A med-spa owner posted a screenshot that’s been making the rounds in aesthetics circles this year. Her front desk asked a new patient the usual question — “How did you find us?” The answer: “I asked ChatGPT where to go for Botox and fillers.”
That’s the shift in one sentence. For a decade, aesthetic patients found their injector on Instagram. Now a growing share of them open ChatGPT, Perplexity, or Google’s AI answers and ask who to trust — and the American Med Spa Association put a name on it with its Medical Aesthetics AI Visibility Index 2026. Meanwhile the old playbook is fading: searches for “med spa marketing” are down about 70% year over year, because the search is quietly moving into the chatbot.
Here’s the catch nobody in the “50 ChatGPT prompts for med spas” crowd leads with: you run a medical business. The moment you paste a patient’s photo, name, or chart into consumer ChatGPT, you may have a HIPAA problem. So before the get-found-in-AI playbook, we have to draw the line you can’t cross — because getting recommended by ChatGPT is worthless if it costs you a privacy violation.
What changed: patients are asking AI who to trust for Botox
Aesthetic medicine is unusually AI-exposed for a simple reason: it’s high-stakes, high-cost, and full of people who research obsessively before letting someone put a needle in their face. AmSpa-linked data pegs it plainly — a large majority of med-spa patients research treatments online before they ever book a consult, and provider credentials and experience are rated “very important” by nearly nine in ten (American Med Spa Association).
When those people ask an AI “who’s the best for Botox near me,” the AI doesn’t look at your reel aesthetic. It looks at structured, verifiable facts — your Google Business Profile, review marketplaces like RealSelf and Healthgrades, named practitioner credentials, and clear treatment information. The Visibility Index found citations concentrate heavily on a handful of well-optimized brands, which means the unoptimized clinic — however good its actual work — is simply invisible in the answer (Haute MD / AmSpa AI Visibility Index 2026). As one marketer put it bluntly online: “the window for being the default answer is closing,” and “most healthcare practices treat AI search like it doesn’t exist yet.”
Good news: you don’t need a $2,000-a-month agency to fix that. You can do the core of it in an afternoon with free ChatGPT. But only if you keep patient data out of it.
First, the line you can’t cross: is ChatGPT HIPAA-compliant?
Short answer: the public, consumer version of ChatGPT should be treated as not HIPAA-compliant. HIPAA requires that any vendor handling protected health information (PHI) on your behalf sign a Business Associate Agreement (BAA) and implement specific safeguards. Consumer ChatGPT doesn’t offer a BAA, so pasting identifiable patient information into it risks a violation (HHS guidance on Business Associates; peer-reviewed analysis of LLMs in healthcare, PMC).
PHI, in HHS’s definition, is any individually identifiable health information — anything that relates to a person’s care and could reasonably be used to identify them. In a med spa, that net is wider than people assume. It’s not just names. It’s faces, dates, procedures, and combinations of details that add up to one specific patient.
So here’s the practical rule: do not paste anything that lives in your chart, your EMR, your photo gallery, or your billing system into consumer ChatGPT.
The 5 things you must never paste
Names and contact details. Patient names, initials, phone numbers, emails, home or work addresses — anything that ties a message to a real person’s treatment. “Draft a follow-up for Sarah Jenkins about her lip filler” is a leak. Strip the name.
Facial photos, selfies, and before-and-afters. HHS explicitly lists full-face images as identifiers. A before-and-after — even a great one you’re proud of — is PHI if it relates to care. Cropping the eyes or blurring doesn’t reliably fix it: a unique tattoo, a scar, or a recognizable background can still identify the person. Your patient photos never go in the chatbot, full stop.
Charts, EMR screenshots, and clinical notes. Injection maps, units administered, vitals, consent forms, progress notes — any export or screenshot from your charting system. This holds even if you delete the name, because the date, visit pattern, and location can re-identify someone.
Appointment and billing details tied to treatment. Appointment times, invoices, the last four digits of a card, membership IDs — combined with what the treatment was, these can constitute PHI even without a name attached.
Highly specific “story-style” cases. This is the sneaky one. A free-text description like “41-year-old teacher in [your town] who had a bruising complication after off-label filler on May 12” has no name — and is still re-identifiable, and still PHI. If the details could point to one real person, don’t paste them.
What’s actually safe to use it for
The line isn’t “never touch ChatGPT.” It’s “never feed it a real patient.” De-identified and generic work is fine — and it’s most of the value anyway (PMC healthcare-LLM guidance):
- Patient-education content. “Explain what a neuromodulator does, in plain language, for a first-time Botox patient.” No real patient involved. Great for your website and FAQs.
