The gap matters more with A2A because agents act, so a misread definition becomes a wrong action at machine speed. Where Dawiso fits: the governed context layer beneath both protocols - catalog, business glossary, classification, interactive lineage across 40+ platforms - served to any MCP-compatible agent via the Context Layer and MCP Server, so the context A2A coordinates around is already governed at the source. Two dark-palette SVG diagrams (the vertical-vs-horizontal axes; the six-dimension comparison with a shared governance-gap banner) and a callout that both protocols move context but neither governs it. CTA to the Dawiso MCP Server. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Data Catalog Tools for 2026](https://www.dawiso.com/blog-post/best-data-catalog-tools-2026) - A market map of the 2026 data catalog space, structured like a Gartner-style report rather than a numbered listicle: a market-dynamics chapter, then four categories, then the tools within each. Market dynamics: (1) the acquisition wave that began in 2024, when HCLSoftware acquired Zeenea, and accelerated through 2025 (Snowflake acquired Select Star in November 2025, Atlassian acquired Secoda in December 2025, ServiceNow acquired data.world, Salesforce acquired Informatica) is now, in 2026, starting to bite: the near-term impact was muted, but as founding teams are absorbed into their buyers, roadmaps slow and support attention shifts at the acquired catalogs, so who owns a catalog and where its roadmap is headed is now part of the decision; (2) the AI and context-layer turn reshuffled the whole market, and vendors handled it so differently that the map from two years ago no longer holds - Alation dropped the data-catalog label entirely and now sells the "Alation Intelligence Operating System" (an AI-governance platform, effectively a new product), Collibra reads as a longtime leader caught by the platform shift and carrying architecture it cannot easily catch up from (bolt-on acquisitions do not close the gap), while a couple of newer players used the turn to move ahead and became the product leaders. The four categories: market leaders (Dawiso and Atlan - cross-platform, AI-native, fast to deploy; on product nothing else sits with them); legacy players (Collibra, Alation, Ab Initio, Zeenea - broad and capable but behind on the AI turn, older architectures catching up; Zeenea is a Paris metadata/discovery platform acquired by HCLSoftware in 2024); niche solutions (Microsoft Purview, Snowflake Horizon Catalog, Databricks Unity Catalog - platform-native, strong inside their platform but single-vendor lock-in; and Entropy Data, formerly Data Mesh Manager, a data-product marketplace and data-contract tool that sits on top of a catalog rather than being one); open-source (DataHub and OpenMetadata - self-serve go-to-market built for smaller teams, free tier stripped back, managed cloud not cheap, limited for large enterprise). Dawiso is placed as a market leader: a data catalog whose same catalog is the context layer for AI (Data Catalog, Business Glossary, Interactive Data Lineage, classification, automated upkeep, served over MCP), differentiated not by having context - that is table stakes now - but by reach, cost, and independence: transparent per-user pricing versus the five- and six-figure entry prices of the enterprise-grade options, 40+ platform connectors, first use cases in weeks, and no ownership by any warehouse or suite, so it fits teams from smaller companies up to enterprise and the governance you build stays yours. The "How to Choose" section covers where your data lives and the lock-in trade-off of platform-native catalogs, how agents read metadata safely over MCP, and the ownership question. Three dark-palette SVG diagrams (a consolidation diagram of catalogs being absorbed into platforms versus an independent catalog you own; a four-quadrant market map with Dawiso and Atlan as leaders; a total-cost-of-ownership comparison across the wider market with Dawiso the lowest). External sources cited via neutral outlets only (InfoWorld for Select Star, TechTarget for Secoda, PRNewswire for the HCL-Zeenea deal); no direct links to competitor websites. Cross-links the companion "How to Choose the Right Data Catalog for Your Maturity Level" article and Dawiso's competitor comparison pages. CTA to the Dawiso Data Catalog. Author: Samuel Nagy (VP of Strategic Growth, July 2026). - [8 Semantic Layer Tools for BI and AI Agents in 2026](https://www.dawiso.com/blog-post/best-semantic-layer-tools-2026) - A buyer's-guide listicle organizing the 2026 semantic layer market into four categories and listing eight tools, numbered by category rather than ranked, each with key capabilities and (where it matters) limitations woven into the prose. Opens by defining a semantic layer (the layer between raw tables and the people or agents asking questions; it holds the definitions of what "revenue" means and how "active customer" is calculated, so a number means the same thing everywhere) and sets an honest frame: most of these tools now market themselves for AI, enforce access rules, and ship an MCP server, so the difference is not who has "context" but scope. A semantic layer governs the metrics you model in it; a context layer governs the whole estate around them. The four categories: standalone layers you run yourself (1. dbt Semantic Layer - MetricFlow, version-controlled YAML, early OSI backer, best for teams already in dbt; limits: assumes you run dbt, metrics-only; 2. Cube - headless layer over SQL/REST/GraphQL/AI APIs with its own access rules and MCP server, warehouse-agnostic; limits: engineering-heavy, fewer pre-built BI integrations, self-hosting overhead off Cube Cloud; 3. AtScale - universal layer with MDX and aggregate-aware acceleration that markets a "deterministic context" AI story; limits: enterprise pricing, configuration overhead, OLAP-cube heritage fit); warehouse-native layers (5. Snowflake Cortex Analyst + Semantic Views governed by Horizon Catalog, excellent inside Snowflake but platform-bounded, two-way with Dawiso via OSI; 6. Databricks Unity Catalog metric views + AI/BI Genie with row-level security flowing into answers, real governance but Databricks-scoped); BI-native layers (7. Looker/LookML, mature, extended with Gemini, BI-tool lock-in outside Looker; 8. Microsoft Power BI/Fabric semantic models, DAX-based, surfaced to Copilot, most at home in the Microsoft estate). Entry 4 is Dawiso, framed as more than a semantic layer and the layer on top: it does the semantic job itself (Business Glossary defines metrics and terms; it generates governed semantic views through OSI) and goes past it. The differentiation is scope: Dawiso catalogs every asset across 40+ platforms, classifies sensitive data wherever it lives, traces lineage end to end, records ownership, and can sit above the semantic layers it connects to (such as dbt, Snowflake, Databricks, and Power BI) and bring their definitions into one governed view, then serves governed context to any MCP-compatible agent via the Context Layer and MCP Server. The "How to Choose" section gives two decision questions (where your stack lives; how agents read definitions safely) and makes the composability/lock-in argument: keeping the context layer separate and owned (portable via OSI) gives a composable, best-of-breed stack with low switching costs, so if a warehouse's pricing climbs you can swap that one layer without re-governing the business, avoiding vendor lock-in at the hardest layer to rebuild. Deliberately excludes governance-catalog competitors (Collibra, Alation, Atlan, DataHub) because those are a different category, not semantic layers. Two dark-palette SVG diagrams (the four-category landscape with the context layer governing the estate beneath the three metric-defining categories and serving agents over MCP; a composable stack where semantic layers are swappable beneath a context layer you own) plus a five-question FAQ with FAQPage schema. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Why Text-to-SQL Breaks in the Enterprise](https://www.dawiso.com/blog-post/why-text-to-sql-breaks-in-the-enterprise) - A value-forward technical article on why natural-language-to-SQL tools that dazzle in demos produce quietly wrong answers on real enterprise warehouses. The thesis: the model is excellent at SQL syntax and blind to what your data means, and in the enterprise meaning is everything. Documents the production gap with the Spider 2.0 benchmark (2025): leading models that score ~87% on small, clean academic schemas solve only ~10% of realistic enterprise text-to-SQL tasks, where schemas average hundreds (sometimes 3,000+) columns, names are abbreviated and ambiguous, and the business knowledge needed to use them is scattered across ~1.5M tokens of documentation. Breaks the breakage into three failure modes: (1) it doesn't know what your metrics mean - "revenue" and "active customer" are definitions, not columns, so the model guesses; (2) it picks or joins the wrong table - raw vs staging vs deprecated vs certified marts look alike, and wrong joins silently double-count; (3) it can't tell what's authoritative or sensitive - no concept of which source is official or which columns are PII, so it reports unofficial numbers or leaks restricted data. The fix is not a smarter model but to stop making it guess: a governed semantic layer the tool reads before writing SQL - business glossary (meaning), catalog + lineage (right source), classification + policy (authority and sensitivity) - delivered over MCP so you govern meaning once and serve it everywhere. Stays value-forward on warehouse-native tools: Snowflake Cortex Analyst (90%+ real-world BI accuracy, ~2x a single-prompt LLM, credited to its semantic model not the raw model), Databricks Genie (reads semantic metadata in metric views), and BigQuery Gemini are excellent inside their platforms precisely because they consume a semantic model; the limitation is per-platform scope, which Dawiso fills with cross-platform governed context. Two light-palette SVG diagrams (the text-to-SQL pipeline and its three blind spots; ungoverned guessing vs governed context over MCP) plus three embedded product screenshots (glossary term, governed AI search answer, MCP assistant) and a five-question FAQ. CTA to the Dawiso Business Glossary. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [What 'AI-Ready Data' Actually Means: A 7-Point Checklist](https://www.dawiso.com/blog-post/what-ai-ready-data-actually-means) - A practical, checklist-style guide that replaces the vague slogan "AI-ready data" with a concrete, measurable definition: data an AI can find, trust, and use correctly without a human supplying the missing context. Deliberately targets the how-to/checklist intent and links "up" to the AI-ready data glossary entry for the definition itself, avoiding overlap with the context-engineering guide. Breaks AI-readiness into seven checks, grouped as find it / trust it / use it, each with a one-line test for how to measure it: (1) Cataloged - the AI can discover it in a searchable inventory; (2) Defined - its business meaning is documented in a glossary, the highest-leverage check because undefined data is what makes models guess (cross-linked to the text-to-SQL post); (3) Classified - sensitive fields are tagged with policy so AI can't leak regulated data; (4) Lineage-traced - any answer can be traced back to an authoritative source; (5) Quality-scored - measured freshness/completeness/accuracy (referencing ISO/IEC 25012) and a certified-or-deprecated status; (6) Owned - a named owner and steward keep the other checks true as the business changes; (7) Accessible over MCP - governed context is exposed through an open protocol any assistant or agent can read, or you rebuild silos one tool at a time. Closes with how to score readiness (rate each check in place / partial / missing; AI-ready = all seven in place; run per domain, get high-value datasets to all-seven before widening). Two light-palette SVG diagrams (the seven checks grouped find/trust/use; a per-dataset readiness scorecard) plus two embedded product screenshots (glossary term, interactive lineage) and a five-question FAQ. Where Dawiso fits: catalog, glossary, classification, lineage, quality/certification, ownership, and an MCP Server assembled in one governed layer across 40+ platforms, served via the Context Layer. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [What Dawiso Adds to Databricks Unity Catalog](https://www.dawiso.com/blog-post/dawiso-databricks-unity-catalog) - Practical article on the Dawiso and Databricks integration, written for teams already running Unity Catalog. (1) The Access Gate handles approval of access to data products: a person or an AI agent requests a data product in the catalog, and Dawiso runs the process from approval through to automatically granting the access in the target platform, so the grant lands in Databricks. Unity Catalog enforces grants but does not record who asked, who approved, for what purpose, or when access expires; the Access Gate produces that record as GDPR, DORA and NIS2 evidence. (2) Scanner scope: a Databricks account has a Unity Catalog metastore axis and a workspace axis, and most third-party catalogs read only the first. Dawiso ingests catalogs, schemas, tables, views, ML models, constraints, Unity Catalog tags, column masks, row filters, data classification, external locations and storage credentials, plus AI/BI dashboards with their datasets and Genie Agents (formerly Genie Spaces) with their sample queries. Lineage comes from system.access events with a configurable window and workspace or metastore scope; the ingestion engine loads millions of objects and scanners run from the Dawiso cloud or as command line utilities inside the customer network. (3) Unity writeback: ownership, GDPR and security classification are mastered in Dawiso and written back onto the Databricks column as a Unity Catalog tag or property under a chosen key, so business owners curate in Dawiso while data engineers read the result in Databricks. (4) Data quality: Databricks Data Quality Monitoring (formerly Lakehouse Monitoring, anomaly detection in public preview February 2026) runs the checks on Databricks compute while Dawiso tracks results at the metadata level and surfaces business-readable dashboards, removing the need for a third data quality tool. (5) MCP: Dawiso was listed in the Databricks MCP Marketplace in May 2026 as one of the first independent tools outside the launch partners, and the connection both reads governed context to Genie Agents and lets agents write documentation back into Dawiso scoped to the running user. Dawiso is a validated Databricks technology partner. Author: Samuel Nagy (VP of Strategic Growth, September 2026). - [Govern Snowflake Cortex Agents and Semantic Views](https://www.dawiso.com/blog-post/snowflake-cortex-agents-semantic-views-dawiso) - Practical article on the two-way connection between Dawiso and Snowflake Cortex. Explains how a Cortex agent answers questions by orchestrating tools (Cortex Analyst over structured data, Cortex Search over documents), and how Cortex Analyst reads a semantic view rather than raw tables, so the chain is agent -> semantic view -> tables. Part 1: Dawiso now scans both Cortex agents and semantic views directly from Snowflake and builds interactive lineage automatically, showing exactly which agent reads which data down to the table (illustrated with a scanned HR_BUSINESS_PARTNER agent, its scanned description and response instructions, and the two semantic views feeding it). Covers the scale problem of departmental "talk to your data" agents (sales, HR, customer care), each reading its own subset of semantic views. Part 2: because a Dawiso data product carries business definitions, glossary terms, and data mapping, Dawiso generates a new governed semantic view back into Snowflake via the Open Semantic Interchange (OSI) - the vendor-neutral semantic-model standard Snowflake announced in September 2025 with Salesforce, dbt Labs, BlackRock, and RelationalAI. The result is visibility on the way in and AI-ready context on the way out, managed from the Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Context Silos: Data Silos Reborn for the AI Era](https://www.dawiso.com/blog-post/context-silos-ai-era) - Thought-leadership argument that the data silos enterprises spent a decade dismantling are returning one layer up, where business context lives. Defines a "context silo" as an isolated store of business context (definitions, metrics, relationships, access rules, trust signals) that one AI tool builds and uses, invisible to every other tool - the data is centralized, but the meaning fragments. Anchors the term to the emerging industry discourse ("from data silos to context silos"). Shows how warehouse-native AI builds them fastest: Snowflake Cortex Analyst from a YAML semantic model and semantic views (governed by Horizon), Databricks AI/BI Genie from Unity Catalog, Gemini in BigQuery - each governed only within its own platform, so the same metric ("active user") ends up defined differently in each, with no system reconciling them. The agent era (OpenAI Frontier, launched Feb 2026, plus Microsoft and Salesforce copilots) multiplies the silos and adds stakes, because agents act rather than answer. Lays out the four costs of fragmented context (inconsistent answers erode trust, governance gaps under the EU AI Act and GDPR, duplicated effort, and context lock-in - the most expensive because it compounds), then the fix: govern context once (catalog, business glossary, lineage, classification) in infrastructure you own and serve it to every tool through MCP, so platforms become consumers of your context rather than its custodian. Two dark-palette SVG diagrams (three fragmented silos defining one metric three ways; one governed context layer served via MCP to Cortex, Genie, Copilot, Frontier, and any MCP agent). Where Dawiso fits: 40+ platform connectors, one governed foundation, served via the Context Layer and MCP Server. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Getting OpenAI Frontier-Ready: A Governance Checklist](https://www.dawiso.com/blog-post/openai-frontier-ready-data-governance-checklist) - Practical checklist riding the February 2026 launch of OpenAI Frontier, OpenAI's open enterprise platform for building, deploying, and managing AI agents that act across an organization's data and systems. Frames the core distinction: Frontier provides access governance (per-agent identity, permissions, guardrails, monitoring - who an agent is and what it can do) but assumes context governance (what data means, whether it can be trusted, how it is defined, where it came from) already exists. So "are we ready for Frontier?" is, underneath, "is our data governed and AI-ready?" Walks six platform-agnostic steps, each mapped to a Dawiso capability: (1) inventory the data and systems agents will reach via a data catalog; (2) define what data means with a business glossary and semantic layer so agents compute metrics your way and no new context silo forms; (3) establish trust with interactive lineage and quality signals so agents and auditors can see where an answer came from; (4) classify sensitive data (PII, regulated, confidential) and attach policy so Frontier's guardrails enforce real boundaries; (5) decide who owns the context - keep glossary, lineage, and rules in infrastructure you control rather than inside the agent platform, to avoid context lock-in (enforcing governance and owning it are not the same thing); (6) serve governed context to any MCP-compatible agent over the Model Context Protocol via the Context Layer and MCP Server. Two dark-palette SVG diagrams (access vs context governance; the six-step checklist) plus a five-question FAQ (Frontier availability and early customers HP/Oracle/State Farm/Uber, what Frontier governs on its own, platform-agnostic readiness, the biggest risk of a permissioned agent on ungoverned data, and how Dawiso helps). Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [Context Engineering for Enterprise AI: The Missing Layer Behind Trustworthy AI Answers](https://www.dawiso.com/blog-post/context-engineering-enterprise-ai) - A guide arguing that the bottleneck in enterprise AI is not model capability but context: what the model can see about your business before it answers. Opens on the MIT finding that 95% of GenAI pilots produced no measurable return (cross-linked to Dawiso's "why 95% of GenAI pilots fail"). Defines context engineering as the practice of deciding what a model sees before it answers and making that information accurate, relevant, and governed - spanning system instructions, retrieved knowledge, business meaning, tool definitions, memory, and governance signals. Distinguishes it from prompt engineering (optimizes a single prompt, owned by an individual; context engineering optimizes the whole pipeline, owned by the data team) and from RAG (a retrieval technique; context engineering is the discipline that governs what is retrieved and supplies the meaning RAG cannot infer). Presents the five layers of enterprise context (system, retrieval, semantic, governance, provenance) and argues the bottom three governed layers are the ones most teams skip and a model cannot reconstruct. Explains why a bigger context window backfires (context rot, citing Chroma's 2025 study and the "lost in the middle" effect), how stakes rise when agents act instead of answer, where MCP fits as the delivery layer that turns a governed context layer into something every tool can consume, and why context quality is downstream of governance quality. Where Dawiso fits: 40+ platform connectors, one governed foundation (catalog, glossary, classification, lineage), served via the Context Layer and MCP Server. Two dark-palette SVG diagrams (prompt vs context engineering; the enterprise context stack) plus a five-question FAQ. CTA to the Dawiso MCP Server. