how to rank in AI search (it's not what most people think)

we noticed something odd about six months into running FlowIntent. certain pages were getting cited by Perplexity and ChatGPT search that weren't ranking well in traditional Google results. and other pages — well-structured, properly optimised, ranking in positions 3–7 — weren't being cited at all.

that discrepancy sent us down a rabbit hole. we started tracking which pages got cited, by which AI engines, and more importantly: what those pages had in common that the non-cited pages didn't.

how AI citation differs from ranking

traditional search ranking is a function of relevance signals combined with authority signals. AI citation works differently. authority matters less than you'd expect for newer AI search engines. Perplexity and ChatGPT search don't weight domain authority the way Google does — they weight answer quality in the moment. a six-month-old page on a low-authority domain can get cited consistently if its answer structure is cleaner than the alternatives.

what matters is: can the AI system extract a clear, accurate, usable answer from this page without ambiguity?

the 4 signals that consistently trigger AI citations

1. the direct answer appears in the first two sentences of the relevant section. not in the first paragraph. in the first two sentences of the section where the answer lives. AI systems extract from the beginning of sections. if your answer is buried after three sentences of context-setting, it will often be skipped.

2. the answer is specific, not hedged. "it depends" is the death of AI citation. AI systems are looking for declarative statements they can attribute and cite with confidence. hedged answers get passed over in favour of specific, confident statements. instead of "keyword difficulty varies by tool," write "keyword difficulty scores between 0–30 are generally achievable for new sites within 60–90 days."

3. the page has a defined topical boundary. pages that try to cover too much are poorly structured for AI citation because there's no clear extraction point. AI systems prefer pages with a single, clear topical focus.

4. there is at least one verifiable factual claim with a source. pages that include cited statistics or named studies get cited significantly more often than pages with only opinion and advice.

platform-by-platform differences

Perplexity is the most citation-transparent AI search engine. it shows sources and pulls consistently from pages with direct answers and clear topical boundaries. it's also the most likely to surface lower-authority domains if content quality is high — making it the most achievable target for new sites.

ChatGPT search has shown a stronger bias toward established domain authority. it cites well-known publications more consistently than Perplexity.

Google AI Overviews pull primarily from pages already in Google's index with established trust signals. new pages typically take 4–8 weeks after indexing before appearing in AI Overviews.

the shift worth making

AI search is not a separate strategy from SEO. the content that gets cited in AI search is the content that best answers the question. the content that best answers the question also tends to rank better in traditional search. the mental shift is from "what keywords am I targeting" to "what question am I completely answering." make that shift and both problems get easier at once.