What patients see when they ask AI to recommend a doctor

When a patient asks ChatGPT or Google AI to recommend a surgeon, dentist or clinic, they get a short list of named providers, each with a one-line reason, built from your website, Google Business Profile, directories and reviews. If those sources are thin or disagree, your practice is left off the list, and the patient may book a consult elsewhere without ever seeing your site.

How many patients actually ask AI about their health?

About a third of adults, and the share is no longer small enough to ignore. KFF's tracking poll, fielded February 24 to March 2, 2026 among 1,343 US adults, found that 32% used AI chatbots for health information or advice in the past year, about the same share that use social media for health.

The reasons matter more for a practice than the headline. In the same KFF poll, 65% of AI users wanted quick answers and 41% were looking things up before deciding whether to see a provider. Another 18% said they had no regular doctor or could not get an appointment. Those are people at the top of your consult funnel, and the AI answer is the first thing they read about your specialty.

KFF did not ask how patients pick a specific surgeon or clinic, so no one should claim a precise share of procedures that start in ChatGPT. What the poll does show is that the research step your patient coordinators used to hear about ("I read a lot of reviews first") now often happens inside a chat window.

What does an AI provider recommendation look like?

Ask "who is a good board-certified plastic surgeon near me for a natural-looking rhinoplasty" and the answer is usually a short list of names rather than ten blue links. Each name tends to come with a short reason: credentials, a procedure focus, a review theme, sometimes a location. Answers differ between engines and from one day to the next, so treat any single answer as a snapshot.

Notice what the reasons are made of. They are facts the engine could read somewhere and confirm somewhere else: a certification on a bio page, a procedure on a directory listing, a pattern across reviews. An engine rarely repeats a claim it found in only one place, and it cannot repeat a claim it never found at all.

Where do ChatGPT and Google AI get their answers?

From the same public web your patients can see. Google's documentation says AI Overviews and AI Mode have no additional requirements and no special optimizations: a page needs to be indexed and eligible to show with a snippet. Google also says these features may use "query fan-out," running several related searches across subtopics, which is why one patient question can pull from your procedure page, a directory profile and a review site at once.

For anything near "near me," Google's local systems feed the answer too. Google's Business Profile help page says local results are mainly based on relevance, distance and prominence, and that more reviews and positive ratings can help your local ranking. The same page says there is no way to request or pay for a better local ranking.

Health is also held to a higher bar. Google's guidance on helpful content says its systems give even more weight to strong experience, expertise, authoritativeness and trust on "Your Money or Your Life" topics, which include health, and that trust is the most important of the four. In practice that rewards pages where it is obvious who wrote or reviewed them and what their credentials are.

ChatGPT search runs its own web retrieval and shows citations, but OpenAI publishes very little about how it ranks sources, so be wary of anyone who claims to know its exact formula. The safe assumption is the same: be easy to find, easy to verify and consistent everywhere.

Why are providers named instead of practices?

Patients ask for a person. For example, "Dr. Lee at Riverside Plastic Surgery" is a more confident answer than "Riverside Plastic Surgery," so engines tend to recommend the provider and mention the practice as where to find them. That puts the weight on individual profiles: each surgeon, dermatologist, dentist or injector needs the same name format, credentials and procedure list on your site, Google, Healthgrades, RealSelf and anywhere else they are listed.

Provider changes are the most common way practices lose answers. When someone leaves, AI keeps naming them until their bio, directory listings and review profiles change. When someone joins, they are invisible until those profiles exist and agree with each other.

Why does the procedure in the question change the answer?

Because engines match the procedure to the provider whose sources mention it most clearly. "Best dermatologist near me" and "best dermatologist for acne scarring" can return different practices, and so can "Botox" versus "lip filler" at the same med spa. A general specialty page tells the engine what you are; a procedure page tells it what you are good for.

The procedure pages that get quoted answer the questions patients bring to consults: who is a good candidate, what recovery looks like week by week, realistic results, risks, and what drives cost. Lead each section with a direct answer, then the detail. That is the paragraph an AI engine can lift.

What about reviews and compliance?

Reviews that name the provider and the procedure carry more information than "great experience," so it is fine to ask happy patients to mention what they had done and who treated them. The limits are firm. The FTC's final rule on fake reviews, announced August 14, 2024, bans fake and AI-written reviews, incentives conditioned on positive or negative sentiment, undisclosed reviews from staff or their relatives, and threats used to get negative reviews removed. It also lets the FTC seek civil penalties against knowing violators.

HIPAA adds a second limit: never reply to a review in a way that confirms the reviewer is a patient or discusses their care. And none of this work needs patient data. AI visibility is built on public pages, so keep tracking tools off login-gated and intake pages; our guide to HIPAA and tracking pixels in 2026 covers what your site can still run.

How to check where you stand

You can run a useful version of this test yourself in under an hour. Then fix what it shows, in this order:

1. Ask the patient's question. In ChatGPT, Perplexity and Google, type the question a patient would ask for each of your top procedures, with your city. Write down who is named, the reason given and any sources linked.
2. Search each provider by name. Note any profile with an old address, missing credentials, a former employer or a provider who no longer works with you.
3. Complete and verify your Google Business Profile. Google's help page lists verification, complete info (address, hours, categories), replying to reviews and adding photos as the steps that help local ranking.
4. Put credentials in plain text on every bio. Board certification, fellowship, affiliations and procedures performed, with a byline or "medically reviewed by" line on procedure pages. If you use structured data, make it match the visible text.
5. Give each high-value procedure its own page. Candidacy, recovery, results, risks and cost drivers, each section opening with a direct answer.
6. Bring directories into agreement. Same provider names, procedures, locations and hours on Healthgrades, RealSelf, Zocdoc, Vitals and specialty society listings.
7. Ask for specific, honest reviews. No incentives tied to a rating, no staff reviews without disclosure, and no replies that confirm patient status.

If you want the full picture, our free AI visibility audit runs your real procedure and provider questions across the major engines and shows where competitors are named instead of you. More on how we work with surgeons, dentists, dermatologists and med spas is on the medical practices page.

Frequently asked questions