How ChatGPT, Perplexity and Google's AI decide which clinics to recommend

By Michael Cameron, founder of Cited · October 30, 2026 · 5 min read

Short answer: AI assistants recommend the clinics they can find, read and corroborate. They pull from a web index, favor pages that directly answer the question asked, and lean on information that agrees across many sources. A clinic that blocks their crawlers, has inconsistent listings, or has no page about the service a patient is asking about is easy for them to leave out.

What "AI search optimization" means

When a patient types "best med spa for Botox near me" into ChatGPT or Perplexity, they don't get ten blue links. They get a short answer that names a handful of clinics. AI search optimization is the work of making your clinic one of those names. You'll also see it called generative engine optimization (GEO), answer engine optimization (AEO) or AI visibility. The labels differ; the goal is the same.

It is not a separate discipline from SEO. It is SEO's fundamentals, plus a few things that matter more when a machine is writing the answer: whether each engine's crawler can reach your site, whether your pages state facts plainly enough to be quoted, and whether the rest of the web agrees with what your site says about you.

Where the answers come from

Each engine retrieves from an index of the web and then writes an answer from what it retrieved. That means the first question is simply whether your pages are in the index each engine uses. Here is what the companies themselves say:

• Google AI Overviews and AI Mode. Google's guidance for site owners says the same SEO best practices apply, and that there are no additional requirements or special optimizations needed to appear. To be shown as a supporting link, a page must be indexed and eligible to show a snippet in Google Search. Google also points to keeping your Google Business Profile up to date. (Google Search Central)
• ChatGPT search. OpenAI runs a separate crawler for search, OAI-SearchBot, and documents that sites which opt out of it will not be shown in ChatGPT search answers. Blocking GPTBot (its training crawler) is a different setting. (OpenAI crawler documentation)
• Perplexity. Perplexity recommends allowing PerplexityBot in robots.txt if you want your site to appear in its results, and says changes can take up to 24 hours to take effect. (Perplexity crawler documentation)

The practical takeaway: check your robots.txt and any firewall or "bot protection" setting on your website host. Some website builders and security plugins block AI crawlers by default, and a clinic can disappear from ChatGPT search without anyone deciding it should.

What we've logged in Houston: the engines disagree

Every night we ask the engines the questions patients ask, in fresh, unpersonalized sessions, and log who they name. One pattern shows up constantly: ChatGPT and Perplexity often recommend different clinics for the exact same question.

Single sessions, one day. AI answers vary between runs, so treat this as an observation, not a ranking.

Two lessons follow. First, checking one engine tells you little about the others. Second, the clinics named by both engines tend to be the ones whose information is easy to find and confirm in many places, which is the part a clinic can actually work on.

The signals a clinic can influence

No engine publishes a recipe, and anyone who claims to know the exact weighting is guessing. These are the fundamentals that the engines' own guidance and our logs point to:

1. Crawler access. Let Googlebot, OAI-SearchBot and PerplexityBot reach your public pages. It's the cheapest fix and the most often missed.
2. Business information that agrees with itself. Name, address, phone, hours and services should match across your website, Google Business Profile, Apple and Bing listings, Yelp and the review sites patients use. Conflicting facts give an engine a reason to pick someone else.
3. A page for each service, in plain sentences. "We offer Botox for forehead lines and crow's feet at our Heights clinic; treatments take about 15 minutes" is quotable. A gallery with a "Book now" button is not. Engines can only recommend you for what your pages say you do.
4. Who does the work. Named providers, their credentials and the medical director's name give an engine (and a patient) something concrete to trust.
5. Reviews and third-party mentions. Recent, honest reviews that mention specific treatments, and mentions in local press or credible directories, are evidence that exists outside your own site.

What not to over-invest in

Some popular advice outruns the evidence. Google's own guidance says you don't need new machine-readable files, AI text files or special markup to appear in its AI features. Structured data (schema) is still worth having because it states your facts unambiguously, and an llms.txt file is cheap to publish, but neither is a switch that makes an engine recommend you. Be wary of anyone selling guaranteed placements or "AI rankings." The engines don't sell them.

How to check your own clinic in five minutes

1. Open a temporary or logged-out chat in ChatGPT, Perplexity and Google (AI Mode).
2. Ask the question a new patient would: "best med spa for [service] in [your neighborhood]" or "best cosmetic dentist for veneers near [area]".
3. Write down every business named, in order, and the date.
4. Repeat for your two or three most profitable services.
5. Ask "[your clinic name] reviews" and check whether the facts it states about you are right.

If you'd rather have it done for you, our $497 AI Visibility Audit runs 20 patient questions across five engines, records who is named and why, and gives you a prioritized fix list.