How to get your med spa recommended by ChatGPT

A patient can ask ChatGPT for a med spa the same way they used to ask a friend. If your name is missing, the honest next step is not a trick. It is to make the public record of the practice accurate, consistent and easy for a crawler to read. This note is that work, for an owner in Austin, Dallas, Miami or another US city. It does not promise that any assistant will name you.

On October 18, 2026 we asked a logged-out ChatGPT one question per city, for Austin, Dallas and Miami. The answers named specific practices, pointed at directories, and in Dallas mentioned a local award in the prose. Review counts did not follow a simple rule. That run is written up in what ChatGPT said about Austin, Dallas and Miami. Use it as a picture of one afternoon, not as a ranking and not as a quota.

How the three products differ is in ChatGPT, Perplexity and Google AI Overviews.

Google Business Profile

Google’s guide to generative AI features on Search says a Google Business Profile can help a local business be visible in AI responses and in other Search results. Its page on AI Overviews and AI Mode lists an up to date Business Profile among ordinary SEO habits. Neither page says a finished profile produces a mention.

Read the profile as a new patient would. The name should be the name on the door. The address, phone, hours, website, services and photos should describe the practice you run today. If you moved, changed a number or dropped a treatment, the old detail is still what a model can repeat. Choose categories for the work you do.

Yelp, and the same facts everywhere else

In the October 18 captures, every Austin listing ChatGPT showed carried a yelp.com chip, and the Dallas listings did too. That does not prove Yelp caused the answer. It shows Yelp was one of the places the answer pointed. A thin Yelp page is a public page about you that you can correct.

Use the same name, address and phone as the Google Business Profile. Add hours, the treatments you offer and photos of the real rooms. If Yelp still lists a line you no longer answer, that is the number a patient, and an answer, can repeat.

Directories and the website footer drift in small ways: a suite number on one listing, a tracking phone on another. Pick one spelling of the name, one street address and one phone number. Walk Google, Yelp, the contact page and any directory you already joined, and change the ones that differ. Do not open empty new accounts just to be everywhere. Where a source chip in our captures was cut off, fix the profiles you can identify and do not guess the hidden name.

Reviews, without a magic count

That one ChatGPT run did not show a minimum that “works.” In Austin, the five named spas showed between 30 and 146 reviews in the listing quoted. The label was “reviews in the local business listing,” next to Yelp chips, not “Google reviews.” In Miami, the top pick was described with more than 2,300 Google reviews, and another named practice with more than 500 reviews. Dallas showed no review counts at all. A modest count was named. A very large count was named.

Ask patients who had a real visit for an honest review on the profile they already use. Do not buy reviews. Google’s optimization guide says chasing inauthentic mentions is less useful than it sounds, because quality systems still sit under its AI features.

A real local mention

The Dallas answer text said Sanjiva Medical Spa was recognized in the Dallas Observer’s 2024 Best of Dallas readers’ choice awards. That was one sentence, in one answer, on one day. It was not a source chip, and it is not evidence that awards are required.

If a local paper runs a real process and the practice belongs in it, take part, then cite the outlet, the year and the award name. Do not pay for a made-up badge. Google’s guide says seeking inauthentic mentions is not the shortcut it appears to be.

Service pages in sentences

Google’s AI features page says important content should be available as text, with no extra technical requirement beyond being indexed and eligible to show a snippet. The longer guide says ordinary SEO still applies, including people-first pages, and that a separate page for every tiny wording of a question can cross into scaled content abuse.

Write a page per real service: who does the work, what the visit is like, what you tell patients afterward, and how to book. Say which concerns you treat and which you refer. You do not need a separate page for every neighborhood and every synonym.

Put those facts in sentences. A photo of a price board is invisible to a reader that only has the text. If you publish prices, publish the ones you charge. Link each service from the homepage. Google lists internal links as a basic habit.

Crawler access

Three products, three switches. OpenAI’s crawler documentation says OAI-SearchBot surfaces websites in ChatGPT’s search features. A site that opts out of OAI-SearchBot will not be shown in ChatGPT search answers, though it can still appear as a navigational link. OpenAI recommends allowing OAI-SearchBot, and allowing its published IP ranges, if you want the site eligible for those results. A robots.txt change can take about 24 hours to be picked up. Eligible is not the same as named.

GPTBot is separate. OpenAI says it crawls content that may be used to train foundation models, and that the two settings are independent. You can allow search and refuse training. Read the page before you paste a robots.txt block from memory.

ChatGPT-User is separate. OpenAI says it visits when a person triggers a fetch, that robots.txt may not apply, and that it is not used to decide Search inclusion. Blocking it is not the same decision as blocking OAI-SearchBot.

Perplexity’s crawler documentation says PerplexityBot surfaces and links sites in Perplexity search results and is not used to train foundation models. They recommend allowing it, and they say a change can take up to 24 hours. Perplexity-User fetches a page when a person asks a question, and that page says this fetcher generally ignores robots.txt because a person requested the fetch.

Google’s AI features page says the crawl control for Google Search, including AI Overviews and AI Mode, is robots.txt for Googlebot. A page has to be indexed and eligible for a snippet before it can be a supporting link. Meeting that bar does not mean Google will crawl, index or show it. There are no additional technical requirements for those features.

If a firewall challenges unknown bots, a robots.txt allow is not enough. OpenAI and Perplexity publish IP ranges on those same pages. Allow the ranges for the bots you mean to admit, and check the live robots.txt on the public domain. A leftover disallow will drop the pages you just rewrote.

Files and markup Google says it ignores

Google’s optimization guide says you do not need a new machine-readable file, an AI text file, special markup or Markdown to appear in Google Search, including generative features, because Google Search does not use them. It says an llms.txt file neither helps nor harms visibility in Google Search. The AI features page says there is no special schema.org markup for AI Overviews or AI Mode.

Structured data can still help ordinary rich results, and Google says the markup should match the text a person sees. It is not a ticket into an AI answer. We publish Organization and FAQPage markup on Pixelmosa for that ordinary reason. We do not claim it gets a med spa named. An llms.txt file can be a tidy summary for tools that choose to read it. Treat it as optional housekeeping.

Free AI visibility check

The check asks one question for your city and reports the answer we saw. It does not promise a mention or a ranking.