Glossary
RAG — retrieval-augmented generation
RAG — retrieval-augmented generation — is the architecture behind most answer engines: retrieve documents relevant to the query, then have a language model compose its answer from them, grounding the response in sources it can cite.
The two stages define the optimization problem. Retrieval decides whether content is in the room when the answer is written; generation decides whether it gets quoted and credited. Content that is never retrieved can never be cited, which makes retrievability — clear structure, self-contained passages, unambiguous entities — the precondition for everything else in AEO and GEO.
Grounding is the architecture's promise and its pressure point: an answer assembled from retrieved sources can cite them, but it inherits whatever the retrieval brought in — which is what makes evidence quality, and evidence poisoning, live research problems.
See also: Answer engineGEOAI crawlers