We asked ChatGPT 155 times about Toronto med spas, then studied the winners. They share three things.
We asked ChatGPT 155 questions a Toronto med spa client would ask, and every answer named a business, about 5 per answer. One clinic appeared in 40 of the 155 answers; of the 742 med spas with a Google listing we hold, 19% were named at all. ChatGPT searched before every answer, so for med spa marketing the listing and website it reads are the work.
How many Toronto med spas does ChatGPT actually name?
Every one of the 155 answers named at least one business, about 5 per answer. That is the whole shelf. A customer who asks gets a short list, not a page of results, and a clinic is either on it or it is not.
Across all the answers, ChatGPT named 172 distinct businesses as proper entries, a card with a name and usually a link, and about 342 if we count the ones a second model read out of the running text, which we treat as an estimate. Against the 742 med spas with an open Google listing in our Toronto catalogue, that comes to 144 named at least once, or 19%, and 598 never named.
A further 75 businesses were named for which we could find no Google listing at all: largely injectors and physicians named under their own name, and brands named without a location. They are real recommendations a customer would act on, and they are invisible to any comparison built on listings, including this one.
Who are the winners, and is it the same clinics every time?
One clinic, Bar Beauty Medical, appeared in 40 of the 155 answers, roughly a quarter of everything ChatGPT said about med spas in Toronto that week. Tight Clinic and ID Cosmetic Clinic were the next most often named. Those three are the clinics ChatGPT recommends, and we name them for that reason.
The surprise is how long the tail is behind them. Because each answer names about 5 businesses and the names change with the question, the most-named clinic took only 4% of all the names ChatGPT gave, and the top three together 9%. There is a clear winner, and there is also room: nine names in ten were spread across the rest of the 172 clinics ChatGPT named.
Who it did not recommend is a count in this report, never a list: 598 clinics, some of them with more reviews and a higher rating than the typical clinic it did name. A chain named by brand counts as named at every location we hold, because a customer told the brand can go to any of them; that weights chains up in the named group.
How did ChatGPT find them?
By searching, every time: ChatGPT ran a web search before answering, and 0% of the answers came from memory alone. That matters for a newer clinic, because an assistant answering from memory can only name what it already knew; one that searches can find a page written last month.
Where it looked is harder to read. Of the pages ChatGPT cited, 8% were Google profile pages, 7% were directories, 3% review sites and 3% booking marketplaces; publishers (magazines, news) were 1%. The large remainder, 78%, our classifier could not sort into a type, and we are not going to guess at it here.
What we can say with confidence is who got cited for themselves. In 69% of the named med spas, at least one answer cited the clinic's own website as a source. The winners are not only being talked about elsewhere; their own pages are part of what ChatGPT reads.
What do the winners share?
Three things, and reviews come first: 62% of the named med spas have a hundred or more Google reviews, against 21% of the unnamed. The medians are 136 reviews to 34. Not rating: the unnamed clinics' typical rating is 5 stars, the named clinics' 4.9. A perfect score is what a short review history looks like.
Second, a Google listing that is finished. Almost every winner's listing links to its website (99%, against 74% of the rest), and 36% carry fifty or more photos, more than twice the unnamed share of 15%. Third, a website an assistant can read: 70% of the winners' sites clear seventy on our readiness check, against 48% of the rest.
We report what the winners have in common; we do not know what ChatGPT weighed, and the big clinics tend to have all three at once. The detail is in three reports on this hub, measured across all three Toronto beauty trades: Google profiles, the 962 websites, and what changes for a business under a hundred reviews.
If you are a Toronto med spa and want to know which list you are on, Radar asks these 155 questions and more every week and shows you where you were recommended, where you were mentioned and where a competitor was named instead. The profile and website checks say which of the three things you are missing. Plans from $99 USD a month.
See where you standCommon questions
How many questions did you ask, and were they about my clinic?+
155 questions in one published run on 20 October 2026, from our Toronto med spa prompt pack: the way a customer asks an assistant where to go, who is good for a concern, and what it costs. None of them named a clinic. A clinic is "named" when the answer brought it up on its own.
Is this Toronto or the GTA?+
The questions say Toronto. The 742 clinics are the open Google listings our catalogue holds for the Toronto market, and the assistant decides for itself what counts; a clinic in Vaughan or Richmond Hill can be named if ChatGPT thinks of it as Toronto.
Would another week, or another assistant, give the same winners?+
Not exactly. The same question asked twice can return different names, and a clinic named once in 155 answers may be noise; a clinic named in dozens is not. This run asked only ChatGPT, with web search on, and the assistants do not agree with each other, so these are ChatGPT's winners, not AI's. We will re-run monthly and add Gemini and Google's AI answers, and report how much the list moved.
Is this part of med spa SEO?+
It overlaps. The Google listing and the website that med spa SEO works on are what ChatGPT read here, so a clinic's SEO budget is the natural home for this work. The difference is the outcome: an answer names about 5 clinics, not a page of ten, and a clinic can rank on Google and still be absent from every answer.
How do I find out whether I was named?+
Radar asks these questions every week and records whether you were recommended, mentioned or absent, and who was named instead. That is the one product line in this report, and the price is in the box below.
cite0 ran and paid for this measurement. On 20 October 2026 we asked ChatGPT, with web search on, the 155 prompts in our Toronto med spa pack and read every answer: 155 answers from 1 published run of 1 assistant. A business rendered as an entry in the answer is counted as named; a business a second model read out of the prose is counted separately and flagged as an estimate. Shares of names are over every naming in every answer; the count of answers a clinic appeared in is over the 155 answers. We then compared the 742 med spas with an open Google listing in our Toronto catalogue against the answers: Google rating, review count, photos and completeness from the listing as bought, and the website through cite0's web readiness checks over 582 sites. Names in the prose vary between runs of the same question, so a single count is a snapshot, not a ranking. What it cannot tell you: why a clinic was named, what another assistant would say, or anything about the 75 businesses named without a findable listing.
- cite0 Radar — 155 ChatGPT answers collected for Toronto med spa prompts on 20 October 2026, and the Google listings and websites of 742 Toronto med spas — October 2026; who-was-named counts and source shares from the collected answers, the listing and website comparison from the data note behind the Toronto beauty report on this hub
- OpenAI, "Overview of OpenAI crawlers" — OpenAI's description of how ChatGPT fetches web pages when it searches, the mechanism behind the "how did it find them" section