Retrieval-augmented generation (RAG)
A pattern that retrieves approved information before an LLM generates an answer.
Why it matters
RAG can improve freshness and traceability, but it does not guarantee correctness. Retrieval quality, source ownership, permissions, chunking, and answer evaluation determine whether the pattern is useful.
Use it well
- Give important sources owners and review dates.
- Preserve citations or source references where users need to verify.
- Evaluate retrieval and generation separately.