Artificial intelligence

AI assistants, RAG, AI agents, support, documents, human-in-the-loop and quality assessment.

How to tell whether your business needs AI

AI should not begin with choosing a model. The right question is where your team does repetitive knowledge work every day that could be sped up without losing control.
Read more
Why an AI project starts with the process, not the model

Most failed AI projects break not because the model was weak but because the process was never described.
Read more
RAG in plain words

RAG lets AI answer from the company’s own documents and sources instead of generating text out of thin air.
Read more
AI agents: where they help and where they are dangerous An AI agent can check data, call APIs, create tasks and prepare reports. But the more actions it takes, the more the oversight matters.
- simple scenarios
- the risk of autonomous actions
- logs
- human confirmation
- the limits of responsibility
How to prepare a company for adopting AI Before AI you need processes, data, documents, access rights and realistic expectations in the team.
- an audit of what the company knows
- the quality of the documents
- access rights
- the role of the human
- a pilot scenario

Want to test an AI scenario on your own process?

Describe where your team searches for information every day or repeats the same answers. We will work out whether AI fits here and where to start.