Skip to main content
The On-Page GEO Patterns tab [Pro] shows the recurring on-page patterns across a product’s most-cited pages — the up to 4 pages AI cited most this week, not a full-site audit. Use it to see where a product concentrates its on-page effort, then adapt the patterns that fit your own pages.
On-Page GEO Patterns tab with focus areas and pattern audit

On-Page GEO Patterns for a ranked product

The tab is explicit about what is fact and what is inference. Patterns marked OBSERVED were found in the cited pages’ actual HTML; anything marked ESTIMATED — including “why it might matter” interpretations — is analysis, not observation.

Overall summary

Where They Focus GEO condenses the audit into one narrative: which of the three focus areas — Content Structure, Meta Information, and Technical — the product leans on for its AI visibility, with the observed patterns listed under each. Below it, What we observe · Why it might matter splits the reading in two: what the cited pages actually do, and what that likely means for how AI extracts and cites them.

The pattern-by-pattern audit

Each focus area has its own tab, opening with a note on what that signal group does for AI citation — and every pattern is a card showing:
  • the pattern’s definition (e.g. Answer-first structure: the first sentence under the H1 gives a direct definition or summary of the topic)
  • how many of the cited pages it was observed on (e.g. 4/4)
Expand a card for the Pattern Summary — how the product applies the pattern — and Cited pages with evidence: each cited URL with its citation count and the actual HTML snippet where the pattern appears. Patterns the pages don’t use are audited too, listed as Not observed — so you see not just what a top-ranked product does, but what it gets cited without.

Reading it for one page type [#pricing-pages]

The tab is easiest to use when you come in with one of your own pages in hand. Pricing pages are the common case, since they’re cited more often than most other page types (the data is here) and AI reads company-size fit off them. Filter your attention to the patterns whose evidence comes from a leader’s pricing URL, and read the Cited pages with evidence panel rather than the pattern name — the HTML snippet shows how that page presents its plans and prices in the markup a bot reads. That’s the comparison worth making against your own pricing page. The GEO Guide links here from What are the cited pages doing on the page?.