Is ChatGPT
Recommending You?

Why “AI visibility” is the channel nobody's measuring

When a developer asks ChatGPT “what's the best CI/CD platform for a startup,” the three or four names it returns are the new search results. If you're not in them, you don't exist at the moment of intent — and unlike Google, there's no page two to climb.

The catch most tools hide: there are two completely different numbers. Ask an engine “tell me about [your product]”and it'll happily describe you — that's branded recall, and it's near-100% for almost everyone. Ask it a non-brandedbuyer question and see if you get named on merit — that's organic visibility, and for most products it's close to zero. This checker only measures the second one, because it's the one that maps to revenue.

How the score works

• We generate 6 organic, non-branded buyer questions from your category.
• We ask four engines through their API models, not the consumer apps: ChatGPT (gpt-4o-mini), Claude (claude-haiku-4.5), Gemini (gemini-2.5-flash), Perplexity (sonar). That is 24 checks.
• The score is the share of answered checks that named you, from 0 to 100. Perplexity counts 1.5 times because it searches the live web; the other engines count once.
• If an engine fails to answer, those checks are left out of the score and the engine shows as "no data", so a partial run is scored on fewer checks.
• We tally every competitor the engines named instead: your share of voice.

This is not a full AI-visibility baseline. A full baseline runs forty-plus questions against each engine, keeps branded recall on its own line and reports four numbers per engine with no blended score. A score here and a baseline there cannot be compared.

Built by GTM Labs — who ran this on their own site, scored ~1% organic, and are fixing it in public. The methodology mirrors the GEO baseline work behind The Invisible Hand. The score is a directional read from a lightweight sample, not a full audit.