How Do Patients Find Doctors Through ChatGPT?

Patients ask AI assistants the way they would ask a friend: "who treats sciatica near me and takes my insurance?" The assistant answers from web pages it can find and read. Practices that state their services, conditions treated, locations, insurance, and providers in plain text, on pages that already rank in search, are the ones an assistant can recommend.

How many patients use AI to look for health care?

More every year. In KFF's June 2026 poll, 29% of US adults said they use AI tools or chatbots for health information and advice at least monthly, up from 17% in 2024. In August 2026, Pew Research Center found that 34% of Americans have used AI chatbots for at least one health or medical reason, including 15% who used one to help decide whether to go to a doctor.

OpenAI's own numbers are larger and global. It says more than 300 million people ask ChatGPT health-related questions every week, and its January 2026 report found that seven in ten healthcare conversations in ChatGPT in the US happen outside normal clinic hours.

No reliable survey yet measures how many patients use AI to choose a specific doctor. But patients are already asking, often while your office is closed.

What do AI assistants read when they recommend a practice?

Mostly the same pages search engines rank. As covered in why a business does not show up in ChatGPT, AI answers increasingly cite pages that already rank in organic search, and in healthcare the overlap is especially high. An assistant can only recommend a practice whose facts it can read as text: what the practice treats, where it is, who the providers are, and how to book.

That is why the work is mostly on the practice's own website, plus the profiles that repeat its facts, such as Google Business Profile.

What did one clinic change before patients found it on ChatGPT?

BioCure Health, a longevity telehealth clinic, had little organic traffic in late 2025: about 218 desktop visits a month. Cream Digital rebuilt its site with clear service pages, structured data, and an llms.txt file, and a blog system that has published more than 160 articles answering the questions patients ask.

After the rebuild, the clinic's founder described the result on a call: "we've been found on ChatGPT like four times already," with Google sending more patients too. That is one clinic's report, not a guarantee, but it shows the mechanism: clear pages that answer real questions get read and repeated.

What should a practice website say so AI assistants can recommend it?

Write down the facts a patient would ask a front desk, in plain text on the site:

• Services and conditions, each on its own page, in the words patients use
• Locations, with the full address and phone number, identical everywhere
• Insurance accepted, listed by plan name
• Providers, with credentials and what each one treats
• Hours and how to book, including telehealth if you offer it
• Answers to common questions, written as questions and answered in the first sentence

Ask Health shows the same pattern on two sites: a parent site with ten care-team bios, structured data, and llms.txt, and a patient site built to make no third-party requests, so outside services cannot track visitors.

How can you check whether ChatGPT recommends your practice?

Ask it, the way patients would, every month. Try a dozen versions: the condition plus your city, your specialty plus "accepts" and a plan name, and your practice name directly. Note whether you appear and which pages are cited. Then check the basics: that your robots.txt does not block search crawlers, and that your key pages rank in Google. If you want a full review, Cream Digital's lead generation operations includes building practice websites for AI search.