Key Takeaways
- Pew Research Center found 34% of US adults have used AI chatbots for at least one of eight health or medical reasons, in a survey of 3,488 adults fielded in June 2026.
- In KFF polling, 41% of adults who used AI for health called looking things up before deciding whether to see a provider a major reason, and 18% said they had no regular provider or could not get an appointment.
- An MGMA Stat poll of 218 practices found 55% have no strategy for AI search visibility, 35% have one and 11% are unsure.
- Across about 4,950 patient-style ChatGPT prompts covering 200 independent practices, MGMA found none that were both named and cited in the answers.
Healthcare marketing has spent a decade optimising for one moment: the search results page a patient sees after typing a symptom or a specialty into Google. A growing share of patients now skip that page. They describe the problem to a chatbot, read a synthesised answer, and arrive at a shortlist of providers without ever clicking a blue link. The survey evidence on how many people do this has firmed up through 2026, and the evidence on how practices are responding suggests most are not yet visible where the decision is starting to happen.
The Chatbot Has Joined the Care Journey
The broadest measure comes from Pew Research Center, which surveyed 3,488 US adults from June 22 to 28, 2026 and found that 34% have used AI chatbots for at least one of eight health or medical reasons. The most common was getting health information quickly (28% of adults), followed by figuring out what is causing symptoms (25%). Some 22% used chatbots to get information at little or no cost, the same share who used them to learn about treatments, and 15% used them to help decide whether to go to a doctor at all. Among those users, 47% found the information extremely or very helpful, while only 29% were extremely or very comfortable sharing personal health information with a chatbot.
Polling from KFF, fielded among 1,343 adults between February 24 and March 2, puts past-year use at 32% and places it on par with social media as a health source. Healthcare professionals (80%) and search engines (68%) still lead. But the reasons people give are the ones that matter to a provider's front door. Among AI health users, 41% called looking up information before deciding whether to see a provider a major reason, and 18% said they had no regular provider or could not get an appointment, rising to 38% among adults aged 18 to 29. More than half of physical health users, 58%, later followed up with a doctor. The chatbot is not replacing the visit; it is increasingly where the route to one begins.
Of US adults have used AI chatbots for at least one health or medical reason, according to Pew's June 2026 survey.
Named Is Not the Same as Cited
The route matters because AI answers do not draw on the same signals as a ranked results page. Research from reputation firm rater8, reported by MediaPost, surveyed 1,024 adults and found that 32% had used AI tools such as ChatGPT or Google's AI summaries for the first time while researching new providers in the past year. Some 26% said AI directly influenced their choice of provider, close to review sites (29%) and primary care referrals (28%). The information those AI users most wanted the tools to surface was verified patient reviews, at 39.9%.
On the supply side, MGMA reported in June that 55% of practices in a poll of 218 have no strategy for AI search visibility, against 35% that do. Its own testing explains why that gap is costly. Across about 4,950 patient-realistic ChatGPT prompts covering 200 independent practices sampled from the national provider registry, none were both named and cited. A follow-up analysis of more than 33,000 citation events found the tools leaning on hospital rosters, specialty-association credentialing pages, state medical boards, regional "Top Doctors" lists and practice-owned pages, with national medical directories a smaller share than many practices expect.
That is a different map from the one most healthcare SEO programmes were built on. A practice can rank well for its own name and still be absent when a patient asks for a recommendation, because the answer is assembled from third-party pages the practice has never treated as marketing assets. Press Ganey's Pranav Desai put the stakes bluntly in an interview with TechTarget: "if you're not seen by these AI engines, then you don't exist for the patients."
Of medical practices in an MGMA Stat poll have no strategy for appearing in AI search answers.
Visibility Meets Privacy and Cost
Two other findings shape what a sensible response looks like. The first is privacy. KFF found 77% of adults are concerned about the privacy of medical information given to AI tools, yet 41% of AI health users, equal to 13% of all adults, have uploaded personal medical information such as test results or doctor's notes. Patients arriving from an AI answer are often further into a sensitive decision than a casual searcher, which raises the bar for how their visits and bookings are tracked once they land on a provider's site.
The second is cost. Pew's 22% using chatbots for low-cost information and KFF's finding that 19% of AI health users could not afford a provider, rising to 32% in households earning under $40,000, both point to price as part of the question patients are asking. Practices that publish clear pricing, insurance and payment information on pages they own give AI tools something concrete to cite, and give a cost-conscious patient a reason to book rather than keep asking.
- Audit the answers, not just the rankings. Run the questions patients actually ask about your specialties and region through the main chatbots, and record whether your practice is named, cited, both or neither.
- Fix the third-party pages AI relies on. Hospital rosters, credentialing listings, state board records and regional awards pages are the citation sources MGMA identified; keep each one accurate and consistent with your own site.
- Make owned pages answerable. Provider bios, conditions treated, locations, insurance accepted and appointment availability should be written as plain, specific answers a model can quote directly.
- Treat reviews as AI input. Verified patient reviews were the information AI users most wanted surfaced, so review generation now feeds AI visibility as well as local search.
- Publish price and payment details. With cost a stated reason for chatbot use, transparent pricing, insurance and HSA or FSA payment options help both the AI answer and the booking decision.
- Measure the new path compliantly. Patients arriving from AI answers will not show up cleanly in last-click reports, and attributing them must not route health data to ad platforms that cannot receive it.
The patients most likely to start with a chatbot are the ones without a regular provider, which makes them exactly the new patients practices compete for. Pew and KFF show that behaviour is already mainstream; MGMA shows most practices have not adjusted. The work is less about a new channel than about tending the pages AI already trusts, and the practices that start now will be the ones named when a patient asks who to see.
