RAG retrieves relevant documents and feeds them into the model's context window so it can answer questions using your data — without fine-tuning. It's the most common production AI architecture, and the most commonly broken one.
Why this matters
RAG is the default enterprise AI pattern. But 'Dumb RAG' — naive retrieval without evaluation — is the #1 technical killer of production AI (Composio). Getting it right requires monitoring what most teams skip.