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [How to Implement an Enterprise Context Layer for AI: A Step-by-Step Guide](https://www.dawiso.com/blog-post/how-to-implement-enterprise-context-layer-for-ai) - A practical, step-by-step guide to building an enterprise context layer, the governed layer that gives AI the business meaning, relationships, trust signals, and access rules it needs to use data correctly. Clarifies the semantic-layer vs context-layer distinction (a semantic layer is one floor; the context layer adds lineage, governance, and provenance, then delivers it). Walks six steps, each mapped to a capability: (1) build the foundation with a data catalog of what exists; (2) add meaning with a business glossary that defines metrics and entities once and links them to catalog assets; (3) establish trust with interactive data lineage and quality signals so answers can be traced; (4) attach classification and access policy so PII and sensitive data are governed at the source under GDPR and the EU AI Act; (5) deliver context over MCP so any compatible assistant or agent reads the same governed definitions; (6) keep it alive with lifecycle - approval workflows, versioning, drift detection (including schema drift), and a feedback loop. Gives realistic timelines (catalog + glossary within a day, full context layer in one to two weeks on a platform vs multi-quarter from scratch) and the three platform capabilities that make it practical (unified metadata, automated enrichment, native governance and delivery). Where Dawiso fits: 40+ connectors, one governed foundation served via the Context Layer and MCP Server. Two dark-palette SVG diagrams (six-step build flow; the lifecycle loop) plus a five-question FAQ. CTA to the Dawiso Context Layer. Author: Samuel Nagy (VP of Strategic Growth, June 2026). - [10 Questions Every Executive Should Ask Before Scaling AI](https://www.dawiso.com/blog-post/10-questions-every-executive-should-ask-before-scaling-ai) - Framework-style article for executive teams scaling AI. Walks through ten honest questions covering AI ambition (real strategic intent vs. theater), ROI (when and how to measure), foundations vs. tools (build the substrate before buying point solutions), governance (as accelerator, not brake), context layer (the missing semantic infrastructure for trustworthy AI), ownership (who is accountable for outcomes), people and change management, intent (why this initiative, what problem solved), autonomy (how far AI agents are allowed to act), and workforce impact (what happens to the team on the other side). Each question pairs a Key Consideration with Sample Good and Bad Responses. Includes an inline SVG "Foundations of Scalable AI Value" infographic (Trust / Governance / People). Author: Petr Mikeška (CEO, May 2026). ~2,270 words. - [The Build Threshold Just Moved: Helpdesk in 7 Days](https://www.dawiso.com/blog-post/the-build-threshold-just-moved) - CEO thought-leadership on what shifts the line between "build it" and "buy it" once agentic development can ship customer-facing systems. Covers Dawiso's customer help desk: first functional version in 7 days (portal + agent UI + ticket CRUD + bidirectional email via Microsoft Graph + Slack via Events API + KB infrastructure), two months of evolution into a pilot system with a RAG-grounded ChatWidget answering with native KB citations, the KB seeded from 577 help.dawiso.com articles (1,107 chunks, 620 images via vision models) and continuously extended by Slack Q/A monitoring with PII anonymization, intent-classified ticket offers, manual quality gate, and `@helpdesk remember` agent shortcuts. Argues the speed isn't the point - the point is what shipping that in a week does to the build/buy boundary. Lays out the three things that made it work (preparation upfront, test discipline, session cadence), the compounding KB loop (five sources feeding one knowledge base), and implications for peer founders and enterprise buyers (vendor moat narrowing to integration, governance, compliance, scale, security, operational trust). 564 commits, 701 tests across 54 test files, 16 PostgreSQL tables, 80+ React components, two customers in pilot, rolling out through May 2026. Author: Petr Mikeska (CEO, May 2026). - [4 Weeks to Migrate. 2 Weeks to Replace How We Ship.](https://www.dawiso.com/blog-post/web-acceleration) - CEO thought-leadership on what changed at Dawiso after rebuilding dawiso.com in code. Covers the moment marketing stopped asking for a CMS and started shipping through feature branches and preview deploys, the metrics from the four weeks after go-live (256 commits, 88 merge requests merged, 76 features shipped), how A/B testing went from a year-long roadmap item to a one-week project, why named "skills" replaced what a CMS would expose, the Pavel test (a new hire walking into the new infrastructure), and why agentic marketing operations require governed context (definitions, ownership, lineage). Frames the rebuild as an infrastructure shift, not a website change. Closes with the loop: better context → better agentic work → more shipping → more context to govern. Author: Petr Mikeska (CEO, April 2026). - [How We Went All-In on AI Agents at Dawiso](https://www.dawiso.com/blog-post/how-we-went-all-in-on-ai-agents-at-dawiso) - CEO retrospective on Dawiso's eight-month AI adoption journey from "we all use ChatGPT" to "AI agents are part of every team." Covers the AHA (AI Hyper Accelerator) program, two hackathons in Prague and Brno, the projects that emerged (competitive intelligence platform, bug detection toolkit, AI-generated user profile pages, AI video production, automated Jira ticket bot from Slack threads), and the four cross-cutting patterns that determined success: specification quality, templates and standards, session discipline, and transferable mistakes. Argues that the new bottleneck is specification, not headcount, and that organizations that build the muscle to absorb new AI capabilities as fast as they appear will outpace those that pick a single vendor. Author: Petr Mikeska (CEO, March 2026). - [Agentic AI and Metadata Governance: Why Your AI Agents Need Business Context](https://www.dawiso.com/blog-post/agentic-ai-metadata-governance-enterprise-guide) - Enterprise AI agents fail when they lack business context. Explains why metadata governance is the critical foundation for reliable agentic AI - covers data lineage as the trust layer, business glossaries as the semantic layer, and MCP as the delivery mechanism. Includes a practical enterprise guide for building agentic AI infrastructure with governed context. - [Agent Systems and Natural Language Query: Transforming AI and Data](https://www.dawiso.com/blog-post/agent-systems-and-natural-language-query) - Two forces reshaping AI and data management: agent systems (multi-step autonomous AI) and natural language queries (asking data questions in plain language). Explains what both mean for enterprise data strategy and how they depend on governed, well-described data assets. - [Why 95% of GenAI Pilots Fail: The Hidden Data Crisis Behind Enterprise AI](https://www.dawiso.com/blog-post/why-95-percent-of-genai-pilots-fail) - 95% of GenAI pilots fail, and data is usually the reason. Covers the hidden data crisis behind enterprise AI struggles: missing context, untrusted data, no lineage. Explains how data governance - catalogs, glossaries, lineage - helps organizations beat the odds. - [Contextual Metadata Layer for AI: Business Context for Reliable AI](https://www.dawiso.com/blog-post/contextual-metadata-layer-for-ai-why-business-context-is-the-key-to-reliable-ai) - Business context is the key to reliable AI. Explains what a contextual metadata layer is, how it grounds AI in real enterprise knowledge (definitions, ownership, lineage), and why AI without business context generates plausible but untrustworthy answers. - [The New AI Success Formula: Business Context Fueled By Automation](https://www.dawiso.com/blog-post/the-new-ai-success-formula-business-context-fueled-by-automation) - Business context fueled by automation is the new AI success formula. Covers how metadata plus AI context drives more accurate and trustworthy enterprise results - and why automating context capture is the prerequisite for scalable AI. - [Data Garbage In, AI Garbage Out: Why Governance Matters More Than Ever](https://www.dawiso.com/blog-post/data-garbage-in-ai-garbage-out-why-governance-matters-more-than-ever) - Only 36% of business leaders trust their company's data accuracy (Salesforce 2024). Explains how poor data quality flows into AI outputs and why data governance - ownership, definitions, quality checks - is the first line of defense against AI failures. - [How to Use AI in the Data Landscape: What Actually Works in 2025](https://www.dawiso.com/blog-post/how-to-use-ai-in-the-data-landscape) - Practical strategies for data teams to leverage AI tools and deliver measurable business value. Covers what actually works in 2025: AI-assisted documentation, automated lineage, intelligent search, and natural language querying of governed data. - [AI Governance Blind Spot: The Hidden Risks of Model Reuse](https://www.dawiso.com/blog-post/ai-governance-blind-spot-the-hidden-risks-of-model-reuse) - Reusing AI models introduces hidden governance risks. Covers blind spots in model provenance, training data quality, compliance drift, and how to spot these risks before they become regulatory or operational liabilities. - [10 AI Risks in the Enterprise: Key Dangers and How to Manage Them](https://www.dawiso.com/blog-post/10-ai-risks-in-the-enterprise-key-dangers-and-how-to-manage-them) - Ten AI risks threatening enterprise operations - from bias and hallucination to security breaches and compliance failures. Practical guidance on how data governance helps identify, monitor, and manage each risk category. - [10 Essential AI Governance Questions for Board Members](https://www.dawiso.com/blog-post/10-essential-ai-governance-questions-for-board-members) - Ten oversight questions a board can put to management about AI, each with a sample good and bad answer. Grounded in the EU AI Act as it stands after the 2026 Omnibus: the compliance calendar through August 2028, the AI literacy obligation on deployers (Article 4), deployer duties and named human oversight (Article 26), the fundamental rights impact assessment required of credit scoring and life and health insurance pricing (Article 27), data governance for training data (Article 10), technical documentation and logging (Articles 11 and 12), the individual right to explanation (Article 86), and the penalty tiers (Article 99). Covers AI use case inventory including bought systems, risk-tier classification, bias detection, data lineage, accountability by name, audit readiness, and suspension paths. Written for board oversight, as distinct from the executive scaling questions. - [How RAG and Semantic Layers Work Together in Enterprise AI](https://www.dawiso.com/blog-post/how-rag-and-semantic-layers-work-together-in-enterprise-ai-with-dawiso-context-layer) - RAG and semantic layers explained. Shows how they complement each other in enterprise AI - RAG retrieves relevant chunks, the semantic layer provides business meaning - and how Dawiso's Context Layer serves as the governed semantic foundation for both. - [RAG vs Semantic Layer: What's the Difference?](https://www.dawiso.com/blog-post/rag-vs-semantic-layer-whats-the-difference) - RAG vs semantic layer explained. Covers the key architectural differences, when to use each, and why enterprise AI systems often need both to deliver accurate, business-grounded answers at scale. - [The Complete Guide to MCP (Model Context Protocol)](https://www.dawiso.com/blog-post/what-is-mcp-model-context-protocol) - Model Context Protocol (MCP) connects AI agents directly to data sources. Covers how MCP works, why it matters for enterprise AI, and how Dawiso implements MCP to give AI agents access to governed business context - definitions, lineage, ownership. - [What Is a Semantic Layer and How Does It Help AI Understand Data](https://www.dawiso.com/blog-post/what-is-a-semantic-layer-and-how-does-it-help-ai-understand-data) - A semantic layer translates raw database tables into business terms AI can understand. Covers what semantic layers are, how they bridge technical schemas and business meaning, and why they are the critical infrastructure for reliable enterprise AI. - [DORA Compliance and Data Governance: What Every Financial Institution Needs to Know](https://www.dawiso.com/blog-post/dora-compliance-data-governance-financial-institutions) - DORA enforcement is live. Covers how data lineage, business glossaries, and metadata governance help financial institutions meet the four DORA pillars: ICT risk management, incident reporting, digital operational resilience testing, and third-party oversight. Includes mapping of DORA articles to governance capabilities. - [EU AI Act Compliance Deadlines and Timeline](https://www.dawiso.com/blog-post/eu-ai-act-compliance-deadlines-what-businesses-need-to-know-and-how-to-prepare) - The AI Act timeline as it stands after the simplification regulation adopted by the European Parliament on 15 June 2026 and the Council on 29 June 2026 (the Omnibus VII package). In force: 2 February 2025 prohibited practices and AI literacy, 2 August 2025 general-purpose AI obligations and governance, 2 August 2026 transparency obligations and the start of enforcement by the AI Office and national authorities. Still ahead: 2 December 2026 for the new prohibition on AI-generated non-consensual intimate imagery and CSAM and the shortened transparency grace period, 2 August 2027 for national regulatory sandboxes, 2 December 2027 for stand-alone high-risk systems (Annex III, moved from 2 August 2026), and 2 August 2028 for high-risk AI embedded in products (Annex I, moved from 2 August 2027). Sourced to EUR-Lex, the European Parliament, the Council and the European Commission. - [Are You Governing AI Safely? AI Act Classification, GDPR, Copyright, and Your Internal AI Policy](https://www.dawiso.com/blog-post/are-you-governing-ai-safely) - Safe AI governance comes down to four pillars. Pillar 1: classify every AI system by AI Act risk tier (unacceptable, high, limited, minimal); obligations apply to deployers, not only providers. Pillar 2: reconcile AI with the GDPR (lawful basis, purpose limitation, data minimisation, transparency, DPIAs, and invisible PII in prompts and training data). Pillar 3: solve copyright on both sides - input (text-and-data-mining opt-outs, the GPAI training-data summary obligation under Article 53(1)(d) with the AI Office template from 24 July 2025) and output (uncertain ownership in the EU without human authorship, plus infringement and IP-leakage risk). Pillar 4: an internal AI policy people follow, naming approved tools, permitted data, human-review triggers, and ownership. All four pillars rest on a governed metadata foundation: AI system inventory, data lineage, ownership, and classification - delivered by the Dawiso Data Catalog, Interactive Data Lineage, access management, AI governance solution, Context Layer, and MCP support. - [Shadow AI: The Ungoverned AI Tools Your Employees Already Use](https://www.dawiso.com/blog-post/shadow-ai-enterprise-risk) - Shadow AI is the use of AI tools an organisation has not approved or does not know about - the AI equivalent of shadow IT. Industry surveys in 2025-2026 estimate roughly half of employees (more in some studies) use unsanctioned AI tools, while only about a third of organisations have a formal AI governance framework; IBM has reported shadow AI as a factor in about one in five data breaches, adding hundreds of thousands per incident. The post argues shadow AI is a data-governance problem, not just IT: every prompt is a data transfer that can leak sensitive data, create GDPR and AI Act blind spots, and erase lineage. Banning does not work (it pushes usage out of sight); the fix is a discover-classify-enable-monitor loop and governed, sanctioned alternatives. Dawiso provides the visibility via the Data Catalog as a living AI/data inventory, data classification and access management, Interactive Data Lineage, and a governed AI path through the Context Layer and MCP. - [The AI Omnibus: What Changed in the EU AI Act and What It Means for Your AI Strategy](https://www.dawiso.com/blog-post/ai-omnibus-eu-ai-act-changes) - The AI Omnibus, provisionally agreed on 7 May 2026 as part of the EU Digital Omnibus, amends the AI Act without rewriting it. Covers the deferral of high-risk obligations (Annex III stand-alone systems to 2 December 2027, Annex I embedded systems to 2 August 2028), the shift of Article 50(2) watermarking to December 2026 and sandboxes to August 2027, the replacement of the Commission's conditional standards trigger with fixed dates, two new Article 5 prohibitions (non-consensual intimate imagery and CSAM, applying December 2026), and simplification measures (trimmed Annex VIII registration, softened AI-literacy wording, SME and small mid-cap relief, AI Office oversight of GPAI and VLOPs). Explains what stays unchanged (risk classification, GPAI and prohibited-practice obligations) and how a governed, AI-ready metadata foundation (data catalog, lineage, AI governance, Context Layer, MCP) turns the deadline runway into lasting compliance capability. - [How to Be Compliant and Ready for the EU AI Act: A 2026 Guide](https://www.dawiso.com/blog-post/how-to-get-ai-act-ready-2026) - A practical, step-by-step guide to using the AI Omnibus deadline deferral (Annex III high-risk to 2 December 2027, Annex I to 2 August 2028) as runway rather than relief. Recaps what changed in May 2026 and what still applies from 2 August 2026 (deployer transparency, GPAI obligations), then walks five steps to AI Act readiness, each mapped to a Dawiso AI Governance capability: (1) inventory every AI use case via a centralized list of AI systems feeding Article 11 / Annex IV technical documentation; (2) classify risk against Article 5 and the Annexes with structured, repeatable risk assessment; (3) map data provenance and interactive lineage for Article 10 and Article 26 data-governance duties; (4) standardize meaning with a semantic layer and business glossary for Article 14 human oversight; (5) make governance continuous with workflows and human-in-the-loop review. Argues compliance is the floor not the ceiling - the same catalog, lineage, and glossary built for the Act ground trustworthy AI context through the Context Layer and MCP, so the work that makes you compliant by 2027 makes your AI dependable in 2026. - [BCBS 239 in 2025: Why Now Is the Time to Strengthen Compliance](https://www.dawiso.com/blog-post/bcbs-239-in-2025-why-now-is-the-time-to-strengthen-compliance) - Only 2 of 31 global banks met BCBS 239 requirements by 2023. Covers the 11 BCBS 239 principles, the most common failure points (data lineage gaps, manual reporting, siloed definitions), and why a modern data catalog is the fastest path to compliance. - [BCBS 239: 5 Data Catalogs to Support Risk Data Governance](https://www.dawiso.com/blog-post/bcbs-239-data-governance-5-data-catalogs-to-support-risk-data-governance) - Five data catalog capabilities that help financial institutions strengthen risk data governance and BCBS 239 reporting: automated lineage, business glossary, data ownership, quality monitoring, and audit trails. Includes a comparison of catalog approaches. - [What Banks Need to Know in 2025: Why BCBS 239 Matters More](https://www.dawiso.com/blog-post/what-do-banks-need-to-know-in-2025-and-why-bcbs-239-is-more-relevant-than-ever) - Only 2 of 31 global banks met BCBS 239 requirements by 2023. Covers why regulatory pressure is intensifying, what examiners are looking for in 2025, and how banks must strengthen risk data aggregation and reporting capabilities. - [9 Financial Data Compliance Challenges Banks Must Solve in 2026](https://www.dawiso.com/blog-post/9-financial-data-compliance-challenges-banks-must-solve-in-2026) - From BCBS 239 to DORA and the EU AI Act - a practical walkthrough of the nine toughest data compliance challenges facing financial institutions in 2026. Covers cross-border data tracing, regulatory reporting automation, and AI model governance for banks. - [BCBS 239, GDPR, and the Lineage Problem: Audit-Ready Data Governance in 2026](https://www.dawiso.com/blog-post/bcbs-239-gdpr-lineage-audit-ready-data-governance-2026) - Three regulatory pressures (BCBS 239, GDPR Article 30, DORA) viewed as one underlying capability gap. Covers what ECB supervisory review now demands, why PII tagging needs automation to stay current, and how Dawiso's interactive lineage, impact analysis, business glossary, audit trails, and catalog automation answer each pressure with a single governance layer. Includes the KB Group scale reference (30+ financial institutions, 370,000+ tables, 6,500 business terms). - [Data Governance Maturity Model and How to Measure It](https://www.dawiso.com/blog-post/what-is-data-governance-maturity-and-how-to-measure-it) - Most organizations score their governance on artifacts they own rather than on what changes because they own them. Covers the five maturity levels (Unaware, Aware, Defined, Managed, Optimized) with the observable symptom that identifies each, the six dimensions to score separately (ownership, policies, metadata and discoverability, quality, architecture, literacy), a diagnostic question set a team can run itself, and the test added at the top of the scale in 2026, whether governance is machine-readable and applied at query time rather than written down. Cites the EDM Association 2026 Global Data Management Benchmark and Gartner data and analytics predictions. - [Why Robust Data Governance Is Essential for Unstructured