From the Journal AI Search: AEO & GEO

How to Show Up When Patients Ask ChatGPT for a Medspa

A practical approach to ChatGPT discovery that separates search access from model training and turns real treatment details into useful, accessible pages.

How to Show Up When Patients Ask ChatGPT for a Medspa

At a glance

  • OAI-SearchBot serves search discovery; GPTBot addresses a separate model-training use case.
  • Publish verified treatment, location, and consultation information on accessible public pages.
  • Separate sampled ChatGPT answers, referral visits, booking clicks, and confirmed consultations.

Patients can use ChatGPT to compare local options before opening a clinic's website. An owner naturally wants the practice to be part of that conversation. The work starts by separating two things that often get mixed together: a model generating an answer and a search-enabled experience finding current information on the web.

This guide focuses on what a medspa can reasonably improve for search discovery. We cannot directly edit another company's answer or inspect every signal behind it. We can make the website accessible, the treatment information clear, and the business facts consistent. Those tasks sit within our AI visibility service and the wider approach in our medspa SEO management.

Know which experience you are observing

Start by recording whether the response is using search and whether it provides linked sources. A general response about choosing a medspa is different from a current local shortlist that draws on web results. Do not describe every answer as a fresh scan of your website.

Write down the request, date, known location context, named clinics, and linked pages. If the interface or search context is unclear, record that limitation. A useful observation log preserves what happened without guessing how the system produced it.

This matters when someone reports that a clinic is “in ChatGPT.” Was it named, linked, or simply present in a source the response used? Was the answer about an actual treatment and city? Those distinctions determine whether the next task is a technical access check, better local information, or a more useful treatment explanation.

Check search access separately from training preferences

OpenAI's crawler documentation identifies OAI-SearchBot as the crawler used to surface websites in ChatGPT's search features. It identifies GPTBot separately for content that may be used to train foundation models. The documentation states that these controls are independent.

Ask the person managing the website to review the actual public robots.txt rules and the hosting firewall. An allowed crawler in robots.txt can still encounter a server-level block, a bot challenge, or a login wall. Compare the rule with the current official documentation rather than copying an old crawler list from an unrelated blog.

  1. Confirm the intended access policy for public treatment and location pages.
  2. Review OAI-SearchBot rules and any matching hosting restrictions.
  3. Check that the important page returns useful public HTML rather than a challenge page.
  4. Keep private account, staff, and patient information protected.
  5. Document the search-access decision separately from the training-access decision.

A website owner does not need to make every file public to improve discovery. The objective is appropriate access to public business information, not the removal of privacy or security controls. Never expose a scheduling account or patient record in the name of AI visibility.

Give the system a clear page to discover

A navigation item called “Treatments” may contain many services but little specific information about any one of them. For a question about a particular service in a particular place, choose a page that explains that real offering. Avoid making an outside system reconstruct the service from a gallery, a pricing widget, and several social captions.

State what the service is, where it is available, and what a consultation involves. Identify the actual team appropriately and explain how to request an assessment. Use the practice's verified information, not assumed details about how every medspa operates.

Our one-treatment, one-page structure provides a practical content outline. It helps keep the commercial destination focused while educational articles answer adjacent questions. The service page should remain useful even if the reader arrives directly rather than through the homepage.

Use questions that match patient decisions

Patients rarely begin with the technical vocabulary used inside a marketing report. They may ask which local clinics offer a treatment, what to discuss during a consultation, or how two options differ. Use front-desk questions and real search data to identify the decisions the website should help them make.

Do not treat a long list of invented prompts as evidence of demand. Start with recurring questions the practice can substantiate. Group questions that need the same answer, and distinguish questions requiring different pages or qualified clinical review.

For each chosen question, write a clear initial answer and then the context needed to make it responsible. A question about suitability should explain the role of assessment instead of diagnosing the reader. A question about location should point to a real location, not an invented service area created solely to capture a city keyword.

