Your next patient has already asked an AI assistant whether she is a candidate for the treatment you do best. She asked what it costs, how long the downtime is, what could go wrong, and whether a nurse injector or a surgeon should do it. By the time she fills in your form, the consult has already half happened, and you were not in the room.
That is the shift. AI assistants are absorbing the pre-consult, the long private stretch of questions a patient used to bring to your consultation suite. And they answer those questions by pulling from sites that actually explain procedures, give honest cost context and structure their service pages clearly. If your website is brand copy, a hero video and a booking button, you are not part of that answer.
What patients actually ask before they ever call you
Not "best med spa near me." That query still exists, and it still matters, but it is the end of the journey, not the start.
What happens earlier looks like this: I am 43, I have hollowing under my eyes and some laxity along my jaw, I do not want to look done, what are my options and what do they cost. Then: is a thread lift worth it or should I wait for a deep plane. Then: what questions should I ask at the consultation.
The pattern is consistent: patients use the chat window to get fast answers, to size up their options before deciding whether to see a provider at all, and to ask things they would rather not say out loud to a receptionist.
Read that last one twice. Aesthetic patients are asking about their own face, their body, their ageing, their budget and their vanity. Privacy is not a side benefit for this audience. It is the reason the chat window beats the phone call.
Why AI answers now sit where your consult used to
The behavior is not fringe and it is not slowing. Asking an AI assistant about a treatment has become a routine step, sitting alongside a search engine and a conversation with a clinician rather than replacing either.
So AI has not replaced you or Google. It has slotted in front of both.
The number that should change how you think about your website is the trust split. Among people who have used AI for physical health information, 69% trust it a great deal or a fair amount. Among people who have not used it, that drops to 18%. The patients doing this already believe the answer they get. They are not treating it as a first draft to verify with you. They are treating it as a briefing.
Which means the practice named in that briefing starts the relationship ahead, and every other practice starts behind, pitching against an answer the patient already accepts.
What makes a practice quotable
AI assistants cite sources that resolve a specific question cleanly. Luxury brand copy resolves nothing. "Artistry meets science" cannot be quoted in an answer about whether a patient with melasma should have that laser.
The practices that get pulled into answers tend to have three things on the page.
Procedure explanations written for the undecided. Not what the device is called, but what it does, who it suits, who it does not suit, what the recovery actually feels like day by day, what the realistic result is at three months, and what the alternatives are. The willingness to say "this is not right for you if" is the single strongest signal of a genuine clinical source, and almost nobody publishes it because it feels like turning patients away. It is not. It is pre-qualifying them.
Real pricing context. Not necessarily a price list. Context: what the range depends on, how many units or sessions a typical plan involves, why one practice charges more than another, what is included. A patient who can self-assess affordability before they book is a patient who turns up to the consult ready to proceed.
Structured service pages. One procedure per page, clear headings, consistent naming, a genuine question and answer section, location, and proof close to the claim. Scattered information across a single long services page gives an AI nothing specific to lift.
On our plastic surgery work we call this the Trust Stack: trust is built in layers, and patients commit when everything reinforces everything else. Miss a layer and confidence breaks down. That was true when the only audience was a human reading your site. It is now also true for the machine summarising you.
What AI cannot see on your site right now
Most premium aesthetic sites are optimised for the way a luxury brand looks, which is exactly how they should look, and that is precisely where the gap opens.
Common ones we find in marketing audits:
- Procedure detail locked inside a video or a carousel, with no text equivalent.
- Before and after proof presented as images only, with no written case context, no treatment named, no timeframe.
- Pricing referenced as "investment varies" and nothing else.
- A single Treatments page covering eleven procedures in two sentences each.
- Reviews living on third party profiles while the site itself carries no outcome evidence.
- An inconsistent practice name, address and phone across the website, Google Business Profile and directory listings, so the model cannot confidently tie the entity together.
None of these are design failures. They are translation failures. The site communicates beautifully to a human and says almost nothing to a system that reads text.
The honest numbers
Here is where most agency articles overclaim, so let us be straight with you.