- De-identified scenarios. Strip every identifier and generalize: “healthy woman in her 40s considering 20 units for mild glabellar lines — what questions should a consult cover?” That could describe thousands of people.
- Policy and SOP drafts. Ask it to draft a privacy-policy sample, a staff script, or a consent-language starting point — then have your healthcare attorney adapt it to your state and scope. AI writes the first draft; a human lawyer signs off.
- Rebooking, membership, and review messages. Templated, de-identified: “Write a warm 6-week check-in text inviting a member to book their next appointment.” You merge in the name yourself, in your own software, after ChatGPT is out of the loop.
- Marketing and operations brainstorming. Promo calendars, content topics, training outlines — no PHI required.
How to get recommended by ChatGPT for Botox (without an agency)
Once patient data is safely out of the picture, here’s the DIY version of the get-found-in-AI play. The whole game is giving AI the specific, verifiable facts it likes to quote:
- Write factual service pages. AI answers reward specifics over brand fluff. A line like “Botox is $12–$15 per unit, administered by a board-certified nurse practitioner with nine years of aesthetic experience” is exactly the kind of thing an AI will cite. Vague “premium, results-driven aesthetics” copy is invisible. Prompt: “Draft a factual service page for Botox at a med spa, including a plain-language explanation, a per-unit price range, who administers it and their credentials, and what to expect — no hype, no patient examples.”
- Name your injectors and medical director, with credentials. Medical supervision is a trust signal AI leans on. Put real names, titles, and years of experience on the site.
- Build a treatment FAQ in patient language. Answer the exact questions people type — “how long does filler last,” “does Botox hurt,” “what’s the difference between the toxin brands” — because those match what patients ask the AI.
- Claim and fill your Google Business Profile, RealSelf, and Healthgrades. These are the structured sources AI answer engines pull from most. Complete, consistent, current.
None of that requires a retainer. It requires an afternoon and the discipline to be specific.
What this means for you
- If you’re a solo owner or a single nurse injector: start with three factual service pages (your top treatments) and a named-credentials block. That alone moves you from invisible to citable. Do it before a competitor down the road does.
- If you run 2–3 locations: the leverage is consistency — same structured facts, same credentials, same FAQs across every location’s page and profile. AI rewards the clinic whose facts line up everywhere.
- If you’re the front desk or practice manager: you’ll be tempted to paste real patient messages to “just get help replying.” Don’t. Use de-identified templates and merge the real details in your own compliant system.
- If you’re an MD medical director: the compliance piece is your name on the line. A one-page “never paste this” checklist for your staff is worth more than any marketing tactic in this post.
What ChatGPT can’t do here
- It can’t make itself HIPAA-safe by you being careful. There’s no “private mode” that turns consumer ChatGPT into a BAA-covered vendor. Care reduces risk; it doesn’t remove the legal gap. For anything touching real PHI, you need HIPAA-compliant tooling, not consumer ChatGPT.
- It will state medical “facts” that are wrong. It can misstate a dosing range, a contraindication, or a claim about a product. Never publish clinical content it wrote without a qualified clinician checking every line.
- It can’t give you E-E-A-T you haven’t earned. Getting cited by AI rewards genuine credentials and real expertise. It will not manufacture authority for a page of filler content — and thin, generic pages can hurt as much as help.
- It can’t replace your lawyer. Draft policies and consent language with it; adapt and approve them with a healthcare attorney who knows your state.
The bottom line
Two things are true at once in aesthetics right now: patients really are asking ChatGPT who to trust for Botox, and consumer ChatGPT is genuinely unsafe for patient data. The winning move isn’t to pick a side — it’s to do both correctly. Keep every scrap of PHI out of the chatbot, and use it for the de-identified work: factual service pages, named credentials, patient-language FAQs, and the profiles AI actually cites. That’s how a solo owner becomes the default answer without hiring an agency — and without a compliance headache.
Want the full system? Our Get Found by AI Search course walks a local business through making ChatGPT recommend you, step by step, and Answer Engine Optimization for Small Business goes deep on the structured content AI answers pull from.
Sources
- AI Visibility Report Details How Patients Can Find Your Brand — American Med Spa Association
- Medical Aesthetics AI Visibility Index 2026 — Haute MD
- Business Associates — U.S. Department of Health & Human Services (HIPAA)
- Large language models in healthcare and the question of HIPAA compliance — PMC (peer-reviewed)
- Med Spas in AI Search: How to Get Found in ChatGPT & AI Overviews — Kelly WM
