ChatGPT named zero of 200 independent medical practices when MGMA researchers ran about 4,950 patient-style prompts through the model in May 2026. That same year, 55.0% of medical groups told MGMA Stat they have no strategy at all for showing up when a patient asks an AI tool to find a doctor.
Patients are not waiting for practices to catch up. Nearly half now use AI to research a provider before booking, according to rater8's 2026 Patient Choice Report, and the share who trust an AI-generated answer over a plain search result keeps rising. For a practice administrator, these figures describe a new front door to the practice that most independent groups have not built yet.
The ChatGPT test that found nothing
MGMA drew 200 independent practices at random from the National Plan and Provider Enumeration System, spanning 18 specialties and 20 metro areas. Researchers ran roughly 4,950 patient-realistic prompts through the GPT-5.5 API in May 2026, asking the kind of question a real patient might type, such as which local practice treats a condition, accepts a specific insurance plan, or has an open appointment. Not one of the 200 sampled practices was both named by ChatGPT and cited as a source in the same answer.
The gap matters because patients are already turning to these tools for health questions generally, not just provider search. A KFF tracking poll of 1,343 U.S. adults, fielded online and by phone in English and Spanish from February 24 to March 2, 2026, found about a third had used an AI chatbot for health information or advice in the past year, including 29.0% for physical health questions and 16.0% for mental health questions.
Where the citations actually go
MGMA followed the practice-level test with a broader citation study, running about 35,000 queries and identifying roughly 18,000 distinct named entities across more than 33,000 individual ChatGPT citation events. The model mentioned a provider or organization about 41,000 times across nine recurring question patterns, but which specialties got mentioned varied widely by category.
Dermatology, ENT, and ophthalmology practices captured a 16.0% to 17.0% share of mentions within their categories, while plastic surgery, cardiology, and family medicine practices captured about 8.0%, roughly half as often. Mention share also varied by as much as three times across the 20 metro areas MGMA tested. Among the 516 practices that did get cited somewhere in the analysis, 66.1% had none of five basic on-site features MGMA checked for, including structured data that marks up a practice's services, locations, and providers in a format machines can read.
Patients are already asking the robots
rater8's 2026 Patient Choice Report, published June 3, 2026 from a survey of nearly 1,000 U.S. adults, found 47.0% had used AI to research or find a healthcare provider, up from 31.0% just nine months earlier. Among patients who actively searched for a doctor, 36.0% named an AI tool as an influence on their choice, ahead of Google search at 34.0% and a doctor's own referral at 32.0%.
Use skews older than many practices assume: 64.0% of patients age 45 to 60 said they had used AI to research a provider, compared with 28.0% of patients age 18 to 29. When rater8 asked which channel patients trust most for researching a doctor, 37.0% picked Google's AI Overviews, ahead of organic search results at 20.0%, the local map pack at 13.0%, and sponsored results at 7.0%.
That trust has a blind spot. rater8 found 66.0% of patients had seen an AI tool give incorrect information about a provider, yet 60.0% said they still trusted the summary without checking it further. Reviews still carry weight inside that trust equation: 55.0% of patients said they had walked away from at least one doctor because of online reviews, up 15 points from 2025, and 66.0% said a provider's response to a review directly shaped how much they trusted that provider, up 24 points from 2025.
What practices are actually doing
MGMA Stat's most recent poll on the topic, fielded September 15, 2026 among 251 practice leaders, found 17.0% pay for a listing or booking slot on a third-party platform such as Zocdoc or Healthgrades, while 77.0% do not and 6.0% were unsure. Many practices are instead adding staff to handle demand directly: 51.0% of groups said they added net-new advanced practice provider positions in the past year. New-patient wait times held roughly steady for 49.0% of practices and grew longer for 28.0%, and MGMA's DataDive benchmarking data puts the median third-next-available new-patient appointment at about two weeks across many specialties.
An earlier MGMA Stat poll of 218 practice leaders, fielded June 16, 2026, asked directly about AI search readiness: 55.0% had no strategy for improving how their practice shows up in AI search results, 35.0% had one in place, and 11.0% were unsure. The stakes for that gap are measurable outside healthcare too. Marketing analytics firm Seer Interactive has found that brands cited in Google's AI Overviews get about 35.0% more organic clicks than uncited competitors, while uncited brands saw organic traffic fall by roughly 65.0% year over year.
| Channel patients trust most when researching a doctor | Share of patients |
|---|---|
| Google AI Overviews | 37.0% |
| Organic search results | 20.0% |
| Local map pack | 13.0% |
| Sponsored search results | 7.0% |
Source: rater8, 2026 Patient Choice Report, June 3, 2026.
Only 35.0% of practices report having any plan for AI search visibility, and the one controlled test of independent practices run so far, MGMA's 200-practice sample, did not turn up a single practice that ChatGPT both named and cited.
Sources
- MGMA, "Third-party booking platforms: measure your capacity before buying demand," Sept. 17, 2026
- MGMA, "Can AI find your practice when patients ask?," June 2026
- MGMA, "Why ChatGPT has probably never named your practice," 2026
- rater8, "2026 Patient Choice Report," June 3, 2026
- KFF, "New KFF Tracking Poll on Health Information and Trust Finds One in Three Adults Have Used AI Chatbots for Health Advice," March 25, 2026
Talk to us
Fifteen minutes about your practice and the systems you run.