Direct traffic from AI assistants is still small. A working paper studying 973 ecommerce sites with around $20 billion in combined revenue, covering August 2024 to July 2025, found ChatGPT referrals made up roughly 0.2% of sessions, about 200 times smaller than Google organic, with conversion rate and revenue per session from ChatGPT improving while average order value declined (Search Engine Land, October 2025). A separate Amsive study found no significant conversion difference between LLM and organic traffic, with LLM referrals under 1% of total sessions. A SALT.agency analysis found organic search visitors were more likely to engage or complete key actions across most sectors, with LLMs performing better only in a few niches such as health and careers. Vendor-published claims that AI visitors convert far better are marketing material. Treat them as contested. One more thing worth saying out loud: these numbers come from general-population US health polling and an ecommerce referral study, not from aesthetic practices, so read 0.2% as context, not a med spa benchmark.
So no, AI search visibility is not going to replace your paid campaigns next quarter.
The reason to do the work anyway is that the referral click is not the mechanism. The influence is. A patient reads an answer, forms a view on candidacy and cost, and then searches your name directly or walks into the consult already decided. That journey shows up in your analytics as direct traffic, or as nothing at all. The honest measurement is not a dashboard line. It is what patients say in the consult room: how specific their questions are, whether they arrive with a treatment already in mind, whether they mention a number before you do. Start logging that, and you will see the shift before any report does.
The part nobody budgets for
Two things get missed.
First, this is recurring work, not a one-time technical fix. Procedure pages go stale, protocols change, pricing drifts, proof ages, and models recrawl. A page written once and left alone decays quietly. The practices that stay cited are the ones producing new procedure content, new case documentation and fresh proof on a schedule, which is why ongoing content and recurring production shoots belong in the plan rather than in a launch sprint.
Second, your front desk and your clinicians now meet a patient who arrives pre-briefed, sometimes accurately, sometimes not. Someone has to handle "ChatGPT said I need X" without making the patient feel foolish and without validating a bad plan. That is a service design job, and it is where the consult is won or lost now.
One boundary worth naming. Aesthetic advertising sits under medical board and advertising rules that vary by state and by profession, and the requirements around outcome claims, before and after imagery and testimonials differ depending on who regulates you. Get your claims and imagery policy reviewed by someone qualified in your jurisdiction before you publish at volume. Clarity and compliance are not in conflict, but they do need to be designed together.
Prime Aesthetic Media works only with aesthetic practices, which is why our AI visibility work starts with your procedures, your proof and your consult, not a generic checklist. If you want to know what an assistant currently says about your practice and your competitors, that is a good first fifteen minutes.
Frequently asked questions
How many patients are actually using ChatGPT to research treatments?
Enough that you should assume it is happening before most consults. Asking an AI assistant about a treatment now sits alongside a search engine and a conversation with a clinician. It has not replaced search or your consult, it has moved in front of both.
Will AI search visibility actually send traffic to my website?
Not much yet, and we would rather say that plainly. An independent study of 973 ecommerce sites found ChatGPT referrals at roughly 0.2% of sessions, about 200 times smaller than Google organic, with other analyses putting AI referrals under 1%. The value is influence on the patient's decision before they search your name, not the referral click itself.
Do I have to publish my prices to get cited?
Not a full price list, but you do need cost context. What the range depends on, how many sessions or units a typical plan involves, and what is included. A patient who can self-assess affordability before booking arrives at the consult ready to proceed rather than surprised.
How do I know if this is working if the traffic does not show in analytics?
AI-influenced visits usually land in your analytics as direct traffic or not at all, so the dashboard will not tell you. Watch the consult instead: how specific the questions are, whether patients arrive with a treatment already chosen, and whether they mention a price before you do. Log it and the shift becomes visible before any report shows it.
Is there a compliance risk in publishing detailed procedure and outcome content?
There can be. Aesthetic advertising sits under medical board and advertising rules that vary by state and by profession, particularly around outcome claims, before and after imagery and testimonials. Have your claims and imagery policy reviewed by someone qualified in your jurisdiction before publishing at volume.