Data](https://www.dawiso.com/blog-post/why-robust-data-governance-is-essential-for-unstructured-data) - Unstructured data - documents, images, emails, SOPs - makes up 80% of enterprise data. Covers why unstructured data is hard to govern (no schema, PII everywhere, inconsistent formats), how AI makes it usable, and why governance must extend beyond structured databases. - [How to Improve Data and Business Literacy in your Company](https://www.dawiso.com/blog-post/how-to-improve-data-and-business-literacy-in-your-company) - Better data and business literacy eliminates IT-business misunderstandings. Practical steps to align technical and business teams: shared glossaries, data ownership programs, self-service access with context, and literacy training tied to governance workflows. - [Flexible Deployment for Data Governance: SaaS, On-Prem, and Hybrid](https://www.dawiso.com/blog-post/flexible-deployment-for-data-governance-saas-on-prem-and-hybrid) - SaaS, on-premise, or hybrid - which data governance deployment fits your infrastructure? Covers the trade-offs of each model for security-sensitive industries (banking, public sector, healthcare), and how Dawiso supports all three deployment modes. - [What Are Data Products and Why Does Your Business Need Them?](https://www.dawiso.com/blog-post/what-are-data-products-and-why-does-your-business-need-them) - Data products package data with context, quality, SLAs, and ownership - enabling self-serve access and better decisions without waiting for the data team. Covers the core attributes of a data product, the difference from datasets, and how data products fit into a data mesh architecture. - [Data Product Lifecycle: How to Manage Data Products in Dawiso](https://www.dawiso.com/blog-post/understanding-the-data-product-lifecycle-how-to-manage-data-products-in-dawiso) - Data products need lifecycle management - from creation through governance to retirement. Covers the full data product lifecycle in Dawiso: definition, ownership assignment, quality monitoring, access provisioning, versioning, and deprecation. - [Decentralized Data Architecture and How Data Products Fit In](https://www.dawiso.com/blog-post/what-is-decentralized-data-architecture-and-how-do-data-products-fit-in) - Decentralized data architecture distributes ownership to domain teams instead of central IT. Covers the four data mesh principles, why data products are the unit of exchange, and what federated governance means in practice for data teams and platform owners. - [How to Choose the Right Data Catalog for Your Maturity Level](https://www.dawiso.com/blog-post/data-catalogs-comparison-for-2025-best-tools-for-your-business) - A decision-focused guide (companion to the "10 Data Catalog Tools for 2026" listicle, which it cross-links) on how to choose a data catalog that fits your organization's data governance maturity rather than overbuying a platform built for a stage you are not at. Covers why choosing wisely matters, the drivers pushing catalog adoption, the four factors to weigh (fit for your needs, budget and resources, ease of use and adoption, regulatory support), the gap between the governance most organizations have today and the adaptive governance they need, why many rollouts stall (complexity and lack of urgency, per Gartner), the must-have capabilities of a modern platform, and why starting early with a lightweight foundation beats adopting an advanced tool too late. CTA to the Dawiso platform. - [How Data Catalog Automation Saves Hours of Manual Work](https://www.dawiso.com/blog-post/how-data-catalog-automation-saves-hours-of-manual-work) - 402.74 million terabytes of data are created daily. Data catalog automation eliminates hours of manual scanning, tagging, and lineage documentation - metadata management on autopilot. Covers automated discovery, AI-assisted documentation, and lineage extraction. - [The Role of Data Catalogs in Modern Analytics and AI](https://www.dawiso.com/blog-post/the-role-of-data-catalogs-in-modern-analytics-and-ai) - Data catalogs are the missing piece for trustworthy AI. Covers how modern catalogs enable searchability, business context, governance, and AI-ready data across the analytics stack - and why AI without a catalog produces unreliable results. - [How to Get an Overview of Your Database and Understand Data Flows](https://www.dawiso.com/blog-post/how-to-get-an-overview-of-your-database-and-understand-data-flows) - Getting a clear overview of your database should not take weeks. Strategies to map data flows, eliminate blind spots, understand table relationships, and use a data catalog to make your database knowledge accessible to the whole team. - [Data Lineage Techniques and Extraction Methods Explained](https://www.dawiso.com/blog-post/data-lineage-techniques-what-is-data-lineage-and-what-extraction-methods-do-we-use) - Data lineage extraction comes in many forms: SQL parsing, query log analysis, API metadata capture, code scanning. Covers which technique fits which stack, the trade-offs between static and runtime lineage, and how Dawiso combines methods for comprehensive coverage. - [Data Lineage Guide: Where Does Your Data Come From and Go?](https://www.dawiso.com/blog-post/where-does-your-data-come-from-and-go-a-guide-to-data-lineage-with-tips-to-maximize-value) - How data lineage helps you trace origins, understand transformations, and maximize data value. Covers forward and backward lineage, impact analysis, compliance use cases (GDPR, BCBS 239), and practical tips for getting value from lineage in your organization. - [Power BI Visual-Level Lineage: Trace Data to Every Table or Chart](https://www.dawiso.com/blog-post/power-bi-visual-level-lineage-trace-data-to-every-table-or-chart) - Power BI lineage typically stops at report level. Dawiso takes it further, tracing data flow down to every individual table, chart, and visual in your reports - enabling impact analysis, compliance documentation, and root cause investigation at the visual level. - [Data Ingestion Architecture in Dawiso Explained](https://www.dawiso.com/blog-post/data-ingestion-architecture-in-dawiso-explained) - Dawiso ingests metadata via push and pull modes with private connections. Explains the architecture: how the Dawiso Integration Runtime (DIR) enables secure on-premises scanning, how push ingestion works for custom sources, and how all metadata lands in one unified catalog. - [What Is a Data Warehouse?](https://www.dawiso.com/blog-post/what-is-a-data-warehouse-take-care-of-the-central-repository-for-your-data) - A data warehouse, or DWH, is a central repository that collects data from across an organization, reconciles it, and stores it in a structure built for analysis rather than transactions. Covers W. H. Inmon's canonical definition (subject-oriented, integrated, time-variant, non-volatile) and what each of those four terms means in practice, how a warehouse differs from an application database, a data lake, and a data lakehouse, why warehouses decay into expensive rebuilds through schema drift and lost definitions, and the catalog, glossary, and lineage that prevent it. - [How Does Data Governance Help with Data Marts?](https://www.dawiso.com/blog-post/why-build-a-data-mart-streamline-financial-analysis-with-self-service-and-enhanced-security) - Data marts give finance teams self-service analytics, tighter security, and lineage visibility - advantages a full data warehouse alone cannot deliver. Covers why to build a data mart, how governance ensures mart data stays trusted, and the role of lineage in financial reporting. - [Challenges in Traditional Data Modeling and Dawiso's Solutions](https://www.dawiso.com/blog-post/challenges-in-traditional-data-modeling-and-dawisos-solutions) - Traditional data modeling struggles with scale, agility, and AI readiness. Covers the most common pain points - rigid schemas, documentation gaps, slow iteration - and how Dawiso helps teams overcome them with automated lineage and collaborative metadata management. - [What Is a Knowledge Graph and Why It Beats Relationship Diagrams](https://www.dawiso.com/blog-post/what-is-a-knowledge-graph-and-why-does-it-outperform-relationship-diagrams) - Knowledge graphs map entities and relationships that flat ER diagrams miss. Covers why knowledge graphs outperform traditional relationship diagrams for understanding complex data connections, and how they serve as the structural backbone for enterprise AI and RAG systems. - [What Is Data Reconciliation, and How Does It Work in Practice?](https://www.dawiso.com/blog-post/what-is-data-reconciliation-how-data-governance-helps-achieve-consistency-and-trust-with-a-data-catalog) - Data reconciliation is comparing two sets of data that ought to agree and resolving where they do not. Separates the three jobs that share the name (financial reconciliation between ledgers and statements, pipeline or source-to-target reconciliation after a load or migration, and master data reconciliation across duplicate records), each with who runs it and what "reconciled" means for it. Walks the five steps: define what should match and to what tolerance, extract both sides at a comparable cut-off, match on an agreed key, classify the breaks as timing, structural or definitional, then resolve and record the cause. Explains why timing breaks resolve themselves while definitional breaks recur every cycle until a business glossary settles the definition, and how catalog, glossary and lineage prevent them. - [Data Silos and Shadow IT: Break Them for Data-Driven Success](https://www.dawiso.com/blog-post/what-are-data-silos-and-shadow-it-break-it-down-for-data-driven-success) - Data silos and shadow IT fragment your knowledge base. Proven strategies to break down barriers: data catalogs for discoverability, business glossaries for shared definitions, and governance programs that make centralized data worth using. - [Collibra Alternative: 6 Reasons Dawiso Is Better for Governance](https://www.dawiso.com/blog-post/collibra-alternative) - Looking for a Collibra alternative? Covers six reasons why Dawiso is a better fit for many organizations: faster implementation, lower total cost, more business-friendly UX, stronger lineage, better AI integration, and no vendor lock-in. - [Collibra UX vs Dawiso: The Business-Friendly Alternative](https://www.dawiso.com/blog-post/comparing-collibras-ux-with-the-business-friendly-alternative-dawiso) - Collibra vs Dawiso UX compared in detail. Teams switching from Collibra choose Dawiso for its business-friendly interface, faster user adoption, and everyday simplicity - with no sacrifice in enterprise governance capability. - [What Are the Differences Between OpenMetadata and Dawiso Data Catalog?](https://www.dawiso.com/blog-post/what-are-the-differences-between-openmetadata-and-dawiso-data-catalog) - OpenMetadata vs Dawiso: open-source flexibility or enterprise-grade governance? Compares the key differences in architecture, business user experience, lineage depth, glossary capabilities, support, and total cost of ownership. - [Total Cost of Ownership with Alation Pricing](https://www.dawiso.com/blog-post/total-cost-of-ownership-with-alation-pricing-nxdp2) - Alation pricing can reach $963,450 in Year 1 with add-ons. Shows how Dawiso delivers the same data catalog capabilities - discovery, lineage, glossary, AI features - with transparent per-user pricing and no hidden module costs. - [Why 57% of Manufacturers Can't See Their Supply Chain](https://www.dawiso.com/blog-post/why-57-of-manufacturers-cant-see-their-supply-chain-and-how-to-fix-it) - 57% of manufacturers lack supply chain visibility. Covers how data silos, manual processes, and fragmented systems create production blind spots - and how data governance eliminates fragmentation to restore visibility and reduce costly inefficiencies. - [How Universities Can Manage Data and Meet Cybersecurity Requirements](https://www.dawiso.com/blog-post/how-universities-can-effectively-manage-data-and-meet-new-cybersecurity-requirements) - Universities face new cybersecurity requirements for data management (NIS2, ISO 27001, GDPR). Practical strategies to govern academic data effectively: data catalog for asset visibility, lineage for compliance, and governance programs suited to decentralized academic structures. - [Trends Shaping Data and Analytics for 2025](https://www.dawiso.com/blog-post/trends-shaping-data-and-analytics-for-2025) - Four trends reshaping data and analytics in 2025: AI governance as a board-level priority, AI-driven metadata management, data products as the new unit of data delivery, and the enduring importance of data governance as the foundation for all of the above. - [Key Takeaways from the 2025 Gartner Data & Analytics Summit](https://www.dawiso.com/blog-post/key-takeaways-from-the-2025-gartner-data-analytics-summit) - Key takeaways from the 2025 Gartner Data & Analytics Summit: metadata management, AI trust and transparency, and data catalogs as the governance backbone are shaping enterprise data strategy for the years ahead. Dawiso perspective from attendees. - [Customizing Dawiso with Marketplace Packages and Patch Packages](https://www.dawiso.com/blog-post/adapt-data-governance-to-your-organizations-needs-customizing-dawiso-with-marketplace-packages-and-patch-packages) - Tailor Dawiso to your governance needs using Marketplace packages and patch packages. Covers how to customize workflows, metadata models, and dashboards for your organization's specific governance requirements - without writing code. ### What Is Data Quality? **URL:** https://www.dawiso.com/glossary/data-quality Explains data quality through six core dimensions (accuracy, completeness, consistency, timeliness, uniqueness, validity) and positions it as a pillar of data governance. Covers the industry shift toward platform-native DQ capabilities in Databricks, Microsoft Fabric, and Snowflake. Compares two data catalog approaches: open DQ metadata integration (Dawiso, Alation) versus built-in DQ engines (Collibra, Informatica). Addresses common challenges including quality degradation, upstream root causes, and tool sprawl. - About: https://www.dawiso.com/about-us - Careers: https://www.dawiso.com/careers - Join 30 people building the data governance platform behind Societe Generale, CEZ, and Stora Enso. AI-first toolchain, real ownership, startup pace. Open roles are listed and applied to directly on-site (no external ATS); applications go to recruitment@dawiso.com. - AI Agentic Developer (https://www.dawiso.com/careers/ai-agentic-developer): Build AI-powered, agentic features on top of Dawiso's Context Layer and MCP integration. Work across multi-model approaches (Claude Code, GPT/Codex), design agentic workflows that ship to production, and shape how the team develops with spec-driven, test-driven agentic coding. Prague, hybrid, full-time. - Data Governance Consultant (https://www.dawiso.com/careers/data-governance-consultant): Work on the ground with enterprise customers - banks, insurers, energy companies - guiding Dawiso implementations (data catalog, business glossary, data lineage, data products, unstructured data governance) and helping them navigate AI adoption, AI governance, and EU AI Act compliance. Prague, hybrid, full-time. - Solution Engineer (https://www.dawiso.com/careers/solution-engineer): Own technical presales presentations and product demos across a wide range of industries, tailor demos to what each prospect cares about, handle technical Q&A during evaluations and RFPs, and turn technical credibility into closed deals. Prague, hybrid, full-time. - Customer Success Representative (https://www.dawiso.com/careers/customer-success-representative): Guide new customers through onboarding and adoption, consult on requirements and translate them into feature specs, turn customer feedback into product input, support pre-sales, and contribute to onboarding materials and Dawiso Academy. Prague, hybrid, full-time. - Let us know about you (https://www.dawiso.com/careers/let-us-know-about-you): Open/spontaneous application for people who don't see a matching posting but want to help across engineering, customer-facing work, or go-to-market. ### Data Governance Adoption Guide **URL:** https://www.dawiso.com/guide Navigate your data governance journey with Dawiso's adoption guide. Build from awareness to full optimization with practical strategies at every stage - for CDOs, data stewards, and governance leads who want a clear path from initial interest through pilot, scale-up, and continuous optimization. ### Try Dawiso **URL:** https://www.dawiso.com/try-dawiso Start your Dawiso journey. Try a live demo with preloaded data, begin a free 14-day trial, or book a call to find the best option for your team. Three entry paths designed for different evaluation styles - quick exploration, hands-on testing, or guided buying. ### Resources **URL:** https://www.dawiso.com/resources The signpost for everything Dawiso publishes on data governance, metadata management, and AI readiness. Each section shows the newest few items and links to the hub that holds the full list: blog posts, customer case studies, downloadable e-books and guides, company news, and the conferences and webinars where you can meet the team next. It also points to the Dawiso Academy for structured learning and the Help Center for product documentation. ### Partners **URL:** https://www.dawiso.com/partners Dawiso teams up with the best services and technology companies to deliver the most flexible solution on the market. Includes services partners (governance consultancies, integration specialists) and technology partners (cloud providers, data platforms, BI tools). ### Tech Blog **URL:** https://www.dawiso.com/tech-blog Practical guides, tool deep-dives, and technical data insights - covering Power BI, Snowflake, Databricks, Keboola, dbt, Azure Data Factory, Tableau, and more. Aimed at data engineers, BI developers, and technical practitioners who need hands-on knowledge across the modern data stack. ## Glossary (Additional Terms) ### Active Metadata **URL:** https://www.dawiso.com/glossary/active-metadata Active metadata explained. How active metadata differs from passive metadata, why it powers automation in modern data governance, and how active metadata platforms continuously enrich, recommend, and act on metadata across the data stack. ### A/B Testing **URL:** https://www.dawiso.com/glossary/a-b-testing Complete guide to A/B testing methodology for data-driven optimization and business growth. Covers experiment design, sample size, statistical significance, and how governed data underpins reliable A/B test results. ### Agile Development **URL:** https://www.dawiso.com/glossary/agile-development Complete guide to agile development methodology for DevOps teams and software engineers. Sprints, iterations, retrospectives, and how agile practices apply to data engineering and analytics teams. ### AI-Powered Business Intelligence **URL:** https://www.dawiso.com/glossary/ai-powered-business-intelligence AI-powered BI adds machine learning, NLP, and predictive analytics to traditional dashboards for proactive, conversational insights - replacing reactive reports with intelligent assistants that surface anomalies and explain trends. ### AI-Ready Data **URL:** https://www.dawiso.com/glossary/ai-ready-data AI-ready data is enterprise data that is accurate, well-documented, discoverable, and governed - prepared for reliable use in AI models, LLMs, and agentic workflows. Covers what it takes to get there: catalog, lineage, glossary, and quality controls. ### Analytics Tools **URL:** https://www.dawiso.com/glossary/analytics-tools Complete guide to analytics tools for BI, data visualization, and business intelligence platforms. Covers categories from spreadsheets and BI tools to advanced analytics, ML platforms, and real-time analytics engines. ### Apache Iceberg **URL:** https://www.dawiso.com/glossary/apache-iceberg Apache Iceberg is an open table format for large-scale analytic datasets. It brings ACID transactions, schema evolution, and time travel to data lakehouses. Explains how it compares to Delta Lake and Hudi. ### Artificial Intelligence **URL:** https://www.dawiso.com/glossary/artificial-intelligence Artificial intelligence transforms industries through machine learning and intelligent automation. Covers the AI landscape - supervised/unsupervised learning, deep learning, NLP, computer vision - and what enterprises need to deploy AI responsibly. ### Auto Recovery **URL:** https://www.dawiso.com/glossary/auto-recovery Complete guide to automated system recovery, resilience patterns, and self-healing infrastructure. Covers checkpointing, replication, automatic failover, and how observability triggers recovery in modern data platforms. ### Auto Remediation **URL:** https://www.dawiso.com/glossary/auto-remediation Complete guide to automated issue resolution, intelligent remediation, and self-healing systems. Covers detection, classification, automated playbooks, and the role of metadata in routing remediation actions. ### Azure Databricks vs AWS vs GCP - Cloud Platform Comparison **URL:** https://www.dawiso.com/glossary/azure-databricks-vs-aws-vs-gcp-cloud-platform-comparison Databricks on Azure vs AWS vs GCP: what actually differs in integration, pricing, and feature availability - and how to govern data across clouds with a unified catalog. ### Business Intelligence **URL:** https://www.dawiso.com/glossary/business-intelligence Complete guide to Business Intelligence strategy, implementation, and data-driven decision making. Covers BI architecture, key roles, KPIs, and the shift from reporting to AI-augmented BI. ### Business Intelligence (BI) Debt **URL:** https://www.dawiso.com/glossary/business-intelligence--bi--debt BI debt management strategies for better decision-making and reduced operational costs. Covers the hidden cost of unmaintained dashboards, undocumented