Retrieval-Augmented Generation (RAG) explained
Retrieval-augmented generation (RAG) combines two steps. Retrieval finds documents relevant to a question — from a search index, a crawler or a company’s own knowledge base. Generation uses a large language model to write an answer grounded in those documents.
RAG helps AI systems give more current and verifiable answers than a model alone, and it’s the basis of most AI search products’ web answers. Google uses the term grounding for a similar idea.
For businesses, the lesson is practical: if your pages can’t be retrieved — blocked, unindexed, or with content hidden behind JavaScript — they can’t be used, however good they are.
Retrieval-Augmented Generation (RAG): an example
Asked about this month's interest rates, an AI assistant searches the web, reads current bank pages and writes an answer citing them — rather than relying on what it learned months ago.
Why it matters for search and AI visibility
RAG is why crawlable, current, clearly written pages matter for AI visibility: they are what gets retrieved.
Last reviewed by the Prisol team.