Five Health AIs Launched in the Last Year. Only One Helps Patients Find a Clinic.

Quick answer: It's not just a ChatGPT thing anymore. Five consumer health AIs launched or announced in the last fourteen months — OpenAI's ChatGPT Health, Anthropic's Claude for Healthcare, Microsoft's Copilot Health, Amazon's One Medical assistant, and Alphabet's Verily Me. Four of them help patients understand their own records. One of them — Copilot Health — helps patients find a clinician, by specialty, location, and insurance. And it reads from a directory feed you've never heard of. If your listings are wrong there, you're the clinic that doesn't get shown.

Quick thing for anyone running an IVF, dental implant, LASIK, hair restoration, cosmetic derm or plastic surgery clinic.

Most people think patient AI use is still "someone asked ChatGPT about IVF." It's bigger than that now. OpenAI launched ChatGPT Health in January. Anthropic launched Claude for Healthcare four days later. Microsoft added Copilot Health in March. Amazon's One Medical assistant went to all members in January. Verily — Alphabet's precision-health company, not Google Search — launched Verily Me last fall.

All of them are built around the same idea: patients bring their questions to AI before they bring them to a clinic. And at least one of them is also built to help people find clinicians, based on insurance and location.

So what does AI see when it looks at your practice? It doesn't only read your website. And the consult you never hear about is the one you lose.

• Five consumer health AIs launched in ~14 months: ChatGPT Health, Claude for Healthcare, Copilot Health, Amazon One Medical's assistant, and Verily Me. All are records-grounded companions — patients connect their own data and ask about themselves.
• Exactly one helps patients search outside its own walls for a clinician: Microsoft's Copilot Health, which filters real-time US provider directories by specialty, location, language and insurance.
• Copilot Health's provider search is powered by a data partner (H1 Ribbon), and HealthEx — a network of 50,000+ provider organizations — feeds records connections to both Microsoft and Anthropic. Your directory data is now an AI input, whether you signed up for that or not.
• Microsoft's own docs say care navigation is "not a recommender system" — no personalization, no algorithm learning your patients. But it does return named clinicians from directory data, and a real-world test produced a named shortlist. Being findable in that data is the new table stakes.
• Check this week: your listings match everywhere, your insurance is listed correctly (that's a filter field), departed providers are gone, and your services are described the same way on every directory.

The five, in one place

A JMIR overview of the consumer health AI wave covers all five launches in one citable place. Here's the practice-side version:

The pattern in that last column is the whole story. Four of these products are built to help a patient understand their data. One is built to help a patient find you. And it's the one run by the company that also sells Microsoft 365 to your front desk.

The nuance, stated carefully

Two things are true at once, and most coverage only prints one.

These products are mostly not recommendation engines. ChatGPT Health is explicitly "not intended for diagnosis or treatment" — it's a place to make sense of your own records. Amazon's assistant books you into One Medical's network because that's Amazon's business. Verily pairs its AI with Verily's own clinicians. Nobody is paying for placement inside these answers, because there's nothing to buy.

But Copilot Health does return named doctors. Microsoft describes care navigation as searching real-time US provider directories — and says, in writing, it is "not a recommender system": no personalization, no learning from user behavior. All options presented impartially. A hands-on test asked it for an ENT in Manhattan matching an insurance plan and got named doctors, one highlighted. Impartial doesn't mean random — it means the data decides who shows up.

And the volume behind all of this is not hypothetical: OpenAI says people ask ChatGPT health-related questions 300 million times a week, and Microsoft's consumer products alone field 50 million health questions a day. The health tabs are where those questions go when patients want privacy. The general assistants — ChatGPT, Gemini, Perplexity — are still where your next patient asks which clinic to trust.

What AI actually reads about you

Here's the part that should change how you think about your Google Business Profile.

Copilot Health doesn't maintain its own map of American medicine. Microsoft's support documentation names the source: provider data sourced from H1 Ribbon, a healthcare data company that maintains clinician directory information. And HealthEx — the partner powering record connections for both Microsoft and Anthropic — covers more than 50,000 US provider organizations.

So the chain looks like this: a directory feed gets your listing. Copilot Health searches it when a patient filters by their insurance plan and your city. If your phone number differs on one directory, a doctor who left two years ago is still listed on another, your insurance isn't updated, or your services are described three different ways — you're the practice that gets filtered out. Not ranked lower. Absent.

