What is Retrieval-augmented generation (RAG)?

Also called RAG.

Retrieval-augmented generation is a technique where a language model retrieves relevant documents at query time and uses them to generate its answer, rather than relying only on knowledge learned during training.

RAG is why AI visibility can change without a model being retrained: if the retrieval step surfaces different sources, the answer changes. It is also why being present on retrievable, trusted pages matters more than model recency.

All glossary terms