Google AI Overviews, ChatGPT, and Perplexity choose which pages to cite by weighing four things: structured data, a clear self-contained answer, trustworthy and authoritative signals, and freshness. When someone asks who is a good insurance agent near me, the site that scores well on all four is the one that gets named.
Think about how a buyer actually asks. Who is a good insurance agent near me is a citation question, the AI has to pick a name. How do I choose a Medicare plan is an explainer question, the AI has to pick a source that explains it well and often names an agent to talk to. Both come down to the same underlying decision: which page is the engine confident enough in to hand to the user.
None of the major AI answer engines publish an exact formula, but their public documentation and observed behavior point to the same four signals, in roughly this order of importance.
JSON-LD schema that spells out who you are, what you do, and what questions you answer. It is the difference between an engine guessing at your page and reading it with certainty.
A paragraph that fully answers the question on its own, without requiring the rest of the page for context, so it can be lifted and quoted as a standalone passage.
Consistent business information, real reviews, a claimed Google Business Profile, and a site that looks like it belongs to a real, established local business rather than a thin content farm.
Content that is current and regularly maintained, which matters even more for insurance, where plan years, enrollment windows, and rules change annually.
Before any of the four signals get evaluated, a page has to clear a basic floor: it has to load, it has to be indexable, and its robots.txt cannot be blocking the AI crawler in question. A perfectly written page that blocks GPTBot or PerplexityBot is invisible to that engine no matter how well it would otherwise score.
From our study of 533 independent insurance agent websites.
What Answer Engine and Generative Engine Optimization mean, and the page structure they require.
The data behind why most agent sites cannot be cited yet.
Why buyers now ask AI who to call, and how to be the answer.
The consistency signal that feeds the trust factor engines look for.
The complete system these signals fit into, start to finish.
The step-by-step checklist agents follow to show up in Search and Maps.
Google AI Overviews draws from pages that already rank well for the query, then favors the ones it can extract a clear, self-contained answer from, backed by structured data and signals of trust like consistent business information and real content. It is built on top of Google's normal ranking system, so a page still has to be findable and relevant before it can be cited.
Mostly, yes, with a different mechanism. ChatGPT and Perplexity use live web search and retrieval to find pages, then favor ones that are clearly structured, directly answer the question, and read as trustworthy, similar to what Google AI Overviews looks for. A site that is blocked to their crawlers cannot be cited at all, no matter how good the content is.
You control readiness: writing a direct, self-contained answer to the exact questions buyers ask, adding connected structured data and real FAQ content, keeping your business information consistent everywhere, and allowing AI crawlers to read your site. You cannot control or guarantee that any specific engine names you on any specific query, but readiness is what makes it possible.
Start with a free Agent Visibility Score. See whether your site is structured, self-contained, trustworthy, and current enough to be the source AI picks.