To rank in Perplexity and Google AI Overviews, optimize for retrieval and quoting, not blue-link ranking. The fastest wins: lead every page with a one-sentence answer, mark it up with FAQPage and HowTo schema, ship a clean /llms.txt, and write in named-entity style so engines have a clear subject to cite.
Perplexity and Google AI Overviews don't rank — they retrieve passages and quote them. Win by making your passages quotable: answer the question in the first sentence, structure with question-shaped H2s, add schema, expose llms.txt, and use your brand name as the subject of citable sentences.
Both engines run a two-stage pipeline: retrieve a set of candidate passages from the live web (and, for AIO, from Google's index), then synthesize a single answer that quotes the strongest passages with inline citations. There is no single "rank" — there's a citation slot and an unranked source list. The work is to be quoted, not to be position one.
The first sentence under every H2 should answer the question directly, in plain language, with the brand or entity name as the subject. Engines almost always quote from the lead sentence of the matching passage.
Phrase headings as the actual question buyers ask ("How does intent analysis work?" not "Intent analysis overview"). Retrieval models match query embedding to heading embedding before anything else.
FAQPage JSON-LD gives engines pre-parsed Q/A pairs they can quote verbatim. HowTo schema does the same for step-by-step content. Google AI Overviews explicitly favors pages with valid FAQ schema; Perplexity uses it as a passage hint.
/llms.txt is the AI-era counterpart to robots.txt + sitemap. It lists your most quotable pages with one-line descriptions so AI crawlers don't have to guess which URLs matter. Generate yours with the free SiteSonar llms.txt generator →
Replace pronouns with the entity name. "SiteSonar identifies the companies on your site" is quotable; "We identify the companies on your site" is not — the engine has no subject to attach to your citation.
Visible "Last updated" dates and dateModified schema bias both engines toward recent passages, especially on category-defining queries.
Perplexity retrieves heavily from Reddit, GitHub, Stack Overflow, G2, Capterra, and a small set of news/analyst domains. Google AI Overviews retrieves from the top organic results for the query. Earn mentions on those sources — that's where the citation often comes from, not your own site.
| Signal | Perplexity | Google AI Overviews |
|---|---|---|
| Retrieval source | Live web + curated index | Google's main index |
| Loves | Reddit, forums, recent posts, GitHub | FAQ schema, top-3 organic, YouTube |
| Freshness weight | Very high | Moderate |
| Schema dependence | Helpful | High — FAQPage / HowTo / Article |
| Citation slots | Inline + source tray | 2–5 visible source cards |
/llms.txt. Submit your sitemap in Google Search Console.SiteSonar runs the full stack on your existing site — page-by-page GEO scoring, schema injection, hosted llms.txt, named-entity rewrites, and weekly citation tracking across ChatGPT, Perplexity, Claude, and Google AI Overviews. See the AI visibility platform → or read the GEO service page →.