How to get your practice recommended by ChatGPT

The short answer

ChatGPT recommends practices it can connect to a specific treatment and place, using sources it trusts. To get your practice recommended, make that connection obvious and consistent: let AI search crawlers reach your site, publish a clear page for each treatment, show who your clinicians are, keep your listings identical everywhere, earn real reviews and outside mentions, and test the questions patients actually ask, across the whole area you serve. Then retest, because answers change.

1. Test where you stand before you change anything

Start with the questions your patients ask, not the ones you would type. Patients ask about a treatment or a symptom, and usually add a place:

• “Who is the best [treatment] specialist near [town]?”
• “Where can I get [treatment] near me?”
• “Which clinic near [town] do you recommend for [symptom]?”
• “Is there a non-surgical option for [condition] near me?”

Ask each one in ChatGPT, and also in Google AI Overviews or AI Mode, Gemini, and Perplexity. Record which practices are named, in what order, how yours is described, and which sources are cited. Run each question more than once and from more than one town. A single screenshot proves very little; the pattern across repeated runs is what matters.

2. Make sure AI search can reach your website

When ChatGPT searches the web, it relies on its own crawler, OAI-SearchBot, alongside other search sources. If your robots.txt or firewall blocks that crawler, your pages may not be eligible to appear in its search answers. Check with whoever manages your site. Blocking GPTBot, the crawler used for model training, is a separate decision with its own trade-offs.

Also confirm that your site is indexed by Google and by Bing. Set up Google Search Console and Bing Webmaster Tools, submit a sitemap, and fix pages that cannot be crawled. Important details should be visible as text on the page, not only inside images or PDFs.

3. Publish one clear page for each priority treatment

A single “Services” page listing twenty treatments gives an AI assistant very little to work with. Give each treatment you want to be known for its own page that answers, in plain language: what it treats, who it is for, who performs it at your practice, where, what the visit involves, and how to book or get referred.

Use the names patients use as well as the clinical term, for example “uterine fibroid embolization (UFE), also called uterine artery embolization”. Clinical content should be written or reviewed by your physicians.

4. Show who your clinicians are

AI assistants and patients both look for evidence of expertise. Give each physician and advanced practice provider a profile with their specialty, board certification, training, the treatments they perform, the locations where they see patients, and hospital affiliations. Use the same name format everywhere. Remove or update providers who have left.

5. Make your practice details identical everywhere

Conflicting addresses, phone numbers, practice names, or hours make a practice harder to verify. Check your Google Business Profile, Apple Business Connect, Bing Places, Healthgrades, WebMD, Vitals, insurer directories, and hospital or health system listings. Choose the correct version of each detail and make every listing match it. Add matching structured data (schema markup) to your website so machines can read the same facts.

6. Earn reviews and outside mentions

An assistant is more likely to recommend a practice that other sources describe in the same way. Ask satisfied patients for reviews on the platforms they already use, steadily rather than in bursts, and never script or pay for them. When you reply, do not confirm that anyone is a patient or mention their care.

Beyond reviews, look for credible places that would naturally mention you: your hospital’s physician directory, medical society listings, local news coverage, community health talks, and referring physicians. Each one is independent evidence that you offer a treatment in a place.

7. Cover the whole area you serve

Patients travel for the right specialist, but AI answers often favor practices with an obvious connection to the town named in the question. Say clearly which towns and areas you serve, with directions, parking, and travel information where it helps. Avoid thin, copy-and-paste pages for every suburb; they rarely help and can make the site look less trustworthy. Then test your prompts from those towns, not just your own.

8. Retest on a schedule

Some fixes show up after your site or a listing is recrawled. New pages, reviews, and outside mentions take longer, and plan for months, not days. Repeat the same prompt set every few weeks and track how often you are named, how you are described, and which sources are cited. Keep what moves the numbers; fix what remains unclear.

What to avoid

• Anyone who promises a guaranteed or permanent ChatGPT recommendation. No one controls the answer.
• Fake, incentivized, or scripted reviews.
• Mass-produced pages for every town, or AI-written clinical content your physicians have not reviewed.
• Changing your practice name, address, or phone number in some places but not others.

Want us to do this for you?

These are the steps we run for practices, across every treatment and town that matters to them, so the practice gets recommended across the whole area it serves, not just in its own zip code. See AI visibility for medical practices and medical SEO for what the work includes. Steps 1 and 5 are where most practices find their biggest gaps; if you want to see yours first, our free report checks one treatment.