Retrieval
The step where a system finds and fetches your page. If a page is blocked, unreachable, or clearly not about the topic asked, it is never retrieved, and nothing else matters.
Updated August 21, 2026
Retrieval is the first gate: before a model can trust or quote a page, it has to fetch and read it. A robots rule, an aggressive bot filter, or content that only appears after heavy JavaScript can all leave the crawler with an empty or missing page. This is why technical accessibility is scored first — everything downstream depends on it.
Related terms
- Crawler (bot). An automated program that fetches web pages. Search engines and AI systems use crawlers to read the web. Crawlers read the raw page and often do not run all of a site's scripts.
- Citation. When an AI answer credits or quotes your page as a source. To be cited, your claims usually need to be specific and backed by identifiable evidence the system trusts.
- Answer extraction. The step where the system lifts a specific fact from your page into its answer, like a price, a definition, a step or a yes or no. Content written as clear, self-contained statements extracts cleanly. Vague copy does not.