Raiffeisenbank built a governed knowledge foundation for people and AI

The Client

Raiffeisenbank is one of the largest banks in the Czech Republic, serving retail, corporate, and business customers across a complex organizational and product landscape. Every day, employees across the bank rely on a large body of internal standards that define how it operates. These standards evolve over time, follow different approval processes, and become effective on specific dates, involving experts and decision-makers across the organization. As it expanded its use of AI, Raiffeisenbank saw a broader opportunity: a governed knowledge foundation that manages this knowledge across its whole lifecycle and makes the same trusted context available to employees, applications, and AI.

​

The Problem

These standards are far more than files. They live as thousands of interconnected documents, and on SharePoint, managing them turned out to be far more complex than storing and searching content.

The bank wanted to bring the full lifecycle of these standards under stronger governance. Different document types follow different approval routes, a currently valid version can coexist with a new one still in preparation, and at any moment the bank needs to know exactly which standards apply, both now and at any point in the past.

As Raiffeisenbank expanded its use of AI, a second need converged with the first. The bank wanted the same knowledge to serve as a reliable source of context for AI. An early attempt to point AI directly at the SharePoint content showed the limits of working with unstructured documents. Retrieving text is easy. Reliably determining which information is valid, applicable, and relevant to a specific user is a different problem. Raiffeisenbank needed more than a repository, a search engine, or a vector database. It needed a governed knowledge layer.

​

“The lifecycle of internal standards in a bank is highly complex. It covers the creation, approval, distribution, regular review, versioning, monitoring of effectiveness, and archiving of standards, all in line with regulatory and audit requirements. Simply storing documents is not enough. The bank needs to record every step clearly, keep information current and accurate, track changes, and ensure that standards are available to the right employees while also being structured enough for use by other teams across the bank. Without effective governance and transparent management, the result can be confusion, compliance risk, and errors in decision-making.”

Martina Nováková, Head of Support & Processes, Branch Network, Raiffeisenbank

​

The Solution

Dawiso delivered the project together with Profinit, a partner already established in the bank.

It started with migrating around 4,000 documents, roughly 50 GB of content, out of SharePoint into Dawiso. This was not a simple file move. The source documents carried formatting applied at several levels, and the migration had to preserve how information was presented while turning the content into governed objects that could be managed, searched, integrated, and consumed by AI.

On top of the migrated content, Dawiso built a full lifecycle and approval model for internal standards. There is no single approval workflow: the required path is determined by the document type and can coordinate different combinations of reviewers and decision-making bodies across the bank. A single process can involve line managers, document guarantors, specialist approval groups, and senior governance bodies, with different participants entering at different stages. The environment is designed for regular participation by hundreds of users across the bank.

Documents move through dedicated governance zones that reflect their lifecycle state. Content is created in an editing zone, passes through review and approval, and can wait for a future effective date before Dawiso automatically publishes the correct version. Users find only what currently applies, while a valid standard can stay live as its successor is already being prepared and scheduled. Automations orchestrate the movement between editing, waiting, publication, archive and audit zones.

For selected documents, Raiffeisenbank can define groups or individuals who must formally confirm they have read the content, with support for recurring familiarization at defined intervals.

​

Reconstructing the bank’s standards at any point in time

Beyond version history, Dawiso can reconstruct the complete standards landscape as it existed at any selected point in time - including all documents, their applicable versions, properties, and governance context valid on that date. The entire state can then be exported as a point-in-time snapshot for regulatory or internal review.

​

Main features

For Raiffeisenbank, the implementation centers on:

• Automations that move and synchronize documents across editing, waiting, publication, archive, and audit zones.
• Point-in-time reconstruction of the complete standards landscape, exportable as a snapshot for regulatory or internal review.
• AI Search, which answers questions in plain language with cited sources, in any language in and any language out.
• REST API and MCP as the two ways teams consume the content. Some teams pull specific documents through the REST API into their own context layer. Others query objects through Dawiso’s MCP and assemble answers from what it returns.
• Segment-aware retrieval, so that the answer reflects where the user works. The context of the user’s segment is part of how the AI layer is queried, which means a retail question and a corporate question return the right documents.

​

The Impact

Raiffeisenbank now governs the lifecycle of its internal standards in one place, at enterprise scale. Around 4,000 documents have moved from files in a repository to governed knowledge with explicit lifecycle, approval, validity, and organizational context. Employees always work with what currently applies, and the bank can reconstruct the full state of its standards for any past date when needed.

On top of this foundation, the same governed content now feeds AI, governed once and consumed in different ways. AI Search recently went live for users, and a conversational assistant is moving towards testing. The same layer can also serve as context for other AI initiatives across the bank. Instead of connecting every new solution to fragmented content, Raiffeisenbank has one trusted source built to serve people, applications, and AI.

​

“With the new governance and workflow platform, the bank gained a clear and auditable way to manage internal standards in line with regulatory and audit expectations. Information is centralized, easier to maintain, and available to employees without fragmentation, delays, or inconsistent versions. Just as importantly, the bank has embedded the ability to respond faster to changes in regulation, business processes, and the wider financial sector.

The AI dimension adds further value. AI Search does not only make the documentation searchable in a new way; it also opens the same governed knowledge foundation for other AI initiatives across the bank. By connecting internal standards, related working procedures, and the people involved in governance processes, the platform becomes a hub for managing know-how and compliance. Employees, applications, and AI can all draw on the same trusted context, making it easier to deliver the right information to the right audience at the right time.”

Martina Nováková, Head of Support & Processes, Branch Network, Raiffeisenbank

​

Next phase

The knowledge foundation keeps expanding. Content from Raiffeisenbank stavební spořitelna (RSTS) is set to join the environment, extending the governed model to another part of the group. The conversational assistant is moving towards testing, and the questions employees ask will create a feedback loop that shows where knowledge is hard to find or where the underlying content can be improved. As new AI use cases emerge, the bank does not need to rebuild its foundation each time. The governed context is already there.