What Is an AI Visibility Audit?
An AI visibility audit measures whether and how often a business appears in AI-generated answers across ChatGPT, Gemini, Perplexity, and Google AI Overviews for the questions its customers actually ask. It tests real prompts from clean, logged-out sessions, then diagnoses the entity, content, and third-party gaps causing the business to be absent or under-cited.
What the audit is really checking
An AI visibility audit answers one question: when your customers ask AI about businesses like yours, does it say your name? Everything in the audit works backward from that.
The audit runs the prompts your customers use, not vanity prompts. For a med spa that means questions like "best med spa in [city] for [service]," "where should I get [treatment] near [neighborhood]," and the service-and-location combinations patients actually type. It runs them across the major engines, because presence on ChatGPT does not guarantee presence on Perplexity or in Google AI Overviews. Each engine weights sources differently and refreshes on its own schedule.
Why it must run logged-out
If you run these prompts from your own account, the model already knows you. Your history, your location, and your past searches bias it toward naming you. That is a false positive. A real audit runs from a clean session with no memory and with prompt phrasing that does not leak your location, so the result reflects what a genuine prospective patient would see, not what the model shows its owner.
What the audit diagnoses
Being absent is a symptom. The audit finds the cause, which usually falls into a few buckets:
Entity gaps. Inconsistent name, address, phone, or business description across the web, so the model cannot confidently identify you.
Content gaps. No extractable, answer-first content matching the questions being asked. The model has nothing clean to lift.
Consensus gaps. Little or no third-party corroboration. You describe yourself, but no trusted external source backs it up.
Structure gaps. Missing or invalid schema markup, or content trapped in client-side rendering that AI crawlers cannot read.
Mis-citation gaps. The model names you but describes you wrong, attributes the wrong services or location, or confuses you with another business.
What you do with the results
A useful audit does not stop at "you are not showing up." It maps each gap to the specific fix and prioritizes by leverage: which queries are winnable fastest, which fixes unlock the most sub-queries, and where the practice is closest to the citation threshold. The output is a build order, not just a diagnosis.
Want this done for your practice — and verified from a clean, logged-out session?
Book a strategy call