Should You Trust AI to Pick Your Med Spa?

Type “best med spa near me” into ChatGPT and you get a confident, tidy list of three or four clinics with a sentence on each. It reads like advice from a friend who has done the research. It is worth understanding what that list actually is before you book on the strength of it.

A study published in September 2026 by Altaway Solutions, an AI-visibility agency in Utah, set out to answer exactly that. The researchers asked four AI assistants, Perplexity, ChatGPT, Gemini and Google AI Mode, to recommend a med spa in Salt Lake City, using four fixed prompts each, for sixteen answers in total. Then they tracked every source each assistant cited to build its answer. The full study is here, and it is a useful read if you have ever wondered where an AI’s recommendation comes from.

We have a small stake in it: ClinicCompass was one of the sources cited. More on that below, honestly.

What the study found

Three findings matter to a patient.

The assistants did not agree with each other. Across the sixteen answers, six clinics were recommended more than once, but the lists varied by engine and by prompt. The study’s own summary is that the engines “behave very differently.” The same question, asked of a different assistant, produced a different shortlist. If you have been treating an AI answer as the answer, that is the first thing to adjust.

The answers were built from third-party sources, mostly directories. The assistants cited 30 sources in total. Clinics’ own websites were the most-cited category. After that came directories and review platforms: MedSpaScout was cited five times, Medical Spa Locator four, Google Maps four, a local Salt Lake City directory three, Yelp twice, ClinicCompass twice, and one more directory twice. In other words, the assistant is not evaluating clinics. It is summarizing what other websites say about clinics, and then presenting the summary as a recommendation.

No single source dominated. The study notes that no source appeared in even 40 percent of the answers. The assistants were sampling widely and inconsistently, which is another way of saying the recommendation you get depends heavily on which pages the model happened to pull that day.

The authors are careful about the limits of their own work, and so should we be: AI answers “vary from run to run and by account history,” two of the four engines were tested in signed-in sessions, and the whole thing is “one dated snapshot.” It is a small, well-documented study of one city, not a verdict on AI search. Read it as a window into the mechanism, not a league table.

Our part in it, stated plainly

ClinicCompass was cited in two of the sixteen answers. Two other directories were cited more often. We are including this because the study is genuinely interesting and because it would be odd to write about it without saying where we appeared. We are not going to pretend two citations means AI depends on us, and nothing about how a clinic appears on this site, paid or otherwise, has any bearing on what an AI assistant chooses to cite. What the study does confirm is the thing this site is built around: when an assistant recommends a clinic, it is standing on directory pages. The quality of the directory is the quality of the advice.

Why the source matters more than the summary

Here is the practical problem with an AI shortlist. The assistant reads a directory page that says a clinic has a 4.9 rating, offers Botox and fillers, and is in a good part of town, and it turns that into “a highly rated clinic known for injectables.” That is a fair summary of the page. It is not an assessment of the clinic. The assistant did not check whether the 4.9 rests on 30 reviews or 900, whether the injector is a nurse practitioner or someone with a weekend certificate, whether the price quoted online is the price you will pay, or whether the medical director has set foot in the building this year.

Those are the questions that decide whether you get a good result, and they are not in any AI’s answer, because they are not on most directory pages either. So the useful way to use an AI recommendation is as a starting shortlist, then do the checking the assistant skipped.

The four things to check yourself

1. Who actually does the treatment. Botox, fillers, and laser treatments are medical procedures. In most states and provinces they must be performed by, or under the supervision of, a licensed medical professional, and the rules on who may inject vary by jurisdiction. Ask by name: who will treat me, what is their license, and who is the medical director? Our guide to who can legally inject Botox covers what the answers should sound like. A clinic that is vague about this is telling you something.

2. Review volume, not just the star. A 5.0 from 12 reviews and a 4.8 from 900 are not the same signal, and every ranking on this site weights the rating by how many reviews stand behind it for exactly that reason. Look at the count. Read the three-star reviews, which are usually the most informative. And check that the reviews are recent; a clinic that was excellent in 2022 may have changed hands since.

3. The local price range. Prices for the same treatment vary by city and by clinic, and a price far below the local range is a reason to ask why, not a bargain. Know the band before you walk in. For the most common first treatment, our Botox cost guide breaks down what a unit and a treatment cost by city, and every treatment page on this site shows the typical range.

4. A consultation before you pay for anything. Every reputable clinic will assess you, explain what they recommend and why, and quote the full plan before charging. If a clinic pushes you to buy a package on the first call, or quotes a price without seeing your face, keep looking. Our first consultation guide lists the questions to bring, and our med spa red flags piece covers the warning signs that show up in the first ten minutes.

One more habit worth building: ask the assistant where its answer came from. Perplexity shows its sources by default, and ChatGPT and Gemini will list them if you ask. If the recommendation traces back to a single directory page or a clinic’s own marketing copy, you know exactly how much weight it deserves.

Where to do the checking

If you are in Salt Lake City, the city the study looked at, our ranked list of Salt Lake City med spas shows every clinic’s Google rating alongside its review count, so the volume question answers itself, and each profile links to the clinic’s own site so you can find the medical director and the team. Every clinic on the list is ranked by the same rating-and-volume math, and that order is not for sale. For anywhere else, start from the cities index.

None of this means AI is useless for the search. It is a fast way to get a shortlist, and the shortlists are often reasonable. It means the shortlist is the beginning of the work, not the end of it. The assistant read the directories. You still need to read the clinic.