Integrating AI assistants and automation into your systems

Anyone running their own applications asks how AI can work with existing data and systems without confidential documents ending up in third-party services.

We help you run AI models locally, on servers in your own infrastructure. The models then work with your documents, emails and company data, and these stay in your house instead of going to an external AI service.

Your applications call the model through an interface (API). A runtime such as Ollama serves the model, and with retrieval-augmented generation (RAG) it answers from your own documents. Typical uses are answering questions, summarising, sorting incoming mail, searching documents and extracting data.

With AI, we keep two things apart.

AI that we build into your systems

This is the service described above: we connect AI models to your systems, for example an assistant that answers from your documents or automation that sorts incoming email. You order it as a project of its own.

AI tools in our own work on your code

When we maintain, develop or migrate your applications, we work without AI tools by default. Unless you ask for it, no AI tool sees your code, your data or your documents. This applies to everyone who works with us on your project.

Source code holds your business knowledge: processes, rules and interfaces to other systems. That is why it does not go to an external service without your consent.

If you want it, we build new components with AI assistance, for example a new service alongside your existing system. A senior engineer remains responsible for every change, and you can withdraw your consent at any time.