Retrieval & ranking
Retrieval-Augmented Generation(RAG)
Fetch relevant documents first, then have an LLM write an answer grounded in them.
Retrieval-Augmented Generation (RAG) is the pattern behind most AI search: the system first retrieves relevant documents from an index, then feeds them to a language model that writes an answer grounded in what was retrieved. The model is not answering purely from memory; it is summarising sources fetched for this specific query.
RAG is why on-page AEO works at all. If the engine writes answers from retrieved documents, then being in the retrieved set — and being the clearest, most quotable document in it — is the whole battle. It also explains the two-step nature of AI visibility: first you must be retrieved, then you must be selected and cited from among what was retrieved.
Everything from chunking to embeddings to grounding sits inside the RAG pipeline, which is the mental model to hold when you ask why a given page did or did not get cited.
Related terms
Grounding
Constraining an AI's answer to retrieved source documents instead of model memory.
Chunk Retrieval
Fetching the top few most relevant text passages from a vector index for a query.
Vector Embedding
A list of numbers that encodes meaning so related content sits close in vector space.
Retrieval Corpus
The full body of documents a retrieval system searches to answer a query.
Selection Rate
How often an AI cites your page once it's already in the retrieved set.
Want to know where you stand in AI answers?
We’ll show you which engines cite you, and which don’t.
Get a free AI citation audit