Service
Knowledge management & RAG
Chat with your company knowledge. Answers with source references, self-hosted, your data stays with you.
We capture the knowledge that otherwise lives only in people’s heads and scattered files, and make it usable for AI, as a searchable knowledge base your team can query in seconds.
How does chatting with your own company knowledge work?
We connect a language model to your own documents, so it answers questions based on your company knowledge. This method is called RAG, short for Retrieval Augmented Generation: the AI first looks things up in your documents and then answers on that basis. Every answer comes with a source citation, so you can see which document it came from and do not have to trust the AI blindly.
Why is a public ChatGPT not enough for this?
A public ChatGPT does not know your company knowledge, and your internal questions do not belong in someone else’s input field. Yet that is already happening in the shadows: according to Bitkom, in 25 percent of companies staff use AI tools on their own initiative, mostly with no rules for it. Your own knowledge store gives your team the same convenience, but with your content and under your control.
Does the AI make answers up?
The risk can be reduced significantly. Because the AI forms its answer from your stored documents and provides the source, you can check every statement. Fraunhofer IESE also describes RAG as a way to reduce hallucinations and keep sensitive data in house. When there is no source for a question, the system says so rather than guessing.
Does our data stay in house?
Yes. The knowledge store runs self-hosted, your documents do not leave your company. This lets you use even sensitive internal knowledge without handing it to an external cloud provider, and makes GDPR-compliant use much easier than with a US SaaS.
Does everyone then see everything?
No. Access follows your existing permissions, so each person only finds what they are allowed to see. A sales document does not suddenly become visible to everyone through search, and sensitive areas stay protected.
What does an SME use this for in practice?
Typical uses are fast access to internal knowledge in support, onboarding new staff, retrieving quotes and contracts, and questions to manuals and quality documents. Wherever important knowledge sits in individual people’s heads and scattered files, a searchable knowledge store saves onboarding, follow-up questions, and the hunt for the one right file.
How does a project run?
Every project follows three steps, so you get clarity on value and cost early:
- Potential analysis: we look at your documents and questions and show where a knowledge store delivers the most. More on the potential analysis
- Implementation at a fixed price: you get a binding offer for a clearly defined knowledge store, self-hosted on your infrastructure.
- Operation and further development: on request, we keep the system current and extend it with new documents and sources. More on operation
What does it cost, and who owns the code and data in the end?
We work at a fixed price, after a short potential analysis. The code and data are yours, with no per-user licence and no lock-in, so you can maintain or hand over the knowledge store yourself.