Passage Retrieval
Passage retrieval is the practice of finding and returning specific relevant passages from within documents, rather than whole pages, which is why AI engines cite paragraphs instead of articles.
Passage retrieval pulls the paragraph, not the page. Instead of ranking whole documents, modern systems index and return individual passages — the specific chunk that best answers a query — which the engine then quotes or grounds its answer in. The unit of competition in AI search is the passage, not the article.
This single fact reshapes how you write for AEO. It means every paragraph has to stand on its own as a complete answer, because it may be retrieved with none of its surrounding context attached. A passage that depends on the sentence before it, or trails off without making its point, fails on retrieval even if the full page is excellent. Writing self-contained, answer-first passages is the extractability pillar, and it's a direct response to how passage retrieval works.
Example. Search inside a long guide and the engine surfaces one paragraph — say, the exact step that answers your question — and cites that page. The other 1,900 words didn't lose; they simply weren't the passage that got retrieved.
Relevant pillar
Related terms
- ChunkingChunking is how a retrieval system splits your page into smaller passages before indexing it, so AI engines retrieve and cite chunks of a page rather than the whole document.
- RerankingReranking is a second pass in retrieval where an initial set of candidate passages is reordered by a more precise relevance model, deciding which few actually make it into the AI's answer.
- Inverted PyramidThe inverted pyramid is a writing structure, borrowed from journalism, that puts the most important information first and supporting detail after, making each passage answer-first and easy for AI to lift.