Each AI engine builds a med spa recommendation from a different mix of sources. Yext's analysis of about 6.8 million citations found Gemini leans on first-party websites, ChatGPT leans on directory listings and Google Business Profiles, and about 86 percent of citations across ChatGPT, Gemini, and Perplexity came from sources a brand manages. A spa that fixes only one surface will show up in one engine and not the others.
AEO, GEO, AI visibility, and AI SEO are four names for the same work: getting a med spa named inside AI answers. This post uses AEO. It covers the engine-by-engine split. The ClinicAds post on sources AI assistants cite for med spas covers the other cut, which source types matter, and this one does not repeat that list. It also says plainly where the public evidence runs out, because no engine-level study of med spas exists.
- The engines do not share one source list: Yext's analysis of about 6.8 million citations (July to August 2025) found Gemini drew 52.2 percent of its citations from first-party websites, while ChatGPT drew 48.7 percent from listings and leaned on Google Business Profiles.
- Yext found about 86 percent of citations across ChatGPT, Gemini, and Perplexity came from sources a brand manages, such as its own site and directory listings, so most of the fixes sit inside the med spa's control.
- In Yext's healthcare sample, listings made up 52.6 percent of citations, with WebMD, Vitals, and industry directories the most visible names.
- No independent study found for this post splits med spa citations by engine, so a spa should log its own sources per engine; a source cited in one engine can be absent in another for the same treatment prompt.
Do ChatGPT, Perplexity, Gemini, and Google AI Overviews use the same sources for med spas?
No. The engines retrieve from overlapping but different pools, so the same treatment prompt can name three different spas in three engines. Yext's 2025 citation analysis, covering ChatGPT, Gemini, and Perplexity across retail, finance, healthcare, and food service, found the mix of websites, listings, and review sites shifted by engine and by industry.
The practical consequence for a med spa is that one source audit is not enough. A spa with a strong website and thin directory listings can score in Gemini and miss in ChatGPT, and the reverse is also true. The table below summarizes what the public evidence supports for each engine and labels what is ClinicAds observation rather than study data.
| Engine | What the evidence shows | Evidence type | First fix for a med spa |
|---|---|---|---|
| ChatGPT | Relies heavily on listings (48.7 percent of citations) and on Google Business Profiles | Yext, about 6.8M citations, Jul to Aug 2025, cross-industry | Complete every Google Business Profile service field and match name, address, and phone across directories |
| Gemini | Prefers first-party websites (52.2 percent of citations) | Yext, same sample | One page per treatment with a real price, injector name, and location |
| Perplexity | Included in Yext's 86 percent brand-managed finding; no separate share published | Yext, same sample; per-engine split not found | Log the numbered sources it shows under each answer and fix what repeats |
| Google AI Overviews | Not covered by Yext; Google's local data and its own index sit behind the answer | ClinicAds observation, not study data | Keep the Business Profile current and the treatment pages indexable |
Which sources does ChatGPT lean on when it recommends a med spa?
ChatGPT leans on listings and Google Business Profile data. In the Yext sample, listings supplied 48.7 percent of ChatGPT's citations, the largest single category for that engine. For a med spa, listings means the directories and platforms that carry a name, address, phone number, service list, and reviews.
ChatGPT also searches the web on only a minority of answers. Ahrefs' October 2026 study of 66.6 million logged-out responses found 14.4 percent triggered a web search, so a spa's long-run reputation across directories matters as much as any single page it publishes this month.
- Verify that the spa's name, address, and phone read identically on Google, Apple Maps, Yelp, and the booking platform
- List every treatment as a separate service on the Business Profile, not one line reading aesthetics
- Ask reviewers to name the treatment, since review text is the wording a ChatGPT answer can quote
Which sources does Gemini use for med spa recommendations?
Gemini favors a business's own website. Yext's analysis reported 52.2 percent of Gemini's citations went to first-party sites, higher than the other two engines it studied. That makes the treatment page the most valuable asset a spa controls for Gemini.
A treatment page earns that citation when it states the treatment, the price or price range, the injector or provider who performs it, the location, and the booking path in plain sentences. A page that says call for pricing gives Gemini nothing to quote, and the question gets answered from someone else's page.
What about Perplexity and Google AI Overviews?
The public evidence is thinner for these two. Yext counted Perplexity inside its 86 percent brand-managed finding but the coverage reviewed for this post gave no separate Perplexity share, and Yext did not study Google AI Overviews. No independent study found here splits med spa citations for either engine.
ClinicAds' working method fills the gap with direct reading. Perplexity prints numbered sources under every answer, so the source list can be logged run by run. AI Overviews attach linked sources to the summary on Google. In both cases the instruction is the same: run the prompt, record which domains appear, and treat repeats as the surfaces to fix first. That is an observation method, not a published benchmark.
How should a med spa audit sources engine by engine?
A med spa should log sources per engine for the same set of treatment prompts, because a combined list hides which engine each fix helps. Twenty prompts across the top three to five treatment lines is enough for a first read, repeated monthly because answers change between runs.
- Write 20 prompts: treatment plus city, price checks, injector shortlisting, and membership questions
- Run each prompt in ChatGPT, Perplexity, Gemini, and Google AI Overviews, logged out, with the date recorded
- For every answer, record whether the spa is named and every domain the engine cites
- Count which domains repeat per engine and mark each as owned, directory, review site, or competitor
- Fix the most repeated owned or listing surface first, then rerun the same 20 prompts 30 days later
What is the cost case for fixing the sources an engine uses?
The cost case rests on membership value rather than the first visit. At ClinicAds' planning range of $120 to $280 per member per month, a client retained six months is worth $720 to $1,680, against $28 to $80 to book the first appointment through paid media. A listing or treatment page fix that adds even a few booked appointments a month has a short payback. These are agency averages, not guarantees.
The same fix rarely pays in every engine. Because the evidence points to different leading surfaces by engine, a spa on a $3,000 to $8,000 monthly paid budget should spend the first AEO hours on the surface its own log shows repeating, not on a generic checklist.
Does a med spa need a different strategy for each AI engine?
Not a different strategy, but a different first fix. Treatment pages, a complete Business Profile, and consistent directory listings feed all of them. The engine-level split only decides which of those to do first, and a spa's own per-engine source log shows which.
Is there independent data on which sources AI engines cite for med spas specifically?
Not that was found for this post. The strongest public data is Yext's cross-industry analysis of about 6.8 million citations, which includes a healthcare sample. Yext sells location-data services, so its framing should be read as vendor research, and a spa should confirm the pattern in its own prompt log.
Which engine should a med spa check first?
ChatGPT and Google AI Overviews, because they reach the most clients, then Gemini and Perplexity. Whichever engine already names competitors for the spa's lead treatment is the one to log first.