Knowledge is buried in documents.
Policies, procedures, and manuals exist, but nobody finds the right answer quickly.
Product & platform
Custom AI software, built around your own knowledge.
Sometimes an existing tool is not enough. Then I build AI-native software that does exactly what your team needs: a chatbot that answers from your own documents, a dashboard that gives control over model usage, or an internal tool around your process.
- RAG chatbots
- LLM dashboards
- Internal tools
- AI in your product
The starting point
Does this sound familiar?
A general chatbot is not reliable enough.
ChatGPT gives fluent answers, but not from your sources and without showing where the answer comes from.
No visibility into AI usage.
Teams use different models and tools. Cost, quality, and risk are hard to oversee.
Approach
How I approach it.
Start small with real users, then scale to production.
-
01
Use case and sources
We decide which questions the system should answer, for whom, and from which sources.
-
02
Prototype with real users
A working version, fast, tested with the people who will use it.
-
03
Make quality measurable
Source citations, test questions, and clear boundaries: what the system does and does not know.
-
04
Into production
Securely connected to your environment, with agreements on maintenance and further development.
Result
What you take away.
- A working custom AI application
- Answers with source citations
- Test questions and quality criteria
- Agreements on maintenance and development
Outcome Software that does exactly what your team needs.
Frequently asked questions
Questions about AI software & chatbots.
Is your question not listed? Ask it directly in a short introduction.
Go to contactWhat is a RAG chatbot?
RAG stands for retrieval-augmented generation. The chatbot first looks up the relevant passages in your own documents and only then writes an answer, with a reference to the source. That makes answers verifiable and keeps them close to your own policies.
How is this different from ChatGPT?
ChatGPT answers from general knowledge. A RAG chatbot answers from your documents, shows the source, and can be configured with your own rules about what it does and does not answer.
What happens to our data?
Upfront, we agree which sources the system uses, where data is stored, and which models are used. I make those choices together with you, in line with your privacy and security requirements.
What does custom AI software cost?
Implementation projects typically start at €2,000. Custom software depends on scope, integrations, and maintenance. You get a concrete estimate in the introduction.
One concrete bottleneck is enough
Which knowledge should your team find faster?
Tell me which questions keep coming back and where the answers live today. Then I look at what a first version should be able to do.