# How do AI search engines choose which sources to cite?

Canonical: https://seangarschi.com/answers/how-do-ai-search-engines-choose-sources

Updated 2026-08-19. Category: GEO & SEO.

An AI search engine retrieves a candidate set of pages, then favours the ones whose content directly and unambiguously answers the query, can be quoted without editing, agrees with other retrieved sources, and appears current. Retrieval gets you considered; extractability and corroboration get you cited.

## The four filters

It helps to picture it as a funnel where most pages die at the first two stages.

- Retrieval: can the system find and read your page at all? Blocked crawlers, client-only rendering, and slow responses eliminate you before any judgement about quality happens.
- Relevance: does the page address the specific question, or does it address the general topic? Topic pages lose to question pages consistently.
- Extractability: is there a passage that stands alone? A model needs a block it can quote without inheriting context from three paragraphs above it.
- Corroboration and recency: do independent sources agree, and does the content look maintained? Contradiction and staleness both push a model toward a safer citation.

## The practical implication

Write the page so that any single section could be lifted out and still make sense. Use plain declarative sentences for anything you want quoted. Put a visible last-updated date on the page and mean it.

## Related questions

**Do AI engines prefer big brands?**

They prefer corroborated ones, which correlates with size but is not the same thing. A small, specific, consistently described business often outperforms a large vague one on a narrow question.

**Does page freshness matter?**

Yes, particularly for anything with a factual or time-sensitive component. Dated content is a weaker citation and models discount it accordingly.

Source: https://seangarschi.com/answers/how-do-ai-search-engines-choose-sources
Contact: start@growthr.com