metrics, and orphaned reports - and how governance reduces BI debt. ### Business Intelligence Applications **URL:** https://www.dawiso.com/glossary/business-intelligence-applications Complete guide to BI applications and software solutions for data-driven decision making. Covers Power BI, Tableau, Qlik, Looker, and modern semantic-layer platforms. ### Business Intelligence Dashboards **URL:** https://www.dawiso.com/glossary/business-intelligence-dashboards Complete guide to BI dashboards, data visualization, and executive reporting for decision making. Covers dashboard design principles, KPI selection, and how lineage builds trust in dashboard numbers. ### Business Operating System **URL:** https://www.dawiso.com/glossary/business-operating-system Comprehensive framework for organizational excellence and strategic execution. Connects business operating systems to data infrastructure - how governed data, metrics, and metadata become the operating system layer. ### Change Data Capture (CDC) **URL:** https://www.dawiso.com/glossary/change-data-capture Change data capture (CDC) continuously tracks and streams database changes to downstream systems in near real-time. Covers log-based CDC, tools like Debezium, and how CDC fits into modern data pipelines. ### Connecting Power BI to Databricks - Complete Integration Guide **URL:** https://www.dawiso.com/glossary/connecting-power-bi-to-databricks-complete-integration-guide Connect Power BI to Databricks using Partner Connect or the native connector. Covers when to use DirectQuery vs Import, how to optimize SQL warehouse performance, and how Dawiso adds data context to reports. ### Cost Analysis **URL:** https://www.dawiso.com/glossary/cost-analysis Complete guide to cost analysis, financial analytics, and expense optimization for business intelligence. Covers cost categories, allocation models, and how data governance enables accurate cost reporting. ### Cost-Effective Data Management Strategies **URL:** https://www.dawiso.com/glossary/cost-effective-data-management-strategies Strategic approaches to optimizing data management costs while maintaining quality. Covers storage tiering, compute right-sizing, deduplication, and the role of governance in eliminating wasted data spend. ### Cost Efficiency **URL:** https://www.dawiso.com/glossary/cost-efficiency Comprehensive guide to optimizing cost efficiency in analytics and BI operations. Covers FinOps, query optimization, and metadata-driven cost monitoring. ### Cost Measurement **URL:** https://www.dawiso.com/glossary/cost-measurement Complete guide to measuring and tracking analytics costs for better financial control. Covers cost dimensions (compute, storage, licensing), allocation strategies, and showback/chargeback models. ### Cost Monitoring **URL:** https://www.dawiso.com/glossary/cost-monitoring Real-time cost monitoring strategies for analytics and BI financial control. Covers alerting thresholds, anomaly detection in cloud bills, and how lineage and usage metadata feed cost monitoring. ### Cost Reporting **URL:** https://www.dawiso.com/glossary/cost-reporting Comprehensive cost reporting strategies for analytics and BI financial oversight. Covers reporting cadences, audience-specific views, and integration of cost data with governance tooling. ### Cross-Filtering **URL:** https://www.dawiso.com/glossary/cross-filtering Interactive data exploration through dynamic cross-filtering in BI dashboards. Covers cross-filter direction, performance considerations, and how cross-filtering changes user behavior on dashboards. ### Data Catalog **URL:** https://www.dawiso.com/glossary/data-catalog A data catalog is the foundation of modern data governance. Explains what a data catalog is, how it works (automated scanning, business metadata, search, lineage), and why it's the prerequisite for AI readiness, self-service analytics, and trusted reporting. ### Data Classification **URL:** https://www.dawiso.com/glossary/data-classification Data classification organizes data into categories based on sensitivity, type, or business value. Explains how automated classification powers data governance, GDPR compliance, and security. ### Data Governance **URL:** https://www.dawiso.com/glossary/data-governance What data governance is, how to build a framework, and why it's essential for AI-ready organizations. Covers the four governance pillars (people, processes, policies, platforms), federated vs centralized models, and how to launch a governance program that delivers value within months. ### Data Lineage **URL:** https://www.dawiso.com/glossary/data-lineage What data lineage is, why it matters for data governance and compliance, and how to implement it. Covers table-level vs column-level lineage, manual vs automated extraction, and use cases (BCBS 239, GDPR, impact analysis, AI provenance). ### Data Masking **URL:** https://www.dawiso.com/glossary/data-masking Data masking protects sensitive information by replacing it with realistic but fictional values. Covers static vs dynamic masking, common techniques, and how masking supports GDPR compliance and secure analytics. ### Data Mesh vs Data Fabric **URL:** https://www.dawiso.com/glossary/data-mesh-vs-data-fabric Strategic comparison of data mesh vs data fabric architectures for enterprise data leaders. Covers principles, tooling, organizational implications, and when each architecture fits. ### Data Mesh vs Data Products **URL:** https://www.dawiso.com/glossary/data-mesh-vs-data-products Understand data mesh architecture vs data products for scalable enterprise data management. Explains how data products are the unit of exchange in a mesh, and how organizations can adopt data products without committing to full mesh implementation. ### Data Observability **URL:** https://www.dawiso.com/glossary/data-observability Data observability monitors the health of data pipelines through five pillars (freshness, distribution, volume, schema, lineage). Covers tools like Monte Carlo, Soda, and Bigeye, and how observability complements governance. ### Data Product **URL:** https://www.dawiso.com/glossary/data-product What a data product is, how it differs from a dataset, and why data products are central to data mesh architecture. Covers the data product canvas, ownership, SLAs, and how Dawiso's data product platform manages the full lifecycle. ### Data Sharing **URL:** https://www.dawiso.com/glossary/data-sharing Data sharing enables organizations to securely exchange data across teams, clouds, and company boundaries. Covers sharing models, platforms like Delta Sharing and Snowflake Marketplace, and the governance it requires. ### Data Vault **URL:** https://www.dawiso.com/glossary/data-vault Data Vault is a data warehousing modeling methodology built for scalability, auditability, and resilience to change. Covers Hubs, Links, Satellites, and when to choose Data Vault over Kimball or Inmon. ### Databricks Pricing Explained - Real Cost Breakdown for 2025 **URL:** https://www.dawiso.com/glossary/databricks-pricing-explained-real-cost-breakdown-for-2025 Complete Databricks pricing guide covering DBU costs, cloud expenses, optimization strategies, and real examples - for buyers comparing Databricks to alternative lakehouses. ### Databricks vs Snowflake - Which Data Platform Is Right for You **URL:** https://www.dawiso.com/glossary/databricks-vs-snowflake-which-data-platform-is-right-for-you Compare Databricks and Snowflake on architecture, use cases, performance, and how to choose the right platform. Covers lakehouse vs warehouse design, ML workloads, governance, and pricing models. ### Databricks with dbt - Modern Data Transformation Stack **URL:** https://www.dawiso.com/glossary/databricks-with-dbt-modern-data-transformation-stack How dbt and Databricks work together: SQL-based transformations, Delta Lake materializations, automated testing, CI/CD pipelines, and how Dawiso extends dbt documentation into a cross-platform catalog. ### dbt Lineage **URL:** https://www.dawiso.com/glossary/dbt-lineage dbt automatically generates data lineage through ref() and source() functions, creating a DAG of all transformations. Covers column-level lineage, OpenLineage integration, and connecting dbt to enterprise governance. ### dbt Models **URL:** https://www.dawiso.com/glossary/dbt-models A dbt model is a SQL SELECT statement that defines a data transformation in dbt. Covers materializations (table, view, incremental, ephemeral), the staging-marts layer pattern, and how dbt models generate metadata for governance. ### dbt Tests **URL:** https://www.dawiso.com/glossary/dbt-tests dbt tests are automated data quality checks that run after transformations to validate your data. Covers built-in generic tests, custom singular tests, and how dbt testing integrates with data observability. ### Feature Engineering **URL:** https://www.dawiso.com/glossary/feature-engineering Feature engineering transforms raw data into meaningful variables for machine learning models. Covers common techniques (encoding, scaling, aggregation, time-window features) and the role of governed data in reproducible feature pipelines. ### Feature Store **URL:** https://www.dawiso.com/glossary/feature-store Feature stores provide centralized management and serving of ML features for consistent access. Covers offline vs online stores, feature reuse across models, and the governance challenges feature stores create. ### Federated Learning **URL:** https://www.dawiso.com/glossary/federated-learning Federated learning enables collaborative ML training while preserving data privacy and locality. Covers how it works, where it's used (healthcare, financial services, mobile), and the governance implications of training models on data you can't see. ### 5 Reasons Why Companies Are Migrating to Databricks **URL:** https://www.dawiso.com/glossary/5-reasons-why-companies-are-migrating-to-databricks Discover 5 key reasons why companies migrate to Databricks: lakehouse architecture, ML platform, performance, cost efficiency, and team collaboration. ### GraphRAG **URL:** https://www.dawiso.com/glossary/graph-rag GraphRAG combines knowledge graphs with retrieval-augmented generation to give LLMs richer, relationship-aware context. Covers how it reduces hallucinations and enables complex enterprise Q&A. ### How Are Data Products Connected **URL:** https://www.dawiso.com/glossary/how-are-data-products-connected Learn how data products connect to build scalable, integrated enterprise data ecosystems. Covers product-to-product dependencies, contracts between data products, and federation across domains. ### Key Performance Indicator (KPI) **URL:** https://www.dawiso.com/glossary/key-performance-indicator--kpi Complete guide to KPIs, business metrics, and performance management for data-driven decisions. Covers KPI selection, definition governance, and the link between KPIs and business strategy. ### Machine Learning Operations (MLOps) **URL:** https://www.dawiso.com/glossary/machine-learning-operations--mlops MLOps provides practices and tools for deploying and maintaining ML models in production. Covers CI/CD for ML, model monitoring, drift detection, and how MLOps connects to data governance. ### Medallion Architecture **URL:** https://www.dawiso.com/glossary/medallion-architecture Medallion architecture organizes data lakehouse storage into Bronze, Silver, and Gold layers - each with increasing quality and structure. Covers the pattern, its benefits, and how to govern data across layers. ### Natural Language Processing (NLP) **URL:** https://www.dawiso.com/glossary/natural-language-processing--nlp Natural Language Processing enables computers to understand and generate human language. Covers core NLP tasks (tokenization, NER, sentiment analysis, summarization) and the shift to LLM-based NLP. ### Power BI AI Insights **URL:** https://www.dawiso.com/glossary/power-bi-ai-insights-complete-guide-to-intelligent-data-analysis Unlock automated insights and predictive analytics with Power BI AI for smarter decisions. Covers Quick Insights, Decomposition Tree, Key Influencers, and the AI features that augment dashboards. ### Power BI Copilot **URL:** https://www.dawiso.com/glossary/power-bi-copilot-complete-guide-to-ai-powered-business-intelligence Boost productivity with Power BI Copilot AI assistant for automated report generation. Covers natural language report creation, DAX assistance, and prerequisites (semantic models, governance). ### Power BI Data Modeling **URL:** https://www.dawiso.com/glossary/power-bi-data-modeling-complete-guide-to-effective-data-architecture Power BI data modeling guide: star schema design, relationship configuration, measures vs calculated columns, performance optimization, and row-level security - with concrete DAX examples. ### Power BI Deployment Pipelines and Source Control **URL:** https://www.dawiso.com/glossary/power-bi-deployment-pipelines-and-source-control-enterprise-devops-guide Streamline Power BI DevOps with deployment pipelines and source control automation. Covers Dev/Test/Prod stages, PBIP format, Git integration, and CI/CD for enterprise BI. ### Power BI Power Query Tutorials **URL:** https://www.dawiso.com/glossary/power-bi-power-query-tutorials-complete-guide-to-data-transformation Transform messy data with Power BI Power Query for clean, analysis-ready datasets. Covers M language basics, common transformations, query folding, and performance patterns. ### Power BI Predictive Analytics **URL:** https://www.dawiso.com/glossary/power-bi-predictive-analytics-complete-guide-to-forecasting-and-machine-learning Predict future trends with Power BI machine learning and advanced forecasting models. Covers built-in forecasting, Azure ML integration, and AutoML in Power BI dataflows. ### Power BI Real-Time Dashboards **URL:** https://www.dawiso.com/glossary/power-bi-real-time-dashboards-complete-guide-to-live-data-visualization Monitor live data with Power BI real-time dashboards for instant business insights. Covers streaming datasets, push datasets, DirectQuery patterns, and architectural trade-offs. ### Power BI Row-Level Security (RLS) **URL:** https://www.dawiso.com/glossary/power-bi-row-level-security--rls--complete-data-access-control-guide Secure sensitive data with Power BI Row-Level Security for granular access control. Covers static vs dynamic RLS, DAX filter expressions, and how RLS interacts with workspace permissions. ### Power BI Translytical Task Flows **URL:** https://www.dawiso.com/glossary/power-bi-translytical-task-flows-complete-guide-to-modern-analytics-workflows Optimize analytics workflows with Power BI translytical task flows for efficiency. Covers the merging of analytics and transactions, write-back scenarios, and modern operational BI patterns. ### Predictive Analytics **URL:** https://www.dawiso.com/glossary/predictive-analytics Predictive analytics uses statistical models and ML to forecast future outcomes and trends. Covers the predictive analytics workflow, common algorithms, and how to operationalize predictions in BI tools. ### Real-Time Analytics **URL:** https://www.dawiso.com/glossary/real-time-analytics Complete guide to real-time analytics, streaming data processing, and instant business intelligence. Covers Kafka, Flink, streaming SQL, and where real-time analytics adds business value. ### Reverse ETL **URL:** https://www.dawiso.com/glossary/reverse-etl Reverse ETL moves processed data from a data warehouse back into operational tools like CRMs, ad platforms, and customer success systems. Covers data activation, common tools (Hightouch, Census), and what governance it requires. ### SQL COUNT **URL:** https://www.dawiso.com/glossary/sql-count-complete-guide-to-counting-records-in-database-queries Master SQL COUNT function for accurate database record counting and data analysis. Covers COUNT(*), COUNT(column), COUNT(DISTINCT), and performance considerations across engines. ### SQL DELETE **URL:** https://www.dawiso.com/glossary/sql-delete-complete-guide-to-removing-database-records Learn SQL DELETE statement for safe database record removal and data management. Covers DELETE vs TRUNCATE, transactional safety, and cascade considerations. ### SQL GROUP BY **URL:** https://www.dawiso.com/glossary/sql-group-by-complete-guide-to-data-aggregation-and-analysis Master SQL GROUP BY for powerful data aggregation, grouping, and business analytics. Covers HAVING, ROLLUP, CUBE, and GROUPING SETS with practical examples. ### SQL INSERT **URL:** https://www.dawiso.com/glossary/sql-insert-comprehensive-guide-to-database-data-insertion Master SQL INSERT statement for efficient database data insertion and record creation. Covers single-row, multi-row, INSERT-SELECT, and ON CONFLICT patterns. ### SQL JOIN **URL:** https://www.dawiso.com/glossary/sql-join-complete-guide-to-database-table-relationships Master SQL JOIN operations for combining tables and database relationship queries. Covers INNER, LEFT, RIGHT, FULL OUTER, CROSS, and SELF joins with diagrams and examples. ### SQL PARTITION BY **URL:** https://www.dawiso.com/glossary/sql-partition-by-complete-guide-to-window-functions-and-data-partitioning Learn SQL PARTITION BY for advanced window functions and data partitioning techniques. Covers ROW_NUMBER, RANK, DENSE_RANK, and how PARTITION BY differs from GROUP BY. ### SQL SELECT **URL:** https://www.dawiso.com/glossary/sql-select-complete-guide-to-database-queries Master SQL SELECT statement fundamentals for database querying and data retrieval. Covers projections, predicates, joins, and ordering with practical examples. ### SQL SUM() OVER **URL:** https://www.dawiso.com/glossary/sql-sum--over-complete-guide-to-window-functions-and-running-totals Learn SQL SUM OVER for running totals, cumulative sums, and window function analytics. Covers frame clauses, ordering, and performance considerations. ### SQL Aggregate Functions (SUM, AVG, MIN, MAX) **URL:** https://www.dawiso.com/glossary/sql-sum-avg-min-max-complete-guide-to-sql-aggregate-functions Master SQL aggregate functions SUM, AVG, MIN, MAX for data analysis and calculations. Covers NULL handling, DISTINCT, and integration with GROUP BY. ### SQL UPDATE **URL:** https://www.dawiso.com/glossary/sql-update-complete-guide-to-updating-database-records Learn SQL UPDATE statement for modifying database records and data maintenance. Covers UPDATE-FROM, UPDATE with subqueries, and transactional safety. ### SQL WHERE Clause **URL:** https://www.dawiso.com/glossary/sql-where-clause-complete-guide-to-data-filtering Master SQL WHERE clause for precise data filtering and conditional query logic. Covers comparison, BETWEEN, IN, LIKE, NULL handling, and indexability of predicates. ### Visualization Tools **URL:** https://www.dawiso.com/glossary/visualization-tools Complete guide to data visualization tools, BI platforms, and visual analytics for business insights. Covers Power BI, Tableau, Qlik, Looker, and emerging tools, plus principles of effective visualization. ### What Is a Metadata Management? **URL:** https://www.dawiso.com/glossary/metadata-management Learn what metadata management is, why it matters for AI readiness, and how to implement it effectively. Covers metadata categories (technical, business, operational), active vs passive metadata, and modern metadata management platforms. ### What Is a Semantic Layer? **URL:** https://www.dawiso.com/glossary/semantic-layer Learn what a semantic layer is, how it works, and why it's critical for business-friendly analytics and AI-ready data. Covers semantic layer architecture, metric definitions, and integration with BI tools and AI agents. ### What Is Databricks - Complete Guide for 2025 **URL:** https://www.dawiso.com/glossary/what-is-databricks-complete-guide-for-2025 Complete 2025 guide to Databricks: unified analytics platform, lakehouse architecture, and enterprise features. For buyers, architects, and engineers evaluating Databricks for data and AI workloads. ### What Is OpenMetadata? **URL:** https://www.dawiso.com/glossary/what-is-openmetadata OpenMetadata is an open-source (Apache 2.0) metadata platform for data cataloging, discovery, lineage, and governance, open-sourced in 2021 from the team behind Uber's metadata systems. Covers the four-component architecture (Java metadata server, Python ingestion framework, JavaScript/TypeScript UI, relational database plus search index), 120-plus connectors, the unified metadata graph, key features, what "open source" means for operating cost, and how a self-hosted toolkit compares to a governed, managed catalog like Dawiso served to AI agents via MCP. ### OpenMetadata Pricing: Is It Really Free? **URL:** https://www.dawiso.com/glossary/openmetadata-pricing OpenMetadata's software is free under Apache 2.0, but production use carries real cost. Separates software price (free) from total cost of ownership (cloud infrastructure, deployment engineering, and ongoing maintenance for self-hosting). Covers Collate, the managed cloud service from the project's creators, including its free tier (inactive clusters reclaimed after four weeks) and paid plans priced through sales and cloud marketplaces, plus how Dawiso compares on total cost and time to value. ### DataHub Pricing: The Open-Core Model **URL:** https://www.dawiso.com/glossary/datahub-pricing DataHub Core is free under Apache 2.0, but two costs hide behind that: self-hosting a multi-service stack (Kafka, Elasticsearch, the metadata service, the store, and the frontend), commonly 6 to 12 weeks to production plus ongoing maintenance, and the open-core boundary that places governance-lifecycle and enterprise features in paid DataHub Cloud, priced through sales. Covers self-hosting total cost of ownership, DataHub Cloud pricing, why teams running open-core catalogs re-evaluate and look at managed vendors like Dawiso, and how Dawiso compares with the same feature set on every plan and no open-core upgrade tax. ### Snowflake Pricing: