02 / Sessions & Context
RAG
Retrieval-augmented generation. Before answering, the system looks up relevant passages in your own documents and hands them to the model, so the answer is based on your material.
How you’ll hear it
“Can it answer questions about our policies?”
“With RAG, yes. It finds the relevant policy before it answers.”
Avoid
Assuming RAG removes hallucination. It reduces it when retrieval finds the right passage.
The longer version
RAG connects a model to a searchable store of your content. Each question triggers a search, the best passages go into the context, and the model answers from them. Quality depends heavily on that search step, so well-organised, current documents matter more than the choice of model.