You'd never know. There's no report card that says "a patient asked an AI where to get implants, and your name never came up." The consult just never happens. This is the same cross-directory consistency problem I've written about for AI visibility generally — the new part is that a major AI product now documents exactly which kind of data it searches.

What to check this week

You don't need an agency for the first pass. You need an hour and a checklist:

• NAP consistency — name, address, phone identical across Google Business Profile, Healthgrades, Zocdoc, WebMD, and every specialty directory that covers your field. One canonical version everywhere.
• Insurance listed, correctly, everywhere — Copilot Health filters on insurance participation. An out-of-date plan list isn't a minor metadata error; it's a filter you're failing.
• Departed providers purged — old doctors still listed dilute every entity signal around your practice, and waste a patient's click when they land on a bio for someone who left.
• Services described the same way — "dental implants" and "implant dentistry" on two directories is two entities to a machine. Pick the phrase patients use and repeat it.
• Your own site agrees with the directories — AI cross-references. When the website, the profile, and the directory disagree, the machine picks one. Make sure it isn't the wrong one.

Want the fuller version of this check — tested against actual AI answers, not just listings? Our 20-minute visibility walkthrough covers it.

One honest caveat

Nobody can get you "into" Copilot Health's results by promise or payment — Microsoft says the search is impartial, and the underlying directory feed isn't something an agency can buy a slot in. Anyone who tells you otherwise is selling something. What you can control is whether the data it reads is accurate and consistent. That work — entity cleanup, directory consistency, structured data so every AI reading your site agrees with every directory — is the same work that makes you visible in ChatGPT, Gemini, and Perplexity. The health tabs didn't change the job. They raised the cost of getting it wrong.

If you're not sure where you stand, that's what the audit is for: per-provider, per-location, per-service visibility across the platforms patients actually use.

Frequently Asked Questions

What does Copilot Health's find-a-doctor feature mean for my clinic?

Copilot Health searches real-time US provider directories by specialty, location, language and insurance, and returns named clinicians. Microsoft documents the provider data as sourced from H1 Ribbon and states the search is impartial — no personalization or recommendation algorithm. For a clinic, that means your directory listings and insurance data decide whether you appear when a patient searches; there is no optimization that targets the algorithm itself, only the data it reads.

Does Copilot Health recommend clinics?

It helps patients find clinicians through directory search — Microsoft's docs explicitly describe care navigation as "not a recommender system," with no personalization or learned preferences. In practice it still returns named doctors matching a patient's filters (specialty, insurance, location), so being accurate and complete in the underlying directory data is what determines whether your practice appears.

Are ChatGPT Health and Claude for Healthcare recommending clinics to patients?

No. Both are records-grounded companions: patients connect their own medical records and ask about their own results, medications, and appointments. ChatGPT Health (OpenAI, launched January 2026) and Claude for Healthcare (Anthropic, launched days later) are explicitly positioned to support, not replace, care — neither searches the open market for providers. Amazon's One Medical assistant books within One Medical's network, and Verily Me operates inside Verily's own care program.

What are the new AI health apps, and which matter for a medical practice?

Five launched in roughly fourteen months: ChatGPT Health (OpenAI), Claude for Healthcare (Anthropic), Copilot Health (Microsoft), Amazon One Medical's Health AI assistant, and Verily Me (Alphabet's Verily). For practices, Copilot Health matters most today because it's the only one that helps patients search for outside clinicians — the others shape what patients understand before they walk in, not who they choose.

How do AI health assistants decide which doctors to show?

Copilot Health searches healthcare directory data (Microsoft names H1 Ribbon as its provider-data source) filtered by the patient's specialty, location, language, and insurance criteria. The practical lever for a clinic is data consistency: identical listings, current insurance participation, no departed providers, and matching service descriptions across Google Business Profile, Healthgrades, Zocdoc, WebMD and specialty directories.

How do I check what AI says about my practice today?

Ask the platforms patients use — ChatGPT, Gemini, Perplexity, Google AI Overviews — the questions your patients actually type, with your city and procedure in them, and see who gets named. Our 20-minute visibility walkthrough shows the exact prompts and what to look for. If the answer skips you or names a competitor, the fix starts with the consistency checks above.

Curious where your clinic stands? Claim Free Snapshot — I'll check how ChatGPT currently describes your practice before anyone talks retainers, or see how the audit works first.