Are you found by AI? Log in Free scan
Research

What actually gets a business named by AI

We ask all seven answer engines, ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek and Google's AI, the questions real buyers ask, for 3,344 real businesses, and score six readiness signals against who gets named. The page recomputes from the live index on every visit. One signal tracks with being named far more than any other, and the SEO industry sells a different one.

AI names them (758) AI ignores them (1,196)
Brand, off-page recognition41-point gap
79
38
How many independent sites talk about you, and whether the engines already know you as a real entity. This is the one that moves.
E-E-A-T, trust & credentials16-point gap
75
59
Technical SEO2-point gap
80
78
The invisible businesses are as technically sound as the named ones. Speed and crawlability do not get you named.
Schema / structured data10-point gap
76
66
Structured data helps AI read you. It does not make AI recommend you.
Off-page brand recognition is the signal that most clearly separates the named from the ignored in this data. AI recommends the businesses it has seen talked about across the open web. You cannot schema your way onto the shortlist.

Most businesses are barely there

23% named
42% occasional
36% invisible

36% of the businesses we measure are invisible to AI, scoring near zero when a customer asks the engines who to trust. Most owners have never checked.

Where AI gets the mentions it trusts is mapped, per industry, on the citation-source maps, and every category leaderboard lives on the live AI Visibility Index.

More research

The viral 2026 SEO tier list, measuredEvery tactic on the shared tier list checked against the live corpus: backlinks at S confirmed, schema at C confirmed, exact-match domains unmeasured by anyone, and the row the list is missing. The viral Cal.com playbook, checked claim by claimA thread credits four tactics for Cal.com topping AI recommendations. Measured: Calendly 77, Cal.com 40, and zero of Cal.com's 32 comparison pages were cited by any engine. What holds up, what doesn't, and the parts worth copying. Who wins the moment buyers compare you?1,001 high-intent SERPs analysed: brands own 81% of their own pricing searches, and lose 67% of "vs" and 84% of "alternatives" searches to pages they didn't write. If JavaScript puts your schema on the page, AI never sees itAI crawlers read raw HTML and don't run JS, what that means for tag-injected JSON-LD, and the 60-second self-check. The Microsoft GEO model, measuredMicrosoft's official playbook describes how AI picks recommendations, fan-out queries, freshness, commercial signals. Every factor mapped to a live measurement. AI visibility tools, mappedNine tools compared on published pricing checked 9 September 2026, with engine coverage and vendor sources.
See where your business stands.

Free, about a minute, no signup. You get whether AI names you, and the one lever that moves it.

Run your free scan →

Method: 3,344 businesses, each counted once on its best clean measurement, asked the real buyer questions for its category across seven AI engines, multi-sampled with a majority vote. Signal scores are our six-dimension readiness model, 0 to 100. Named, occasional and invisible mean visibility of 60 and over, 25 to 59, and under 25. Figures recompute live. Method open at github.com/techhorizonlabs/thl-open.