What is retrieval-augmented generation (RAG)?
Retrieval-augmented generation is a pattern that retrieves relevant information from an external knowledge source and supplies it to a generative model so the response is grounded in information that is more specific and current than the model's training data alone.
In practical terms
RAG is useful when answers need to reflect enterprise policies, case information, product documentation, or other controlled sources. The retrieval step should be observable and testable: teams should know what was retrieved, why it was selected, and which source supported the final answer. RAG reduces hallucination risk but does not eliminate the need for validation, confidence handling, and human escalation.