Credits, Storage, and What You Actually Pay **URL:** https://www.dawiso.com/glossary/snowflake-pricing Snowflake is consumption-priced across three independent meters: compute in credits (billed per second with a 60-second minimum each time a warehouse starts), storage at roughly $23 per TB per month after compression, and additional services (serverless features and data egress) billed on their own. The per-credit price rises with the edition (Standard around $2, Enterprise around $3, Business Critical around $4 on US East AWS on-demand), and warehouse size doubles credits-per-hour at each step. Covers the consumption model, editions, per-second billing and the 60-second minimum, storage and Time Travel, serverless and cloud-services costs, realistic small/mid/enterprise scenarios, optimization (auto-suspend, right-sizing, capacity commitments), hidden costs, and how Dawiso's catalog, lineage, and glossary cut Snowflake spend by retiring unused pipelines and duplicate transformations. ### Microsoft Fabric Pricing: Capacity Units, F-SKUs, and What You Pay **URL:** https://www.dawiso.com/glossary/microsoft-fabric-pricing Microsoft Fabric is priced by capacity, sold as F-SKUs that double in Capacity Units from F2 (about $260 per month PAYG) to F2048, plus per-user licenses and OneLake storage (about $23 per TB per month). An F64 runs roughly $8,000 to $8,500 per month on demand; a 1-year reservation saves about 41 percent but cannot be paused, while PAYG can be paused to stop the compute meter. The key cliff is F64: below it every Power BI viewer needs a paid Pro or PPU license, at F64 and above free-license users can view content. Covers the capacity model, F-SKUs and CUs, PAYG vs reserved, per-user licensing and the F64 cliff, OneLake storage and paused-capacity billing, scenarios, optimization, hidden costs, and how Dawiso controls Fabric spend. ### Collibra Pricing: What It Costs and Why There Is No Public Price **URL:** https://www.dawiso.com/glossary/collibra-pricing Collibra publishes no public price and sells only through custom quotes (also private offers on cloud marketplaces). Pricing is tiered and scales with users/roles, editions/modules, scale, and term, but the license is typically only about 15 to 20 percent of first-year cost, because implementation, professional services, and ongoing administration and adoption dominate. Third-party procurement data puts a typical mid-market deployment in the low hundreds of thousands of dollars per year, with large enterprise rollouts into seven figures. Covers whether Collibra publishes pricing, how it is priced, total cost of ownership, what teams report paying, why teams re-evaluate, and how Dawiso compares with transparent per-seat pricing, faster time to value, and business-user adoption. ### What Is Model Context Protocol (MCP)? **URL:** https://www.dawiso.com/glossary/model-context-protocol-mcp Learn what Model Context Protocol (MCP) is, how it connects AI agents to enterprise data, and why it matters for reliable AI applications. Covers MCP architecture, the client-server model, server implementations, and how Dawiso's MCP Server exposes the AI Context Layer to Claude Desktop, Cursor, GitHub Copilot, and Keboola. ### What Is RAG in AI **URL:** https://www.dawiso.com/glossary/what-is-rag-in-ai Understand RAG (Retrieval-Augmented Generation) in AI, how it works, and why it's essential for accuracy. Covers retrieval pipelines, embedding strategies, and the role of governed data in reliable RAG systems. ### Context Q&A Topics A series of accessible explainers on AI context concepts: - [What is an example of a context setting](https://www.dawiso.com/glossary/what-is-an-example-of-a-context-setting): Examples of context setting from literature, research, business, and communication. - [What is an example of context in AI](https://www.dawiso.com/glossary/what-is-an-example-of-context-in-ai): Concrete examples of context in AI from language understanding to personalization and decision-making. - [What is an example of contextual AI](https://www.dawiso.com/glossary/what-is-an-example-of-contextual-ai): Real-world examples of contextual AI from virtual assistants to recommendations and fraud detection. - [What is context AI](https://www.dawiso.com/glossary/what-is-context-ai): Context AI explained - how it uses contextual information and why it creates smarter AI systems. - [What is context simply](https://www.dawiso.com/glossary/what-is-context-simply): Simple explanation of context, what it means, why it matters, and how it shapes understanding. - [What is context understanding in AI](https://www.dawiso.com/glossary/what-is-context-understanding-in-ai): How AI systems comprehend situations and why it's crucial for intelligence. - [What is research context](https://www.dawiso.com/glossary/what-is-research-context): What research context means, why it matters, and how it shapes study design and interpretation. - [Who is the founder of context AI](https://www.dawiso.com/glossary/who-is-the-founder-of-context-ai): The background, vision, and impact of the Contextual AI founder on enterprise AI. - [Why is C.AI asking my age](https://www.dawiso.com/glossary/why-is-c-ai-asking-my-age): Why AI systems request age - legal compliance, safety, and personalization. - [Why is context important](https://www.dawiso.com/glossary/why-is-context-important): Why context is vital for communication, decisions, learning, and information. - [Why is context important in AI](https://www.dawiso.com/glossary/why-is-context-important-in-ai): Why context is crucial in AI for accurate decisions, natural interactions, and robust systems. - [Why is context important in NLP](https://www.dawiso.com/glossary/why-is-context-important-in-nlp): Why context is essential in NLP for understanding language, resolving ambiguity, and powering AI. - [What Is a Context Window in AI?](https://www.dawiso.com/glossary/context-window): A context window is the maximum amount of text an LLM can process at once. Covers how tokens work, current 2026 model sizes (Llama 4 Scout 10M, Gemini 3 Pro 2M, Claude Opus 4.6 1M, GPT-5.4 272K-1M, Claude Sonnet 4.6 1M), why bigger windows do not replace RAG, the lost-in-the-middle problem, cost and latency trade-offs, and best practices for building production LLM applications. - [Character.AI Safety Guide for Parents and Teens (2026)](https://www.dawiso.com/glossary/character-ai-safety-guide): Parents' guide to Character.AI in 2026. Covers the November 25, 2025 under-18 chat ban (limiting minors to pre-created scenes only), the age-assurance system using login signals plus selfie verification, Parental Insights, content filters, time notifications, known residual risks, off-platform clones, and practical conversation tips for parents. - [Databricks vs Microsoft Fabric: Complete Platform Comparison](https://www.dawiso.com/glossary/databricks-vs-microsoft-fabric): Side-by-side comparison of the two leading 2026 data platforms across architecture (PaaS lakehouse vs SaaS suite), pricing models (DBU usage-based vs F-SKU capacity), AI/ML capabilities (Mosaic AI vs Copilot), governance (Unity Catalog vs OneLake + Purview), and integrations. Includes when-to-choose guidance for each. - [Databricks Unity Catalog: Complete Guide to Data Governance](https://www.dawiso.com/glossary/databricks-unity-catalog): Unity Catalog is Databricks' centralized governance plane. Covers core concepts (metastore, catalog, schema, table, view, volume, function, model), the three-level namespace (catalog.schema.object), ANSI SQL access control with row filters and column masks, automatic column-level lineage, audit logs, the Catalog Explorer, volumes for unstructured data, and the open-source Unity Catalog project. - [Power BI vs Tableau vs Looker: BI Tool Comparison (2026)](https://www.dawiso.com/glossary/power-bi-vs-tableau-vs-looker): Side-by-side comparison of the three dominant BI platforms. Covers pricing (Power BI from $10/user/month, Tableau Creator at $75, Looker custom-quoted), ease of use, AI features (Copilot, Tableau Pulse + Agent, Looker Gemini), governance and semantic layer approach (DAX models, calculated fields, LookML), integrations, and use-case-based recommendations. - [Power BI Fabric Integration: Complete Enterprise Guide](https://www.dawiso.com/glossary/power-bi-fabric-integration): How Microsoft Fabric integrates with Power BI. Explains OneLake as the unified storage layer, Direct Lake mode (read Delta tables at import-mode speed without copy or refresh), semantic models, workspaces and sharing, performance considerations including DirectQuery fallback, capacity sizing, V-Order optimization, and migration from Power BI Premium. - [Power BI DAX Functions: Complete Beginner's Guide](https://www.dawiso.com/glossary/power-bi-dax-functions-guide): DAX is Power BI's formula language. Covers the difference between calculated columns and measures, essential functions (SUM, CALCULATE, FILTER, RELATED, IF, SWITCH, VAR/RETURN, DIVIDE), time intelligence (SAMEPERIODLASTYEAR, TOTALYTD, DATEADD), the all-important filter context concept, common patterns (YoY %, running totals, % of total, Top N), and performance tips. - [What Is Microsoft Fabric?](https://www.dawiso.com/glossary/microsoft-fabric): Microsoft Fabric is the unified SaaS analytics platform. Covers core components (Data Factory, Data Engineering, Data Warehouse, Real-Time Intelligence, Data Science, Power BI), the OneLake foundation with shortcuts and Iceberg interop, Direct Lake mode, capacity-based F-SKU pricing with 41% reservation savings, comparison with Azure Synapse, and common use cases. - [What Is Snowflake?](https://www.dawiso.com/glossary/snowflake): Snowflake is a cloud data platform built on a multi-cluster shared-data architecture. Covers the three layers (storage on cloud object storage, compute via virtual warehouses, cloud services for metadata and security), virtual warehouse sizing and auto-suspend, Time Travel, zero-copy cloning, key features (Snowpark, Streams & Tasks, Dynamic Tables, Marketplace, Secure Data Sharing, Native Apps), credit-based pricing, Snowflake Cortex AI, and use cases. - [What Is Delta Lake?](https://www.dawiso.com/glossary/delta-lake): Delta Lake is an open table format that wraps Parquet files with a JSON transaction log to add ACID transactions, time travel, schema enforcement, MERGE/UPDATE/DELETE, and Z-Order optimization. Covers how the _delta_log/ works, comparison with Apache Iceberg and Apache Hudi, implementation with Spark, Delta UniForm for Iceberg interop, and use cases including medallion architecture, GDPR deletes, and Direct Lake on Microsoft Fabric. - [What Is Total Cost of Ownership (TCO)?](https://www.dawiso.com/glossary/total-cost-of-ownership): TCO is the full lifecycle cost of acquiring, operating, and retiring an asset --- not just its purchase price. Covers the three cost layers (acquisition, operating, hidden/indirect) using an iceberg model, the distinction between TCO, ROI, and purchase price, a five-step calculation method, why data platforms (Snowflake, Databricks) are dominated by usage-driven and hidden costs, and how data governance lowers TCO by removing the largest hidden cost of all --- data nobody can find, understand, or trust. - [What Is Snowflake Cortex?](https://www.dawiso.com/glossary/snowflake-cortex): Snowflake Cortex is Snowflake's suite of managed, serverless AI features that run on governed data inside the platform. Covers the four layers --- AISQL/LLM functions (COMPLETE, SUMMARIZE, TRANSLATE, SENTIMENT, CLASSIFY_TEXT, EXTRACT_ANSWER, EMBED_TEXT) over hosted foundation models (Claude, Llama, Mistral, OpenAI, Arctic); Cortex Analyst for text-to-SQL on structured data; Cortex Search for hybrid vector + keyword RAG over text; and Cortex Agents that orchestrate them --- why every call inherits Snowflake's role-based access control, and why answer accuracy still depends on governed business context (glossary, lineage, catalog) that Dawiso supplies. - [What Are Snowflake Semantic Views & Cortex Analyst?](https://www.dawiso.com/glossary/snowflake-semantic-views): A semantic view is a native Snowflake object that encodes business meaning --- logical tables, relationships and join paths, facts, dimensions, named metrics, synonyms, and verified query examples --- over physical tables, and carries role-based access control. Cortex Analyst reads the semantic view (not raw tables) to turn plain-English questions into governed SQL, separating a right answer from a plausible wrong one. Covers semantic views vs staged YAML semantic models, the Open Semantic Interchange (OSI) vendor-neutral standard (Snowflake + Salesforce, dbt Labs, BlackRock, RelationalAI, Sept 2025), and how Dawiso scans agents and views into lineage and generates governed semantic views back into Snowflake via OSI. - [What Is the EU AI Act?](https://www.dawiso.com/glossary/eu-ai-act): The EU AI Act (Regulation (EU) 2024/1689, in force 1 Aug 2024) is the world's first comprehensive, horizontal AI law, regulating AI by risk rather than sector. Covers the four risk tiers --- unacceptable (prohibited: social scoring, manipulative AI, untargeted face scraping), high risk (strict obligations: risk management, Article 10 data governance, documentation, logging, human oversight, conformity assessment), limited risk (transparency: label chatbots and deepfakes), and minimal risk (no obligations); the separate GPAI model regime including systemic-risk models; the phased timeline (Feb 2025 prohibitions, Aug 2025 GPAI, Aug 2026 most high-risk, Aug 2027 product safety components); penalties up to -��35M or 7% of global turnover; and why Article 10 makes documented, high-quality, well-governed training data --- catalog, lineage, classification --- a legal prerequisite. - [What Is the EU AI Omnibus (Digital Omnibus)?](https://www.dawiso.com/glossary/ai-omnibus): The Digital Omnibus is a simplification package the European Commission proposed on 19 November 2025 to amend several EU digital laws at once. It split into two tracks. The AI Act amendments ("AI Omnibus") reached a provisional Parliament---Council agreement on 7 May 2026: high-risk obligations postponed --- standalone (Annex III) to 2 December 2027 and product-embedded (Annex I) to 2 August 2028, tied to the readiness of harmonised standards; Article 50 transparency holds at 2 August 2026 with a watermarking grace period to 2 December 2026; SME-style relief extended to small mid-caps; and the AI Office's GPAI enforcement powers strengthened --- with formal adoption expected by mid-2026. The broader GDPR and ePrivacy amendments are a separate package that, as of mid-2026, has not reached political agreement and remains the most contested. Covers what it proposes, the goals (reduce burden, align overlapping duties, consolidate incident reporting across NIS2/DORA/GDPR/CRA), and why postponed deadlines are breathing room not a reprieve --- the obligations remain, and one governed source of truth is what makes meeting them tractable. - [What Is ISO 42001 (AI Management System)?](https://www.dawiso.com/glossary/iso-42001): ISO/IEC 42001:2023 is the first international standard for an AI management system (AIMS) --- a voluntary, certifiable framework for governing AI responsibly across its lifecycle. Covers its management-system nature (certifies the organization's process, not the model), the Annex SL / Plan-Do-Check-Act structure across clauses 4---10 (context, leadership, planning with AI impact assessment, support, operation, performance evaluation, improvement), the Annex A AI-specific controls (AI policy, roles, impact assessment, data for AI systems, lifecycle, transparency, supplier management), certification like ISO 27001, its complementary relationship to the binding EU AI Act, and why a certifiable AIMS rests on documented, governed data (catalog, lineage, quality, ownership) as audit evidence. - [What Is the NIST AI Risk Management Framework?](https://www.dawiso.com/glossary/nist-ai-rmf): The NIST AI RMF is a voluntary, technology-neutral framework (AI RMF 1.0, released 26 January 2023) for managing AI risk across the lifecycle, developed under the National AI Initiative Act through an open consensus process. Covers the four core functions (Govern, a cross-cutting function for culture, policy, roles and accountability; Map, establishing context and identifying risks; Measure, analyzing, benchmarking and monitoring risk and trustworthiness; Manage, prioritizing and acting on risk), the companion resources (Playbook, Roadmap, Crosswalk, profiles), the Generative AI Profile (NIST-AI-600-1, released 26 July 2024) extending the functions to generative and agentic AI risks (confabulation, data leakage, IP exposure), the ongoing revision and the April 2026 critical-infrastructure profile concept note, how it complements the binding EU AI Act and the certifiable ISO 42001 (practice vs legal floor vs certifiable system), and why every function depends on governed context (catalog, glossary, lineage, classification, ownership) to be operable, especially for AI agents. - [What Is Model Risk Management (MRM)?](https://www.dawiso.com/glossary/model-risk-management): Model risk management (MRM) is the discipline of approving, controlling, and monitoring models as ongoing risk-bearing systems rather than one-off analytical projects. Covers model risk as both a wrong model and a misused one; the five activities (identification and inventory, development and testing, independent validation, ongoing monitoring, governance and controls) and the concept of effective challenge; what counts as a model under the US definition (a complex quantitative method applying statistical, economic, or financial theories, excluding spreadsheet arithmetic and deterministic rule engines) and that a vendor model remains your accountability; model materiality (exposure plus purpose) as the driver of proportionate control depth, including the legitimate immaterial tier; the joint Fed/OCC/FDIC SR 26-2 Revised Guidance on Model Risk Management issued 17 April 2026, which supersedes SR 11-7 (2011) and SR 21-8 (2021), is most relevant to banking organizations over $30 billion in assets, and states that its principles apply to traditional quantitative models and non-generative, non-agentic AI models - so generative and agentic AI need AI governance frameworks alongside MRM; the UK PRA SS1/23 (published 17 May 2023, effective 17 May 2024, five principles, all models informing business decisions regardless of technology, including vendor models) and the PRA's 2025 MRM roundtable on AI and machine learning; and why every MRM control resolves into a data question, with the honest boundary that Dawiso is not a model validation platform but supplies the data-side evidence (catalog, governed definitions, lineage, ownership, classification). - [What Is AI Lifecycle Governance?](https://www.dawiso.com/glossary/ai-lifecycle-governance): AI lifecycle governance is the practice of governing an AI system continuously from intake through design, validation, release, operation, change, and decommissioning, on the premise that a launch approval is not governance. Covers why AI systems drift through channels traditional change control does not watch (prompt edits, model version swaps, extended retrieval sources, added agent tools, widening purpose); the three defining properties (continuous, proportionate to risk tier, evidence-producing); the seven stages with a governance question attached to each; the three gates that carry the weight (intake tier and owner, release on validated residual risk, and the change gate most programs lack); the regulatory basis, including EU AI Act Article 17's documented quality management system and Article 72's post-market monitoring obligation with its monitoring plan as part of technical documentation, ISO 42001's Annex A lifecycle controls, and the NIST AI RMF's four continuous functions; why change and retirement are governed worst; the distinction from MLOps/LLMOps (execution), AI governance (the whole framework), and model risk management (the closest sibling, which as of SR 26-2 excludes generative and agentic AI); and how Dawiso supplies the data-side evidence each gate asks for via catalog, business glossary, classification, lineage, and MCP-served context. - [What Is AI Risk Tiering?](https://www.dawiso.com/glossary/ai-risk-tiering): AI risk tiering classifies AI use cases by impact and applies governance controls proportionate to that risk, solving the mismatch between use-case volume and review capacity. Covers the three properties of a usable scheme (few tiers, assigned at intake and revisited, consequential); the origin in model risk management's model materiality, including the insight that a light bottom tier is a legitimate control rather than an absence of governance; the six tiering dimensions (impact on people, decision autonomy, data sensitivity, reversibility, scale and reach, explainability need); what each tier changes across documentation, who validates, human oversight, monitoring cadence, and change control, plus the two anti-patterns (a top tier so heavy that teams misclassify, and a bottom tier with no tripwire); the critical distinction from the EU AI Act's legal categories, where Article 6 and Annex III fix the classification by law with a narrow documented derogation and mandatory high-risk status for profiling, so an internal tier is independent and legal compliance is a floor rather than a tier; a six-step build sequence starting from a real inventory that includes embedded and purchased AI; and how Dawiso answers the data-sensitivity and reach dimensions through classification, lineage, glossary, and ownership without assigning tiers itself. - [What Is Explainable AI (XAI)?](https://www.dawiso.com/glossary/explainable-ai-xai): Explainable AI (XAI) is the ability to reconstruct and justify how an AI system produced a particular outcome, which in practice