Align the business facts patients will check elsewhere

A patient may move between the clinic's website, a local profile, a professional biography, and a review platform. Conflicting information makes that comparison harder. Use the actual practice name, correct contact information, current hours, and accurate service availability wherever the team legitimately controls a record.

Review one location at a time. Check that a link points to the relevant page and that the website does not imply every branch has an identical service menu. If a treatment is unavailable at a location, saying so is more useful than creating a superficially complete but misleading page.

Independent coverage and real reviews can contribute information about a business, but they should not be presented as a way to buy a specific AI answer. Do not create fabricated endorsements or ask writers to imply first-hand clinical experience they do not have.

Write for readers rather than a supposed secret citation format

There is no reason to replace a clear treatment page with an unnatural sequence of keyword variations. Use descriptive headings, readable paragraphs, and relevant links. Support important claims with appropriate sources and keep the clinical review process separate from the marketing approval process.

Structured data should describe facts that are also available to readers. A business entity, a service, and a set of visible questions have different roles. None should be filled with invented credentials, hidden location claims, or review ratings that the practice cannot substantiate.

Google's AI features guidance is useful for understanding Google's product, not as proof of OpenAI's private selection process. We keep these systems distinct. The shared practical foundation is a useful website that can be accessed and understood, not a claim that all AI products rank pages identically.

Measure referrals without turning them into imaginary bookings

Use analytics to examine identifiable referral traffic and the landing pages receiving it. Keep referral reporting alongside, rather than merged with, the manual answer log. Some sessions may have incomplete attribution, so absence from one report does not prove the patient never encountered an AI answer.

Next, distinguish visits from meaningful actions. A visitor reading a treatment page, clicking a consultation link, and completing a booking are separate stages. A click event can identify interest, but confirmation belongs in the actual scheduling or inquiry record where appropriate tracking is available.

Avoid sending names, treatment concerns, contact details, or other sensitive information into analytics event parameters. The team needs to understand the journey, not collect unnecessary personal information to make a channel report look more precise.

Run a focused first-month review

Choose one real service and a small set of local questions. Verify public access, strengthen the destination page, correct business information under the practice's control, and connect relevant internal pages. Keep a dated record of the changes so the next review has a clear starting point.

During subsequent checks, observe the same kinds of questions and examine the sources that appear. When information is wrong, correct the authoritative page or record you can actually change. When a question needs a better answer, improve the relevant section instead of generating dozens of repetitive posts.

The payoff is a stronger discovery path that the practice can maintain. Our AI visibility approach connects those checks to the broader SEO workstreams, while keeping observations, traffic, and business outcomes distinct.

Frequently Asked Questions

Which OpenAI crawler is relevant to ChatGPT Search?

OpenAI identifies OAI-SearchBot as its search crawler. Check both robots.txt and hosting or firewall access for that crawler using OpenAI's current documentation. Allowing access is an eligibility step, not a placement promise.

Must we allow GPTBot to appear in ChatGPT Search?

OpenAI documents independent controls for OAI-SearchBot and GPTBot. GPTBot relates to content that may be used for model training; it should not be confused with the search access decision.

Can we submit our medspa to a guaranteed ChatGPT recommendation list?

This workflow does not rely on a guaranteed recommendation submission. We focus on accessible pages, accurate business information, useful treatment answers, and observation of the sources a search-enabled response actually links.

Does a ChatGPT mention prove that the patient visited our website?

No. A mention, a linked source, a referral visit, a booking-link click, and a confirmed consultation are different events. Measure the stages separately and identify gaps rather than attributing every inquiry to AI.

Sources & citations

  1. Overview of OpenAI crawlers. OpenAI.
  2. Creating helpful, reliable, people-first content. Google Search Central.
  3. AI features and your website. Google Search Central.

Let’s talk about your search.

Tell us what your practice wants patients to find. We’ll look at your treatments, locations, and where search can help.

Book a Consultation