Latest Posts

Your AI project starts here

Want to unlock the potential of LLMs in your company? We'll show you where getting started pays off.

RAG explained: making company knowledge accessible via chat

RAG explained: making company knowledge accessible via chat

Language models know a lot about the world – but nothing about your company. Retrieval Augmented Generation, or RAG for short, closes this gap: before the model answers, the relevant passages are retrieved from your own documents and provided as context. The answer is thus based on your knowledge and can be verified at any time via the source reference.

What matters in practice

  • Data quality: Outdated or contradictory documents lead to poor answers. Preparing the sources is the most important step.
  • Access permissions: The assistant may only show what the person asking is allowed to see.
  • Quality measurement: Test questions and ratings are used to continuously monitor answer quality.

A good knowledge assistant isn't a chatbot with access to everything, but a carefully curated system with clear sources and rules.

Lea Hoffmann

Lead AI Engineer, DeepIonics

Implemented correctly, a RAG assistant saves employees valuable time searching for information every day. Service, sales and the onboarding of new colleagues benefit in particular.