means auditability rather than transparency for its own sake. Covers why explanations are audience-relative (developer, affected person, auditor, regulator) and why they decay unless recorded; NIST IR 8312's four principles published 29 September 2021 (the system produces an explanation, the explanation is meaningful to humans, it accurately reflects the system's processes, and the system expresses its knowledge limits) and why they fail independently; explainability versus interpretability and why an inherently interpretable model can be the cheaper choice for a top-tier decision; why a generative model's fluent rationale is not evidence and LLM explainability lives in retrieval, governed definitions, tool calls, and permissions; the technique families (feature attribution, counterfactuals, surrogate models, example-based, documentation artifacts) with the failure mode of each; where law requires it, including EU AI Act Article 86's right to clear and meaningful explanation of the AI system's role in Annex III high-risk decisions with legal effects, the Act's transparency and documentation duties, GDPR Article 22's restrictions and safeguards on solely automated decisions, and sector adverse-action rules; and why a model-level explanation is unfalsifiable without the data layer, which lineage, a business glossary, the catalog, and classification supply. - [What Is an AI Maturity Assessment?](https://www.dawiso.com/glossary/ai-maturity-assessment): An AI maturity assessment is a structured evaluation of how repeatably an organization can build, govern, and scale AI, as opposed to how much AI activity it currently has. Covers the three characteristics of an assessment that changes behavior (evidence-based rather than self-reported, dimensional rather than a single composite score, tied to an owned roadmap); the five levels of Gartner's widely used model - Awareness (talking, not strategic, no pilots), Active (proofs of concept and pilots, knowledge sharing), Operational (at least one project in production with accessible practice and expertise), Systematic (AI embedded in the design of products, services, and processes), Transformational (AI reshapes decision making, operating models, and competitive advantage) - with most organizations still in awareness; the dimensions assessed (strategy and value, use-case portfolio, data foundation, governance, engineering and operations, people and operating model) and why presenting the lowest dimensions beats averaging; why the step from Active to Operational is the hard one, because a pilot needs a champion while a production system needs an owner, governed data, monitoring, a change process, and a retirement plan; the opposite failure of governance built as a gate with no service; a six-step method for running one honestly, starting from the inventory rather than the survey; and the asymmetric relationship to data maturity, where AI maturity is capped by data maturity but not the reverse. - [What Is AI Transformation?](https://www.dawiso.com/glossary/ai-transformation): AI transformation is the organization-wide shift to becoming a business that operates and decides with AI woven into how work gets done --- the successor to digital transformation. Covers its five interdependent pillars (strategy, people & culture, technology, operating model, governance) all standing on a foundation of AI-ready governed data; the distinction from digital transformation (which produced the data AI now consumes); why most transformations stall on data readiness rather than model quality (data AI cannot find, trust, or understand; pilots that never industrialize; literacy gaps); and why the data catalog, business glossary, lineage, and the Dawiso Context Layer are the load-bearing foundation that sets the ceiling for any AI ambition. - [What Is a Chief Data Officer (CDO)?](https://www.dawiso.com/glossary/chief-data-officer): The CDO is the senior executive accountable for an organization's data as a strategic asset --- its strategy, governance, quality, and value. Covers the role's origin in post-2008 financial regulation (BCBS 239), its responsibilities across a defensive-to-offensive spectrum (governance, risk & compliance, data quality on one side; analytics & AI enablement, value/monetization, data culture & literacy on the other), the distinction from CIO (technology), CISO (security), and the increasingly common CDAO (Chief Data & Analytics Officer), the mandate's shift from defensive control toward offensive value and AI enablement, why CDO tenure is often short without a mature data foundation, and why the catalog, glossary, lineage, and ownership are the CDO's core operating infrastructure. - [What Is a Data Maturity Model?](https://www.dawiso.com/glossary/data-maturity-model): A data maturity model is a staged framework for assessing how capable an organization is at managing and using data. Covers the five-level progression inherited from the Capability Maturity Model lineage (1 Initial/ad hoc, 2 Managed/reactive, 3 Defined/proactive & governed, 4 Quantitatively Managed/measured, 5 Optimized/continuously improving & AI-ready); the multiple dimensions assessed independently (strategy & leadership, governance, data quality, architecture, discoverability/metadata, literacy & culture); established frameworks (EDM Council DCAM, CMMI DMM, DAMA-DMBOK, Gartner and vendor models); why most organizations sit between Levels 2 and 3 and the catalog/glossary jump to Level 3 unlocks reuse and AI; and how a governed catalog, glossary, lineage, and ownership are the concrete mechanics of moving up the curve. - [What Is GDPR?](https://www.dawiso.com/glossary/gdpr): The General Data Protection Regulation (Reg. (EU) 2016/679, in force 25 May 2018) is the EU's data protection law, applying extraterritorially to any organization processing the personal data of people in the EU. Covers the seven Article 5 principles (lawfulness/fairness/transparency, purpose limitation, data minimisation, accuracy, storage limitation, integrity & confidentiality, accountability), the eight data subject rights (information, access, rectification, erasure/right to be forgotten, restriction, portability, object, automated-decision safeguards), the six lawful bases, the controller/processor/DPO roles, penalties up to -��20M or 4% of global turnover with 72-hour breach notification, and why GDPR is operationally a data governance program built on knowing where personal data lives (catalog), what it is (classification), and where it flows (lineage). - [What Is Data Sovereignty?](https://www.dawiso.com/glossary/data-sovereignty): Data sovereignty is the principle that data is governed by the laws of the jurisdiction where it is collected or stored --- concerning whose law applies and which government can compel access. Covers the crucial distinction from data residency (where data is physically stored, a business choice) and data localization (a legal requirement to keep data in-country); why it matters (cloud concentration, GDPR transfer rules, geopolitics, conflicting laws); the laws that drive it (GDPR Chapter V, Schrems II, the US CLOUD Act, national localization mandates); why a local cloud region does not by itself guarantee sovereignty; and how a governed catalog, classification, and lineage let an organization prove, not just assert, that the right data sits in the right jurisdiction. - [What Is Data Residency?](https://www.dawiso.com/glossary/data-residency): Data residency is the physical or geographic location where data is stored and processed. Explains that it is the narrowest of three related ideas: residency is about location (where), data sovereignty is about jurisdiction (whose law), and data localization is a legal requirement to keep data in-country. Covers why residency became mainstream (cloud made data location invisible; GDPR made it a compliance question), why it matters (compliance, customer trust, latency and continuity), and the critical trap that EU residency does not equal EU sovereignty, since a US-headquartered provider can store data in Frankfurt and still be reachable under the US CLOUD Act. To control residency you need to inventory every data store, classify by sensitivity, choose deployment you control, and separate residency from sovereignty. Dawiso provides the governed catalog and classification plus private-cloud or on-premise deployment so an EU residency choice also holds as an EU sovereignty choice. CTA to Enterprise Deployment. - [What Is a Sovereign Cloud?](https://www.dawiso.com/glossary/sovereign-cloud): A sovereign cloud is cloud infrastructure that guarantees the data and operations it hosts stay under a chosen jurisdiction's control, out of reach of foreign law. Explains that sovereignty depends on who owns the operator, who has administrative access, and which laws they obey, not just the data center address. Describes sovereignty as a ladder of three levels: data sovereignty (data stored and processed in-region under local law, the baseline most "EU region" offerings claim), operational sovereignty (only local, vetted staff operate it; admin control stays in-jurisdiction), and technical sovereignty (no dependence on foreign-controlled technology or supply chain, the hardest level). Covers why it matters now (the US CLOUD Act's reach and GDPR conflict; the EU's 2026 Cloud Sovereignty Framework grading providers up to SEAL-4 and a 180 million euro sovereign-cloud procurement; EuroStack; buyer demand), sovereign vs public cloud trade-offs, and the realistic mixed model. Dawiso is European-owned and operated, deployable in a private cloud in your own EU tenant or on-premise, metadata-only and single-tenant. CTA to Enterprise Deployment. - [What Is Single-Tenant Architecture?](https://www.dawiso.com/glossary/single-tenant): Single-tenant architecture gives each customer its own dedicated, isolated application instance and data store, while multi-tenant serves many customers from one shared application and database separated only by logical partitioning. Covers the defining trait (isolation of the data and application boundary per customer, whether on dedicated servers, a private cloud tenant, or on-premise), the key differences (data isolation, blast radius, compliance evidencing, location and control), why it matters for sensitive data (structural isolation removes cross-tenant risk; cleaner GDPR/ISO 27001/SOC 2 mapping; residency and sovereignty control), and the trade-offs (higher per-customer overhead and multi-instance upgrades versus multi-tenant efficiency). Framed as fit, not superiority: multi-tenant for high-scale lower-sensitivity apps, single-tenant for regulated or sovereignty-bound data. Dawiso uses a separate metadata store per customer rather than a shared multitenant database. CTA to Enterprise Deployment. - [What Is Data Localization?](https://www.dawiso.com/glossary/data-localization): Data localization is a legal requirement that certain data must be stored, and sometimes processed, within a specific country's or region's borders. Explains the strictness spectrum (storage-only, local processing, no copies or transfers abroad), the family relationship where localization is the requirement that forces a residency outcome while sovereignty is about whose law governs, and that keeping data in-country does not automatically place it beyond foreign reach if the operator is foreign-controlled (the US CLOUD Act). Covers where it applies (public-sector and government data, health records, financial and payment data, telecom and critical-infrastructure data; the GDPR restricts transfers but does not impose blanket localization, while member-state and national rules vary widely), and how to comply (know what you hold and where, classify by category, trace flows, deploy within the border). Dawiso's catalog, classification, lineage, and private-cloud or on-premise deployment make localization provable. CTA to Data Catalog. - [What Is the US CLOUD Act?](https://www.dawiso.com/glossary/us-cloud-act): The US CLOUD Act (Clarifying Lawful Overseas Use of Data Act, 2018) lets US authorities compel US-based providers to hand over data in their possession, custody, or control regardless of where it is physically stored, including EU data centers. Explains that it amended the 1986 Stored Communications Act to remove doubt about reaching data stored abroad, that it applies to providers subject to US jurisdiction (US companies and, depending on control, their subsidiaries) with control (not location) as the trigger, and the mechanism (a US authority serves a lawful order; the provider must produce the data whether it sits in a US or EU region; it also enables executive agreements between countries). Details the conflict with GDPR Article 48 (a third-country order does not by itself make a transfer lawful; the EDPB's restrictive view leaves a conflict of laws no contract fully resolves), what it means for EU organizations (jurisdiction not geography decides who can reach your data; a catalog is a sensitive map), and why residency alone does not solve it. Dawiso is European-owned and operated, deployable in your own EU tenant or on-premise, metadata-only and single-tenant per customer. CTA to Data Catalog. - [What Is a Data Dictionary?](https://www.dawiso.com/glossary/data-dictionary): A data dictionary is a centralized technical reference for an organization's data elements --- column names, data types, formats, allowed values, constraints, keys, and relationships. Covers what an entry contains; the key distinction from a business glossary (plain-language business terms) and a data catalog (inventory of all assets that links the other two); active dictionaries (DBMS-managed, always current) vs passive ones (standalone documents that drift out of date); and why the function has migrated into automated metadata management and data catalogs that keep technical metadata current and connected to business meaning. - [What Is Data Literacy?](https://www.dawiso.com/glossary/data-literacy): Data literacy is the ability to read, work with, analyze, and communicate with data --- a baseline competency for everyone, not a specialist skill. Covers the four progressive core competencies; why literacy unlocks the value of every other data investment and enables real data democratization; the barriers (inconsistent definitions, undiscoverable data, fear/culture, tool sprawl --- mostly governance problems, not training gaps); and how a business glossary, data catalog, and visible lineage build literacy by giving everyone a shared, trustworthy vocabulary to be literate in. - [What Are the FAIR Data Principles?](https://www.dawiso.com/glossary/fair-data-principles): The FAIR principles (from a 2016 Scientific Data paper and the GO FAIR initiative) make data Findable, Accessible, Interoperable, and Reusable --- for machines as much as people. Covers each principle and its metadata basis (persistent identifiers, standard protocols, shared vocabularies, licence & provenance); the crucial distinction from open data ("as open as possible, as closed as necessary" --- data can be FAIR but access-controlled); FAIR as a blueprint for AI-ready data; and how metadata management plus a data catalog, glossary, and lineage operationalize all four principles as a maintained, governed state. - [What Is Augmented Analytics?](https://www.dawiso.com/glossary/augmented-analytics): Augmented analytics (term popularized by Gartner ~2017) uses AI, ML, and NLP to automate the analytics workflow --- automated data preparation, automated insight and anomaly discovery, and natural-language query and generation. Covers how it works, the contrast with descriptive/manual traditional BI, its place on the path to conversational/agentic analytics (Databricks Genie, Snowflake Cortex Analyst), the benefits (speed, self-service, freed analysts) and the under-appreciated risks (confidently wrong answers, spurious correlations, false confidence, access blind spots), and why --- because a machine rather than a trained analyst interprets the data --- accuracy depends entirely on a governed semantic layer, business glossary, catalog, and data quality. Augmented analytics amplifies whatever data quality it is fed. - [What Is Schema Drift?](https://www.dawiso.com/glossary/schema-drift): Schema drift is the unexpected, uncontrolled change of a dataset's structure over time --- columns added, dropped, renamed, or retyped --- that breaks or silently corrupts downstream pipelines, reports, and models. Covers its causes (source-system changes, third-party API evolution, uncoordinated upstream edits, format/type changes), the crucial distinction from data drift (where structure is intact but values/distribution shift, degrading ML models), the two costs (loud breakage vs dangerous silent corruption), and the defences --- data contracts to agree the schema, data observability to detect changes, column-level lineage to see what a change will break before shipping it, and automated tests (dbt) --- all resting on governed metadata, a documented catalog, and interactive lineage. - [What Is Risk Assessment in Data Management?](https://www.dawiso.com/glossary/risk-assessment): Risk assessment is the structured process of identifying, analyzing, and prioritizing the risks attached to data so limited budget targets the risks that matter most. Covers the formal definition (event × likelihood × impact); the distinction between assessment, risk management, and risk categorization; the five-step loop (identify, analyze, evaluate, treat, monitor) and the living risk register it produces; the types of data risk (privacy, security, quality, compliance, operational); the likelihood × impact risk matrix that forces prioritization into green/amber/red zones; and how a governed catalog, classification, and interactive lineage make honest assessment possible at scale --- you cannot score the risk of data you cannot see, and lineage turns a guessed impact into an evidenced blast radius. - [What Is Risk Categorization?](https://www.dawiso.com/glossary/risk-categorization): Risk categorization sorts risks and the data carrying them into consistent groups by type and severity so each group can be assessed, owned, and controlled the same way. Covers the two questions it answers (what kind, how serious); its place between data classification and risk assessment; categorizing risks (privacy, security, quality, compliance, operational) versus categorizing data by sensitivity tier; the typical four-tier schemes (Public→Internal→Confidential→Restricted; Low→Medium→High→Critical) each mapped to a fixed control set; why it converts policy into automation (controls by rule, consistent prioritization, audit-readiness); the failure mode of partial coverage; and how a governed catalog makes every asset carry its category and lineage so the scheme becomes enforced and queryable. - [What Is AI-Based Data Risk Assessment?](https://www.dawiso.com/glossary/ai-based-data-risk-assessment): AI-based data risk assessment uses machine learning to discover, classify, and score data risks automatically and continuously rather than through periodic manual review. Covers how it inverts the traditional process (automated, human-supported) and is active metadata applied to risk; manual vs AI-based across cadence, coverage, latency, consistency, and cost; the continuous scan → classify → score → detect loop with a human-in-the-loop for treatment and accountability; AI-powered risk detection (exposed sensitive data, anomalous access/movement, quality and structure decay overlapping with observability); and why it must sit on a governed catalog with classification and lineage so every detected risk arrives with its meaning, owner, and downstream blast radius --- connecting to AI governance and the EU AI Act. - [What Is AI-Powered Compliance Automation?](https://www.dawiso.com/glossary/ai-powered-compliance-automation): AI-powered compliance automation uses AI to perform the continuous, high-volume work of regulatory compliance --- mapping data to obligations, monitoring for violations, and generating audit evidence --- turning compliance into a continuously verified state rather than a periodic scramble. Covers what AI adds over rule-based tooling (scale, content understanding of unlabeled/unstructured PII, speed, automatic evidence generation, mapping assets to specific articles); the workflow between regulations, the data estate, and an AI engine producing an always-ready audit trail; the hard limits (ambiguous legal language, model errors, non-delegable accountability --- regulators hold people not software); the assistance-not-autopilot design with human sign-off; and how a governed catalog, classification, lineage, and AI governance make the evidence simply already there. - [CCPA Compliance for Data Teams](https://www.dawiso.com/glossary/ccpa-compliance-for-data-teams): A practical guide to the California Consumer Privacy Act (as expanded by the CPRA) for data teams. Covers what the CCPA is and its broad definition of personal and sensitive information; how it compares to GDPR (CCPA centres on the sale and sharing of data and the right to opt out); who it applies to (for-profit businesses meeting revenue ~$25M, volume 100,000+ consumers/households, or 50%+ data-driven-revenue thresholds, reaching any business handling Californians' data); the five consumer rights (know, delete, correct, opt out of sale/sharing, limit sensitive PII use) plus non-discrimination and penalties; the four data-team capabilities every right reduces to (know what you hold via a catalog, know which data is personal via classification, know where it flowed via lineage, and act on it via workflows); the recurring failure of the forgotten copy; and how a governed catalog makes CCPA readiness a by-product of good governance. - [What Is Cloud Computing?](https://www.dawiso.com/glossary/cloud-computing): Cloud computing is the on-demand