The 8 Citation Signals That Get You Cited

Eight measurable, weighted signals drive whether an AI answer engine cites a page — a synthesized field model, ranked from FAQ schema down to author attribution, with each weight to be verified against a primary source at build.

Which eight signals get a page cited by an AI answer engine?

Eight signals, in descending weight, drive whether an AI answer engine cites a page: FAQ schema (20%), answer-first structure (19%), statistical density (16%), heading structure (16%), freshness (8%), crawler access (8%), schema coverage (7%) and author attribution (6%). Each is a measurable change to a single page.

Source discipline: these weights are a synthesized field model assembled from public GEO/AEO research, not a measured law. Treat the percentages as a relative ranking, each with a verify against primary at build note; the ordering is the durable signal, the exact decimals are not. The one published benchmark behind the model: Aggarwal et al. (2023) measured that adding statistics, quotations and source citations lifted a source's visibility in generative-engine responses by up to 40%, while keyword stuffing produced little to no gain.

What does each citation signal do, and how do you build it?

FAQ schema (20%) is the highest-weighted citation signal because it hands an answer engine a pre-structured question-and-answer pair it can lift verbatim. Write the Q&A visibly and mirror it word-for-word in FAQPage JSON-LD, as this page does. Google limited FAQ rich results to well-known, authoritative government and health sites in 2023 and stopped showing them entirely on 7 May 2026 (Google, 2026); the signal's value is the liftable chunk and the parseable node, not a search snippet.

Answer-first structure (19%) gets a page cited because the engine can quote the opening sentence without rewriting it. Answer engines split a page into Chunks and pick the best-matching one, so sections compete for a citation, not whole pages (Alpar et al., 2026). Lead every section with a standalone answer, then explain.

Statistical density (16%) raises citation likelihood because named numbers, dates and figures are exactly what a generative answer needs to substantiate a claim — the effect Aggarwal et al. (2023) measured at up to 40%. Replace vague qualifiers with quantified, dated, sourced facts.

Heading structure (16%) helps because a most-important-noun-first hierarchy lets the engine map your page to a query. Put the key entity at the front of every H2 and H3, or phrase the heading as the question the section answers.

Freshness (8%) contributes because engines prefer recently verified content for fast-moving topics. Show a visible last-verified date and never change it without changing the content.

Crawler access (8%) is a precondition: an engine cannot cite a page its crawler cannot read. OpenAI states that sites opted out of OAI-SearchBot are not shown in ChatGPT search answers (OpenAI, 2026), and Perplexity documents PerplexityBot as the crawler that surfaces and links websites in its results and recommends allowing it (Perplexity, 2026). Permit the search crawlers in robots.txt and serve content without a render wall.

Schema coverage (7%) adds structured data beyond FAQ — Article, Person, BreadcrumbList — so the engine can resolve authorship, recency and hierarchy in one parse.

Author attribution (6%) raises citability because a named, credentialed author is an E-E-A-T signal the engine can verify. Attach a Person entity with real credentials.

Why are the citation signals also agent-readiness signals?

Every citation signal above is also an agent-readiness signal, so the same investment pays off in both the human-search channel and the LLM channel at once. FAQ schema, answer-first structure and schema coverage are the content and discoverability readiness checks; crawler access is the access-control check; freshness and author attribution are the quality and E-E-A-T checks.

The 8 citation signals — frequently asked questions

Are the eight citation-signal weights measured or modeled?

Modeled. The weights are a synthesized field model assembled from public GEO/AEO research, not a measured constant: the ordering, with FAQ schema and answer-first structure on top, is the durable part, and each percentage is flagged to verify against a primary source at build.

Does FAQ schema still matter now that Google no longer shows FAQ rich results?

Yes, for a different reason. Google stopped showing the FAQ rich result on 7 May 2026, but a visible question-and-answer block mirrored word-for-word in FAQPage JSON-LD is still a chunk an answer engine can lift and a structured node it can parse.

Which citation signal has a published benchmark behind it?

Statistical density. Aggarwal et al. (2023) measured that adding statistics, quotations and source citations lifted a source's visibility in generative-engine responses by up to 40%, while keyword stuffing produced little to no gain.

Which citation signal is a precondition rather than a weight?

Crawler access. An engine cannot cite a page its crawler cannot fetch: OpenAI states that sites opted out of OAI-SearchBot are not shown in ChatGPT search answers, and Perplexity recommends allowing PerplexityBot so pages can be surfaced and linked.

Related: the GEO pillar · getting cited by ChatGPT · getting cited by Perplexity · getting cited by Claude · build the signals as content readiness · declare citable content with llms.txt · audit whether your site ships these signals · GEO defined

← GEO · .md