Patients Now Ask ChatGPT Which Surgeon to See. Here Is How the Answer Gets Picked

When a patient asks ChatGPT, Perplexity, or Google's AI which surgeon to see, the engine does not consult a secret ranking. It cross-references what it can read: your website, your Google Business Profile, your reviews, directories, press mentions, and what other sites say about you, then composes an answer from the practices that show up consistently across those sources. A practice that is thin or contradictory across them simply does not get named.

The question has moved into a chat window

For twenty years, the consideration phase of an elective decision happened across ten open browser tabs: the practice sites, the review pages, a forum thread, a "best of" listicle. The patient did the cross-referencing themselves, and search engine optimization was the fight to be one of the tabs.

AI assistants collapse that work into a conversation. The patient asks a question in plain language, the engine does the cross-referencing, and the output is not a list of links to evaluate. It is a short list of names, often three to five, with reasons attached. Being in the answer is the new version of being on page one, except the answer has fewer slots and comes with the engine's implicit endorsement.

Nobody outside these companies can quote the exact retrieval mechanics, and anyone who claims to guarantee you a spot in AI answers is selling something they do not control. That honesty matters, and it is our standing position. But the inputs these systems draw on are observable, the behavior is testable from the outside, and the pattern is consistent enough to act on.

What the engines cross-reference

Think of the AI engine as an extremely fast, moderately skeptical researcher who has to justify every name it outputs. It looks for the same things a careful human researcher would, and it can only work with what exists in text it can reach.

Your website, as a source it can actually read. The engine needs to establish, from your own pages, who the practice is, who the surgeons are, what procedures you perform, where you are, and what makes you credible. Pages that answer those questions in clear, extractable prose give the engine material to quote. Pages that bury everything in vague brand language, or render critical facts only in images and scripts, give it nothing to work with. A surgeon bio with credentials, board certifications, procedure focus, and years in practice, stated plainly, is exactly the kind of passage these systems lift.

Your Google Business Profile. For any "near me" or city-scoped question, the local layer matters enormously, and Google's AI experiences in particular draw on the same business data that powers Maps. Category accuracy, services listed, hours, photos, and the review stream all sit there. A neglected profile is a hole in precisely the data an engine reaches for first on a local question.

Reviews, in volume, recency, and text. Engines do not just register a star average. They read review text, which means the specifics patients mention, the procedures named, the way the practice responds, all become retrievable evidence. A practice whose reviews are numerous, recent, detailed, and answered gives the engine a rich, current picture. A practice with a handful of old reviews reads as dormant.

Directories and third-party profiles. Health directories, professional association listings, hospital affiliation pages, and local business listings do two jobs: they corroborate that the practice is real and established, and they provide the consistency check. When your name, address, specialties, and surgeons match across independent sources, the engine can merge them into one confident entity. When they contradict each other, confidence drops, and hesitant systems omit rather than gamble.

Independent mentions. Press coverage, local news, conference talks, published articles, podcast appearances, other sites referencing the practice. This is the layer practices control least and engines appear to weight meaningfully, for the same reason a human researcher does: anyone can praise themselves on their own website. Third parties doing it is evidence of a different grade.

The through-line: no single source decides it. The recommendation emerges from agreement across sources. That is why this discipline is not a trick you bolt on, but the accumulated surface of how findable and corroborated your practice actually is.

What the patient's journey actually looks like

An elective consideration inside an AI assistant is not one question. It is a sequence, and each step narrows the field. A realistic journey for, say, LASIK looks like this:

First, condition and options: "Is LASIK safe for someone with my prescription range? What are the alternatives?" The engine answers educationally. If a practice's content is the kind that gets cited in educational answers, the brand can enter the conversation this early, which is the cheapest impression in the whole journey. This early-funnel visibility is a big part of why we treat content strategy and AI visibility as one discipline for specialties like vision correction.

Second, the shortlist: "Who are the best LASIK surgeons in Charlotte?" This is the moment of truth described above. Cross-referenced names come out; everyone else is invisible.

Third, the interrogation: "Tell me more about [practice name]. What do their reviews say? What do they charge? Any complaints?" Note what happened: the AI is now the patient's researcher on you specifically, summarizing your reviews and coverage whether you participate or not. The material it summarizes is the material that exists.

Fourth, the decision support: "What questions should I ask at a consultation? How do I compare these two practices?" The engine hands the patient an evaluation framework, and the patient walks into your consultation carrying it.

The practices that treat this as a checkout lane misread it. It is a consideration journey, the same one patients always ran, compressed and delegated. Which means the fundamentals that always decided consideration, credibility, consistency, and evidence of real outcomes, are still what decides it. The engine just audits them faster than a patient ever did.

The self-audit: run this tonight

You can test your own visibility in under an hour, with no tools beyond the assistants themselves. Open ChatGPT, Perplexity, and Google's AI Mode (or whichever AI surface Google is showing in your market), and run the same prompts in each. Use a browser where you are not signed in to anything associated with the practice, so you are seeing something closer to what a stranger sees.

Ask, substituting your procedure and city:

1. "What are the best [hair transplant / dental implant / LASIK] practices in [your city]?"
2. "Who are the top [procedure] surgeons near [your city]?"
3. "Tell me about [your practice name]. Is it reputable?"
4. "[Your practice name] reviews and complaints."
5. "How much does [procedure] cost in [your city], and who should I consider?"
6. "I'm comparing [your practice] and [a competitor a patient would realistically compare you to]. How do they differ?"

Then read the results the way an auditor would, not the way an owner hopes to:

Are you named at all in prompts one and two? If not, note who is, and look them up. You will usually find they are the practices with the deepest review streams, the most complete profiles, and the most third-party mentions. That is the gap made visible.

In prompt three, is what the engine says about you accurate? Wrong surgeons, outdated locations, procedures you no longer perform, all of these trace back to stale or contradictory sources you can fix.

In prompt four, what evidence is it drawing on? Where the engine shows citations, as Perplexity typically does, read which sources it leaned on. Those citations are a map of which third-party surfaces currently define your reputation to a machine.

In prompts five and six, whose framing wins? If the comparison describes your competitor with specifics and you with generic filler, the engine is telling you whose public evidence is richer.

Run the whole set again in a month. The answers move as the underlying sources move, and the deltas tell you whether your work is landing.

What to do with what you find

The fixes follow directly from the mechanism. Make the website answer the researcher's questions in plain extractable text. Bring the Google Business Profile to full completeness and keep it moving. Build the review stream deliberately and respond to it. Reconcile every directory and listing until the story is identical everywhere. Earn independent mentions, because corroboration is the currency. None of this is exotic, but all of it compounds, and the practices showing up in AI answers today are largely the ones that started earlier.

Two honest cautions. First, this channel rewards the same substance the rest of your marketing rewards, so it should never come at the expense of the fundamentals; a practice invisible to AI but excellent at converting the inquiries it already gets still wins against a visible practice that squanders them. Second, anyone promising guaranteed placement in AI answers is guaranteeing weather. What can be done honestly is to measure your visibility, fix the inputs the engines demonstrably read, and track movement, in writing, with definitions that do not shift between reports.

That measured version is exactly what our AI Search Visibility service does for elective practices: a baseline of how the major engines answer the questions your patients ask, the source-level gaps behind it, and the systematic work of closing them.