delivery of computing resources over the internet, billed by usage rather than owned outright --- the foundation of modern data platforms. Covers the NIST five characteristics (on-demand self-service, broad network access, resource pooling, rapid elasticity, measured service); the CapEx→OpEx shift; the three service models (IaaS rent infrastructure, PaaS rent a managed platform, SaaS rent finished software) and the shared-responsibility line that moves up while data responsibility stays yours; the deployment models (public, private, hybrid/multi-cloud) often driven by data sovereignty; why it matters for data (sprawl, cost surprises, visibility gaps); and how a governed catalog connecting 40+ platforms keeps cloud-scattered data discoverable, compliant, and traceable across cloud boundaries. - [What Is Cloud Migration?](https://www.dawiso.com/glossary/cloud-migration): Cloud migration moves data, applications, and workloads to the cloud --- one of the highest-stakes projects a data team runs because it touches every system at once. Covers why the difficulty is the web of dependencies (pipelines, reports, legacy logic) not the data movement; the 6 Rs strategy framework (rehost, replatform, refactor, repurchase, retire, retain); the four phases (assess and map dependencies with lineage, plan and sequence into waves, migrate and rebuild, validate parity against the source); the two great failure modes (moving data without knowing what depends on it, so downstream breaks silently; and migrating the undocumented mess); and how a governed catalog plus interactive lineage turn a risky big-bang into a mapped, staged, verifiable project. - [What Is Cloud-Native Data Management?](https://www.dawiso.com/glossary/cloud-native-data-management): Cloud-native data management uses services built specifically for the cloud --- elastic, managed, consumption-priced, decoupled --- rather than running legacy databases on cloud servers. Covers the distinction between "in the cloud" and "cloud-native"; the core principles (separation of storage and compute as the foundational idea, managed services, elasticity, consumption pricing, open table formats like Iceberg and Delta); the layered modern stack (lakehouse storage, ELT, orchestration, BI/AI consumption, cross-cutting governance); the governance gap where elastic self-service generates sprawl and cost faster than static governance can track; and why a cloud-native estate needs an equally automated, elastic catalog and lineage layer that discovers data rather than waiting to be told about it. - [What Is an Analytical Pipeline?](https://www.dawiso.com/glossary/analytical-pipeline): An analytical pipeline is the end-to-end sequence that moves raw data through ingestion, storage, transformation, and modeling into analysis-ready datasets that power dashboards, reports, and AI. Covers how it differs from a generic data pipeline (purpose: prepare data for analysis, not just move it); the modern ELT pattern (load raw first, transform in the lakehouse); the five stages (ingest, store, transform, model, serve) and why transform/model concentrate the value and the risk; the batch vs streaming timing choice; and why every trusted metric inherits the correctness of every upstream stage --- making end-to-end lineage from source column to served metric the single most valuable thing you can have over a pipeline. - [What Is a Data Mart?](https://www.dawiso.com/glossary/data-mart): A data mart is a subject-focused subset of a data warehouse optimized for a single team or domain (sales, finance, marketing). Covers its dimensional design and place as the consumption layer; data mart vs data warehouse (scope, size, audience, design); the three types (dependent marts sourced from the governed warehouse, independent silos built straight from sources, and hybrid) and how the choice is really a governance choice; the benefits (speed, simplicity, autonomy) and the classic trade-off of metric fragmentation across proliferating marts; and how a data catalog and business glossary keep definitions consistent across every mart, with lineage tracing which mart and source a number came from. - [What Is a Data Ecosystem?](https://www.dawiso.com/glossary/data-ecosystem): A data ecosystem is the entire interconnected environment --- technologies, data, people, and processes --- that an organization uses to collect, store, process, govern, and consume data. Covers why the connections between parts matter more than any single component; the layers (sources, ingestion, storage, processing, consumption) plus the cross-cutting governance and people/process bands; the distinction between a data ecosystem, a data stack (just tools), and a data architecture (just the design); the difference between a healthy connected ecosystem and a fragmented set of silos; and how a catalog plus lineage form the connective tissue that turns scattered components into a coherent, discoverable whole. - [What Is a Metadata Lakehouse?](https://www.dawiso.com/glossary/metadata-lakehouse): A metadata lakehouse applies data lakehouse principles --- unified open storage, scalable compute, queryability --- to metadata rather than data, consolidating technical, business, operational, and usage metadata into one open, connected, queryable store. Covers the parallel to the data lakehouse; why it emerged (metadata sprawl across siloed tools, plus AI's need for connected machine-readable context); how it works (ingest all metadata, unify and connect it as a graph, query and activate capabilities on top); what it enables (active metadata, cross-cutting impact questions, AI context for GraphRAG and knowledge graphs, governance at scale); and how a modern data catalog built on a unified metadata model --- with lineage, AI enrichment, and a Context Layer/MCP --- is the practical realisation of the idea. - [Semantic Layer vs Traditional Data Marts](https://www.dawiso.com/glossary/semantic-layer-vs-data-marts): Both approaches give business users consistent, analysis-ready data, but in different places. Covers the shared problem (raw warehouse tables aren't business-ready, and ad-hoc queries produce inconsistent metrics); the traditional data mart approach (close the gap physically with per-team curated subsets --- fast and autonomous but each bakes in its own definitions, producing conflicting numbers); the semantic layer approach (close the gap logically and centrally --- define each metric once for all tools and AI agents, so revenue means the same thing everywhere); a head-to-head across where logic lives, consistency, duplication, autonomy vs governance, and AI-readiness; how the two can coexist (marts beneath a semantic layer); and why governed definitions in a business glossary, connected to physical tables by lineage, are what end metric fragmentation. - [What Is a Metrics Layer?](https://www.dawiso.com/glossary/metrics-layer): What a metrics layer (metrics store) is, the part of the stack that defines how each business metric is calculated once so every tool and AI agent returns the same number. Covers the four parts of a metric definition (formula, aggregation, dimensions, filters), how it powers headless analytics, why it ends metric drift, and the key distinction that a metrics layer is a component of a broader semantic layer (which also covers meaning, joins, grain, permissions, and lineage), not a replacement for it. Explains why consistent metrics are necessary but not sufficient for trustworthy AI, and how Dawiso adds governed business meaning, lineage, and classification around metrics, served to any agent via MCP. - [What Is Text-to-SQL?](https://www.dawiso.com/glossary/text-to-sql): What text-to-SQL (NL2SQL) is, converting plain-language questions into executable SQL with an LLM plus schema-aware parsing. Covers how it works (understand the question, read the schema, generate SQL, validate and execute), the benefits (accessibility, speed, valid syntax), and the central finding that accuracy depends on context rather than the model's SQL ability: frontier models score around 85-90% on clean academic benchmarks (original Spider) but collapse to about 20% on enterprise-scale schemas (Spider 2.0), the enterprise cliff, and recover to roughly 85-95% with a governed semantic layer and business context. Covers the enterprise failure modes (complex multi-table queries, wrong table or metric definition, missing access checks) and how Dawiso supplies the governed business glossary, catalog, lineage, and classification that let text-to-SQL resolve to the right, trustworthy query, served via MCP. - [What Is COGS (Cost of Goods Sold)?](https://www.dawiso.com/glossary/cogs-cost-of-goods-sold): COGS is the total direct cost of producing or acquiring the goods a company sold in a period --- materials, direct labour, and production costs --- shown below revenue on the income statement, with Revenue − COGS = Gross Profit. Covers what counts as direct vs indirect cost; that COGS matches cost to goods actually sold (not produced); the formula (Beginning Inventory + Purchases − Ending Inventory) and how costing methods (FIFO/LIFO/weighted average) change it; why it matters (gross margin, pricing, profitability analysis, tax and audit); and the data problem behind it --- inputs spread across inventory, procurement, manufacturing, and ledger systems that need consistent definitions (a business glossary), input quality, and traceable lineage --- making COGS a worked example of why financial reporting is a data governance problem. - [What Is a Context Store for AI?](https://www.dawiso.com/glossary/context-store-for-ai): A context store for AI holds and serves the governed business context AI needs to ground its answers --- definitions, relationships, lineage, ownership, and policy --- rather than guessing from raw schemas. Covers how it is a governed (curated, owned, versioned, access-controlled) store of meaning; what it stores; the distinction from a vector store (embeddings for similarity search) and a feature store (ML features), which it complements; why AI needs it (a confident wrong answer is worse than none, and Gartner-cited GenAI failures trace to missing governed context); the role of MCP in delivering it; and how a governed catalog, glossary, and lineage are, in effect, the context store, packaged by the Context Layer and served via the MCP Server. - [What Is a Metadata Layer for AI?](https://www.dawiso.com/glossary/metadata-layer-for-ai): The governed layer of metadata --- technical (schemas, types), business (definitions, ownership), operational (freshness, quality, lineage) --- that sits between data platforms and AI systems and supplies the meaning AI needs to interpret enterprise data. Covers the distinction between "in the cloud" metadata and a deliberate consolidated layer; what it contains; the passive-vs-active distinction (for AI it must be active: queryable, machine-readable, callable via MCP, not locked in a UI); why it makes data AI-ready and prevents confident hallucination; and how Dawiso harvests metadata from 40+ platforms, makes it active with AI enrichment and lineage, and exposes it to agents through its MCP Server. - [What Is a Metadata Knowledge Graph?](https://www.dawiso.com/glossary/metadata-knowledge-graph): A knowledge graph whose nodes are metadata entities (datasets, columns, business terms, owners, policies, reports, pipelines, models) and whose edges are typed relationships (implements, derives from, owned by, governed by, feeds). Covers why relationship questions (impact analysis, "which term does this metric use?", "what sensitive data feeds this model?") are graph traversals that flat catalogs answer poorly; nodes and edges; why a graph beats relational tables for multi-hop questions (lineage is the graph filtered to flow edges); what it powers (impact analysis, semantic discovery, GraphRAG grounding, governance propagation); and how a connected catalog with AI-proposed relationships is a metadata knowledge graph. - [What Is a Context Graph?](https://www.dawiso.com/glossary/context-graph): A connected model of the business context around data --- concepts, definitions, data, processes, people, and their relationships --- assembled so an AI agent can traverse it to understand and reason about a domain. Covers how it is concept-centred (a business idea surrounded by its definition, related concepts, implementing data, lineage, policy, and owners); how it compares to a knowledge graph (general structure) and a metadata knowledge graph (data-estate view), with the context graph as the AI-facing business-meaning view; what it connects; its central role in grounding GraphRAG and agents (traversing relationships yields connected, governed context rather than isolated snippets); and how Dawiso builds it from the glossary, catalog, lineage, and AI enrichment and serves it via MCP. - [What Is Context Rot?](https://www.dawiso.com/glossary/context-rot): The measurable decline in LLM output quality as input length grows, which begins well before the context window limit (distinct from window overflow). Covers Chroma's 2025 research (all 18 frontier models tested degraded with longer inputs, even on simple retrieval); why a large window is a ceiling, not a usable working volume; the causes (attention dilution, accumulated failure modes, position effects, noise-to-signal drift); why the fix is less, better context rather than a bigger window; and how Dawiso serves tight, relevant, governed single-source-of-truth context via MCP to keep the working context small and trustworthy. - [What Is Context Poisoning?](https://www.dawiso.com/glossary/context-poisoning): When a false or misleading fact enters an AI's context and is then treated as true for the rest of the interaction, compounding. Covers the definition and why it persists and compounds; that the model cannot tell a poisoned fact from a real one (only the source can); its place among the four context-failure modes (with confusion, clash, distraction) per Drew Breunig's 2025 taxonomy; how it happens (self-poisoning by hallucination, bad retrieval, stale data, prompt injection); prevention by grounding in trusted data with provenance; and how Dawiso's governed, traceable context layer served via MCP starves it. - [What Is Context Clash?](https://www.dawiso.com/glossary/context-clash): When two or more pieces of an AI's context contradict each other and the model reconciles them badly, producing inconsistent answers. Covers the factual, definitional, and procedural forms; why its deepest enterprise cause is the absence of a single agreed definition (the semantic gap in AI form); how it happens (conflicting definitions, disagreeing data, stacked instructions, merged retrieval); prevention by defining each term and metric once authoritatively and reconciling data at the source; and how Dawiso's governed glossary and context layer serve one definition per concept to every agent via MCP so context can't contradict itself. - [What Is Context Confusion?](https://www.dawiso.com/glossary/context-confusion): When irrelevant or superfluous information in the context window influences an AI's answer, because models use everything they are given rather than filtering for relevance. Covers why it disproves the "more context is better" assumption; how it differs from poisoning (false) and clash (contradictory); how it happens (over-retrieval, tool overload, kitchen-sink prompting, unpruned history); the fix of deliberately engineered relevance (precise retrieval, scoped tools, governed structure, aggressive pruning); and how Dawiso lets agents pull the right governed slice by meaning via MCP rather than dumping everything. - [What Is Context Assembly?](https://www.dawiso.com/glossary/context-assembly): The step in an AI pipeline where the information that goes into the context window is gathered, selected, ordered, and formatted before a model call - the operational core of context engineering. Covers the retrieve/select/order/format pipeline; why it is the highest-leverage point for output quality (ahead of model choice or prompt wording); what makes it hard (knowing relevance, avoiding contradiction, trust and provenance, staying tight); why its limiting factor is the quality and governance of the sources, not the assembling logic; and how Dawiso gives assembly a single governed source of truth to draw from, served via MCP. - [What Is Context Management?](https://www.dawiso.com/glossary/context-management): How an AI agent curates its context window over the course of a task - what to add, keep, compress, drop, and persist - because windows are finite and more context causes rot. Covers the four core strategies (write to external memory, select what's relevant back, compress history, isolate across sub-agents); why it separates demo agents from production ones; the governance dimension (the keep-or-drop decision depends on knowing what is trusted and current); and how Dawiso acts as the governed source agents manage against, recalling trusted context on demand via MCP instead of hoarding it in the window. - [What Are Context Packs?](https://www.dawiso.com/glossary/context-packs): Curated, reusable bundles of the context an agent needs for a domain or task (definitions, reference data, tools, guidance), packaged once and reused across interactions - framed honestly as an emerging pattern, not a settled standard. Covers what's typically inside a pack; why the pattern is emerging (reduced repeat work, consistency, portability, context as a managed asset); why governance decides their value (a pack multiplies whatever you put in it, so an ungoverned pack standardizes bad answers); and how Dawiso provides the governed source of truth from which trustworthy, reusable packs are built and served via MCP. - [What Is Operational State in AI Agents?](https://www.dawiso.com/glossary/operational-state): An agent's live, high-velocity, ephemeral runtime context - the current values, IDs, and conditions of the task at hand (e.g. "order 4471: pending") - as opposed to the stable governed knowledge (definitions, rules, relationships) it reasons with. Covers operational state vs decision context as two complementary layers; why agents need both (state supplies the facts of the moment, governed knowledge supplies the meaning); the governance challenge (state must be fresh and trusted, and needs governed definitions to be interpretable); and how Dawiso supplies the stable governed meaning that live operational state is read against, served via MCP. - [Context Layer vs Semantic Layer](https://www.dawiso.com/glossary/context-layer-vs-semantic-layer): A semantic layer translates technical data into consistent business definitions and metrics (what data means); a context layer includes the semantic layer and adds lineage (provenance), governance (trust, access, policy), relationships, and documentation --- everything an AI agent needs to use data correctly and safely. Covers both layers defined; the core difference by what each answers about "monthly revenue"; a head-to-head across scope, primary question, primary consumer, trust/provenance, governance, and AI delivery; why AI (which supplies no judgment of its own) needs the context layer not just the semantic layer; and how Dawiso treats the glossary-based semantic layer as the meaning core, wraps it with lineage and governance, and serves it via MCP. - [What Is the Sovereign Context Protocol (SCP)?](https://www.dawiso.com/glossary/sovereign-context-protocol-scp): An emerging concept (explicitly not yet a ratified standard) for delivering governed context to AI while keeping the underlying data inside sovereign, in-jurisdiction, policy-controlled boundaries. Covers the honest caveat about the term; the separation of context (which can travel under control) from raw sensitive data (which stays put); the three-way clash it resolves (use AI on proprietary data, obey data sovereignty, don't leak sensitive data); how it builds on the real MCP standard as a deployment-and-governance discipline (run the context server in your boundary, enforce access and classification there, audit every request); the sovereignty requirements; and how Dawiso meets them with a governed Context Layer served via its MCP Server plus flexible enterprise deployment. - [Context Layer for Snowflake & Horizon Catalog](https://www.dawiso.com/glossary/context-layer-for-snowflake): Giving Snowflake data the governed business meaning AI needs. Covers what a context layer means inside Snowflake; Snowflake Horizon Catalog as a strong native context/governance layer (automated classification and PII/PHI protection, end-to-end lineage, data-quality monitoring, AI and agent governance enforced at the query engine so it applies to humans/BI/AI alike, and Catalog-Linked Databases syncing governance with Apache Iceberg) --- which Snowflake itself positions as a "governed context layer for AI"; how Horizon pairs with Cortex and semantic views to ground Snowflake-native AI; the gap (most enterprises also have data outside Snowflake, and agents need open MCP delivery); and how Dawiso provides a cross-platform context layer that complements Horizon and serves any agent via MCP. - [Context Layer for Databricks](https://www.dawiso.com/glossary/context-layer-for-databricks): Giving Databricks lakehouse data the governed business meaning AI needs. Covers what a context layer means inside Databricks; Unity Catalog as a strong native foundation (unified governance for data and AI, access control, automated column-level lineage plus external lineage in preview to upstream sources and BI, discovery, quality, sharing, auditing --- open-sourced under the Linux Foundation in 2024 and integrating with Spark, Trino, Iceberg, DuckDB); why natural-language AI like Databricks AI/BI Genie needs curated business context; the gap (curated business meaning still has to be layered on, the estate spans more than Databricks, and agents need open MCP