AI glossary
Retrieval-augmented generation (RAG)
A technique where relevant documents are retrieved first and given to the model, so it answers from your sources instead of from memory.
RAG combines search with generation. When a question comes in, the system looks up the most relevant passages — in your documentation, a database or the web — and includes them in the prompt. The model then answers using that material and can cite it.
RAG reduces hallucinations, keeps answers up to date without retraining, and lets a model work with private information it was never trained on.
Example: A support bot that answers from your help-centre articles instead of guessing.
In practice
- Quality depends above all on retrieval: if it does not find the right passage, the model cannot get it right.
- Keep the document base up to date and remove old versions.
- Ask the model to cite the passage it uses and to say “I don’t know” if the answer is not there.
Why it reduces errors: hallucinations and privacy.