delivery); and how Dawiso adds the cross-platform business-aware context layer that complements Unity Catalog and serves any agent (including Genie) via MCP. - [Context Layer for dbt](https://www.dawiso.com/glossary/context-layer-for-dbt): Turning dbt's rich metadata into governed business context AI can use. Covers why dbt is a prime context source (it forces meaning to be written down --- models encode how concepts are computed, tests encode quality expectations, descriptions encode documentation, the ref() graph encodes lineage, all compiled into the machine-readable manifest); the dbt Semantic Layer as a classic semantic layer defining metrics once, which a context layer takes as its meaning core and surrounds with lineage, trust, relationships, and policy; the gap (dbt's metadata is engineer-facing and project-bound, needs governance and business framing, and needs open MCP delivery to reach AI); and how Dawiso ingests dbt models/tests/descriptions/lineage into a cross-platform catalog, elevates descriptions into governed glossary definitions, and serves it all to agents via MCP. - [What Is the Semantic Gap?](https://www.dawiso.com/glossary/semantic-gap): The semantic gap is the distance between how data is stored (technical tables, columns, codes) and what it means in business terms. Covers its origin in CS/AI (low-level representation vs high-level concept) and its concrete data form (physical vs conceptual); where it shows up (inconsistent metrics, misused data, slow discovery, eroded trust --- all symptoms of one underlying gap); why it is acute for AI, which has no implicit knowledge and so fills the gap with confident guesses; how it is closed by making meaning explicit and attached to the data (business glossary, semantic layer, metadata and lineage, ontology); and how Dawiso closes it with a glossary linked to physical data, lineage for provenance, AI enrichment at scale, and a Context Layer that serves governed meaning to AI via MCP. - [What Is an Ontology in AI?](https://www.dawiso.com/glossary/ontology-in-ai): An ontology is a formal, explicit, machine-readable model of a domain's concepts (classes), properties, relationships, and rules --- "an explicit specification of a conceptualization" --- that lets AI reason and infer rather than merely recall. Covers what it specifies and its capacity for inference; the distinction from a taxonomy (hierarchy/is-a only, a tree) versus an ontology (arbitrary typed relationships and rules, a graph); its three roles in modern AI (domain grounding, structure for GraphRAG retrieval, consistency/validation) and the neuro-symbolic trend; the tight relationship to knowledge graphs (ontology is the schema, knowledge graph is the data/instances); and how Dawiso captures ontology-grade structure from a governed glossary and connected catalog, with AI-assisted relationship enrichment, served to agents via the Context Layer and MCP. - [Ontology vs Semantic Layer](https://www.dawiso.com/glossary/ontology-vs-semantic-layer): Both add a layer of meaning over data but solve different jobs. A semantic layer defines business metrics and dimensions mapped to physical data for consistent measurement ("what is the agreed number?"); an ontology models concepts, typed relationships, and rules for reasoning and inference ("how does the domain connect?"). Covers both defined; the core difference by what each lets you ask (measures vs relationships); a head-to-head across core unit, purpose, primary question, shape, AI use, and formality; why traditional BI may need only a semantic layer but agentic AI and natural-language interfaces need ontology-style relationship traversal too; and how Dawiso builds both from one governed business glossary and catalog --- consistent measures and connected relationships --- and serves them to AI via MCP. - [OpenAI Frontier: The Data Governance Problem](https://www.dawiso.com/glossary/openai-frontier-data-governance): OpenAI Frontier, launched February 2026, is OpenAI's enterprise platform for building, deploying, and managing AI agents ("AI coworkers") that act across an organization's data and systems --- with per-agent identity, permissions, and guardrails, and open to third-party agents (Google, Microsoft, Anthropic). Covers what Frontier is and why it matters; the data governance problem (it governs who an agent is and what it may do, not what the data means, is it trustworthy, how it's defined, or where it came from --- so a permissioned agent acts confidently on misunderstandings, and because it acts, the blast radius is a wrong action propagated at machine speed); the distinction between access governance (Frontier) and context governance (a context layer); why Frontier readiness equals data governance readiness (catalog, glossary, classification, lineage); and how Dawiso supplies the governed context Frontier assumes and serves it to any agent via the Context Layer and MCP, paired with AI governance. - [What Is Overall Equipment Effectiveness (OEE)?](https://www.dawiso.com/glossary/overall-equipment-effectiveness-oee): OEE measures how much of planned production time is truly productive, as the product of three factors --- Availability x Performance x Quality --- defined by Seiichi Nakajima within Total Productive Maintenance (Introduction to TPM, 1988). Covers the multiplicative structure that punishes the weakest factor, the three factors and their loss families, Nakajima's six big losses, the 85% world-class benchmark versus a typical 40-60% plant, and the core governance point: OEE looks like an objective percentage but every term (planned production time, ideal cycle time, a "good" unit, the unplanned-stop threshold) rests on a local definition, so raw cross-plant comparisons compare rulers, not performance. How Dawiso governs one shared definition of OEE and every term in it via the business glossary, with lineage for provenance and a catalog of the feeding MES/historian data. - [What Is Industry 4.0 Data Governance?](https://www.dawiso.com/glossary/industry-4-0-data-governance): Applying data ownership, quality, shared meaning, and access control to the operational-technology data of connected manufacturing --- sensor and IoT telemetry, PLC/SCADA tags, process historian series, MES records, PLM/ERP engineering data, and the OEE calculations, digital twins, and predictive-maintenance models built on them. Covers the four questions governance answers for factory data, why it is harder than classic IT governance (volume and velocity, the OT/IT ownership split, semantic chaos where the same measurement is named differently at every plant, quality that degrades as sensors drift, and safety-adjacent security stakes), what good looks like (a single catalog, a shared glossary/ontology, lineage, clear ownership), and the honest scope boundary: governance sits beside the MES and historian, it does not replace them. How Dawiso provides the catalog, glossary, lineage, and ownership layer. - [What Is Audit Trail Review?](https://www.dawiso.com/glossary/audit-trail-review): The periodic, documented, risk-based examination of a GxP audit trail --- the secure, time-stamped log of who created, changed, or deleted regulated data, when, and why --- by a qualified, independent reviewer. Covers what makes it a defined quality activity (periodic, risk-based, independent, documented), what a well-formed audit trail entry captures and the three properties that make it trustworthy (automatic, secure, independent), the regulatory basis (FDA 21 CFR Part 11, EU GMP Annex 11's risk-based review mandate, the MHRA 2018 data integrity guidance and ALCOA+, the ISPE GAMP records and data integrity guide), how review works in practice (which trails matter, frequency by risk, what reviewers look for, routing findings to CAPA), and the volume problem that pushes mature programs to exception-based review. Honest scope: the review is performed inside validated GxP systems; Dawiso supplies the surrounding traceability (lineage), a catalog of GxP-relevant systems, and consistent definitions. - [What Is IDMP (Identification of Medicinal Products)?](https://www.dawiso.com/glossary/identification-of-medicinal-products-idmp): A suite of five ISO standards --- ISO 11238 (substances), 11239 (dose forms, routes, packaging), 11240 (units of measurement), 11615 (medicinal product information), and 11616 (pharmaceutical product information) --- that define a common way to identify and describe medicines so information can be exchanged reliably for pharmacovigilance, supply-chain tracking, and regulatory submissions. Covers what IDMP is and the safety need behind it, the layered model of the five standards, the EU mandate (Commission Implementing Regulation (EU) No 520/2012, articles 25 and 26) and the EMA's phased SPOR master data services (Substance, Product, Organisation, Referential), and the central argument that IDMP is a master data problem: one product is described differently across RIM, R&D/CMC, manufacturing ERP/MES, pharmacovigilance, and QMS systems, and IDMP forces one reconciled description. Honest scope: Dawiso is not a regulatory submission tool; it governs the master data foundation (catalog, glossary, quality, ownership, lineage) that IDMP success depends on. - [What Is Bill of Materials (BOM) Master Data?](https://www.dawiso.com/glossary/bill-of-materials-bom-master-data): The bill of materials --- the structured, hierarchical list of components, subassemblies, materials, and quantities needed to make a product --- treated as governed, authoritative, version-controlled master data rather than a spreadsheet. Covers the BOM hierarchy (multi-level vs single-level, parent-child quantities and units), what treating it as master data adds (stable identity, revision and effectivity, ownership, quality), the crucial distinction between engineering BOM (eBOM, in PLM, as designed) and manufacturing BOM (mBOM, in ERP/MES, as built) plus service and sales views, why BOM data drifts (multi-system fragmentation, revision/effectivity confusion, inconsistent identifiers and units, unclear ownership) and how that converts into wrong purchases, wrong builds, wrong costs, and slow recalls, and how master data governance (authoritative source per attribute, controlled change, ownership, quality, traceability) fixes it. How Dawiso catalogs where BOM data lives, defines the terms, and traces it across PLM, ERP, and MES. - [What Is Knowledge Management?](https://www.dawiso.com/glossary/knowledge-management): The discipline of capturing, organizing, sharing, and applying what an organization knows, so the right knowledge reaches the right people and systems at the right time --- formalized in ISO 30401:2018. Covers the KM lifecycle (capture, organize, share, apply) and KM's recurring failure mode of unmaintained knowledge bases, the key distinction between tacit knowledge (experience and judgment, hard to codify) and explicit knowledge (documented, easy to share), the SECI model (Nonaka and Takeuchi's Socialization, Externalization, Combination, Internalization), and the decisive shift in the data and AI era: classic tooling (documents, wikis, intranets) stored explicit knowledge as prose that does not connect, but AI needs structured, machine-readable knowledge --- defined terms, explicit relationships, ownership, provenance --- which is exactly what a business glossary, catalog, and knowledge graph provide. Framed honestly as an evolution of decades-old disciplines pointed at machine consumers. How Dawiso covers the connected, governed end of knowledge management for the data domain. ## Events ### All Events **URL:** https://www.dawiso.com/events Upcoming Dawiso events, conferences, and webinars. Meet the Dawiso team, see data governance in practice, and stay ahead of the latest trends in data and AI. Covers conferences across Europe (Stockholm, Berlin, London, Warsaw, Prague, Budapest), partner events, and webinars on the Context Layer, AI governance, and BCBS 239. ### Meet Dawiso at AI & Big Data Expo Europe 2026 - Stand 260 **URL:** https://www.dawiso.com/events/ai-big-data-expo-europe-2026 Meet Dawiso at AI & Big Data Expo Europe 2026, part of TechEx Europe, at RAI Amsterdam, October 19-20, 2026. Over two days, 8,000+ attendees, 250+ expert speakers and 200+ exhibitors come together across eight co-located events covering AI and big data, physical AI, cyber security and cloud, IoT, edge computing, intelligent automation and robotics, digital transformation and data centres, with an audience of C-level and senior data and AI leaders from large enterprises. Find Dawiso at Stand 260 for a live demo of the AI context layer: scan your platforms to build the catalog, generate business context with AI (reviewed and approved by your team), and let any LLM or agent query your governed context through MCP. Book a meeting with Samuel Nagy or Michal Bambušek in advance. ### Meet Dawiso at Big Data & AI Paris 2026 - Stand J011 **URL:** https://www.dawiso.com/events/big-data-ai-paris-2026 Meet Dawiso at Big Data & AI Paris 2026, France's largest data and AI event, at Paris Expo Porte de Versailles, Hall 7.2, September 15-16, 2026. The organiser reports 15,000 participants, 200 exhibitors and sponsors, and 350 conferences and workshops. Find us at stand J011 for a live walkthrough of the Enterprise Context Layer for AI on a real data stack: scan 40+ platforms to build the catalog, generate business context with AI (reviewed and approved by your team), and connect any LLM or agent to your governed knowledge through MCP. Book a meeting with Samuel Nagy or Michal Bambušek in advance. ### Data Innovation Summit 2026 - Stand C32, Stockholm **URL:** https://www.dawiso.com/events/dis-2026 Attending DIS 2026? Book a focused 30-minute meeting with the Dawiso team at Stand C32. Live demo, real answers, no sales fluff. Stockholm, May 6-8, 2026. ABM landing page with team booking, capabilities preview, and event countdown. ### ScaleFree Webinar - Context Layer **URL:** https://www.dawiso.com/events/scalefree-webinar-context-layer-2026 Joint webinar with ScaleFree on May 12, 2026, exploring Dawiso's Enterprise Context Layer for trustworthy AI grounding. Live demo, Q&A with the product team, and customer perspectives on how the Context Layer fits into governed AI architectures. ## News & Press ### News Room **URL:** https://www.dawiso.com/news-room News Room index - Dawiso product releases, press releases, partnerships, and insights from the world of data governance, metadata management, and AI. Filterable by tag (Announcement, Product update, Achievement, Event, Data news). ### Dawiso Joins Databricks Marketplace as MCP Provider for Data Catalog **URL:** https://www.dawiso.com/news/dawiso-joins-databricks-marketplace-as-mcp-provider-for-data-catalog Dawiso is listed in the Databricks MCP Marketplace as a data catalog provider with write-back capability - while most integrations offer read-only access, Dawiso lets AI agents actively document, govern, and enrich metadata as they work. Two-Way integration covers Unity Catalog, Databricks DQ Monitoring, and the Databricks scanner. May 20, 2026 announcement with quote from CEO Samuel Nagy. ### Dawiso Launches Context Layer for Trustworthy Enterprise AI **URL:** https://www.dawiso.com/news/dawiso-launches-context-layer-to-help-data-leaders-build-trustworthy-ai Dawiso's new Enterprise Context Layer auto-generates business context and data lineage so enterprise AI systems deliver reliable, trustworthy answers. Announces availability and the architectural approach: combining catalog, glossary, lineage, and MCP into a single AI-ready context surface. ### Dawiso 2025 Achievements and 2026 Vision for AI-Powered Enterprise Data **URL:** https://www.dawiso.com/news/dawiso-announces-2025-achievements-and-2026-vision-for-ai-powered-enterprise-data Dawiso achieved major AI milestones in 2025 and now targets AI agent-driven metadata management and collaborative conceptual modeling for 2026. Year-end announcement covering product, customer, and partnership highlights and the 2026 product roadmap. ### Dawiso 2025.5 LTS: Smarter Automation and Easier Customization **URL:** https://www.dawiso.com/news/dawiso-2025-5-lts-update-delivers-smarter-automation-and-even-easier-customization Dawiso 2025.5 LTS expands automation capabilities, simplifies package customization, and refines permission handling for smoother data governance workflows. Long-term support release with focus on enterprise scale and customization. ### Dawiso Now Available in The Microsoft Azure Marketplace **URL:** https://www.dawiso.com/news/dawiso-now-available-in-the-microsoft-azure-marketplace-7t439 Dawiso is now on the Microsoft Azure Marketplace. Azure customers can purchase and deploy Dawiso directly through their existing Microsoft accounts, simplifying procurement and accelerating deployment for Azure-aligned enterprises. ### Dawiso Recognized in Gartner's 2024 Market Guide for Metadata Management **URL:** https://www.dawiso.com/news/dawiso-recognized-in-gartners-2024-market-guide-for-metadata-management-solutions Gartner's 2024 Market Guide for Metadata Management Solutions recognizes Dawiso for its user-friendly approach to data governance for mid-sized enterprises - a milestone in Dawiso's analyst recognition and market positioning. ### Dawiso Marketplace Launched **URL:** https://www.dawiso.com/news/dawiso-marketplace-launched Dawiso Marketplace brings one-click data governance setup with an updated trial experience, new features, and streamlined onboarding for new users. Marketplace introduces installable packages for industry templates, governance workflows, and metadata model extensions. ### New AI-Powered Features in Dawiso **URL:** https://www.dawiso.com/news/new-ai-powered-features-in-dawiso AI-powered features in Dawiso give organizations deeper insight into large-scale data ecosystems through automated metadata enrichment, intelligent search, and smart suggestions for data quality and documentation. ### The Dawiso 2025.1 Update: Smoother Workflows and Better Insights **URL:** https://www.dawiso.com/news/the-dawiso-2025-1-update-smoother-workflows-and-better-insights Dawiso 2025.1 ships revamped dashboards, advanced data lineage parsing, new integrations, and the AI Catalog for governing AI assets across your organization. ### Dawiso Update: Reflecting on 2024 and Looking Ahead to 2025 **URL:** https://www.dawiso.com/news/dawiso-update-reflecting-on-2024-and-looking-ahead-to-2025 Dawiso's 2024 year in review: new partnerships, security certifications, and industry milestones that set the stage for AI governance and data products in 2025. ### Keboola and Dawiso Partner on Data Lineage at Big Data LDN 2024 **URL:** https://www.dawiso.com/news/keboola-and-dawiso-partner-to-revolutionize-data-lineage-and-business-context-management-at-big-data-ldn-2024 Keboola and Dawiso present their joint data lineage and business context solution at Big Data London 2024, enabling end-to-end data traceability across Keboola pipelines and Dawiso governance. ### 55% of Companies Have Finance Executives Driving Data Governance **URL:** https://www.dawiso.com/news/in-55-of-companies-finance-executives-drive-data-governance-while-the-rest-are-unsure-how-to-start-or-are-unaware-of-it Finance executives lead data governance in 55% of companies, according to Dawiso research. The rest struggle to identify ownership or haven't started yet. Survey-based news on data governance ownership in mid-to-large enterprises. ### Can We Trust Big Data? Dawiso at Data Analytics Summit Berlin **URL:** https://www.dawiso.com/news/big-data-can-we-trust-how-it-is-managed-data-analytics-summit-berlin Dawiso, the main exhibitor sponsor, showcases its data analytics solutions at the Data Analytics Summit in Berlin on June 13-14, with talks on data trust and governance for AI-driven enterprises. ### Dawiso and VŠE: Expanding Partnership with New Student Programs **URL:** https://www.dawiso.com/news/dawiso-and-vse-expanding-our-long-standing-partnership-with-new-opportunities-for-students Dawiso and Prague University of Economics expand their partnership with new student programs in data governance, business intelligence, and analytics - bringing real-world tooling and case studies into the academic curriculum. ### Dawiso at the 9th Data Innovation Summit in Stockholm **URL:** https://www.dawiso.com/news/data-innovation-summit Dawiso presents its data catalog and governance platform at the 9th Data Innovation Summit in Stockholm, the Nordics' largest annual data and AI gathering. Coverage of the showcase, customer meetings, and conference takeaways. ### Dawiso Was Part of the Data Project Challenge 2024 **URL:** https://www.dawiso.com/news/dawiso-was-part-of-the-data-project-challenge-2024 Dawiso joined the Data Project Challenge 2024 in Prague, where students presented real-world data solutions with practical enterprise applications - extending Dawiso's commitment to academic partnerships and data education. ### FIS VŠE and Dawiso Teach Students Data Governance in Practice **URL:** https://www.dawiso.com/news/from-an-abstract-concept-to-real-world-application-fis-vse-and-dawiso-show-students-how-data-governance-works-in-practice Prague University of Economics partners with Dawiso to teach students hands-on data governance, bridging academic theory and real-world practice using the live Dawiso platform. ## Legal - Privacy Policy: https://www.dawiso.com/privacy-policy - Terms of Use: https://www.dawiso.com/terms-of-use ### EU Project - International Expansion **URL:** https://www.dawiso.com/eu-project Dawiso s.r.o. is a beneficiary of an OP TAK grant co-funded by the European Union through the European Regional Development Fund (ERDF). The page carries the funding disclosure the programme requires. ## Contact - Contact page: https://www.dawiso.com/contact-us - LinkedIn: https://www.linkedin.com/company/dawiso/ - YouTube: https://www.youtube.com/@dawisocom