Why patients cannot find good doctors: the case for physician visibility in 2026
Patients behavior has changed. Fast.
In 2026, 47% of patients say they have used an AI tool to research a healthcare provider. That is up from 31% a year earlier, according to rater8's 2026 Patient Choice Report, a survey of 992 adult patients in the United States.
Among patients actively searching for a new doctor, AI tools now influence the decision more than Google search results or a referral from another physician.
The explanation sits in how those systems work.
An AI tool can only assemble an answer from what already exists online, and for most excellent clinicians... very little exists.
When a patient asks ChatGPT which surgeon to see, the model does not evaluate surgical outcomes. It assembles a recommendation from directory listings, review platforms, practice websites, hospital profiles, and social media. A physician with fifteen years of exceptional results and no digital footprint is functionally invisible to the tool making the introduction.
Google was the phone book... now, AI is the concierge.
Happening next week: the AI Referral Challenge, September 21 to 25
Learn how to make AI tools a top referral source.
Everything in this article is what we work through together in a free five-day challenge that starts Monday. We kick off live on September 21 at 5:00 PM CT, and then a short lesson and a daily challenge land each day through Friday. Registration is free, and you keep access for a full year. Limited VIP options available.
How patients actually find a doctor in 2026
The search begins online for nearly everyone.
rater8 found that 72% of patients had actively searched for or chosen a new provider in the previous twelve months, and 55% of those searches started with a local query like "[specialty] near me." From there the research spreads across several channels at once, which is where the meaningful change has happened.
Among patients actively searching, 36% cited AI tools such as ChatGPT and Claude as a top influence on their choice, ahead of Google search results at 34% and ahead of doctor referrals at 32%. Among patients who actually switched providers, AI was the single largest influence at 39%. Those figures come from a base of 714 searchers and 463 switchers within the same survey. They describe people in the middle of an active decision, not the general population.
The pattern holds inside Google as well. When rater8 asked patients which section of a Google results page they trust most when researching care, 37% chose the AI Overview at the top, 20% chose the organic blue links, 13% chose the local map pack, and 7% chose sponsored ads. Practices are generally spending the most on the channel patients trust the least, and almost nothing on the channel patients trust the most.
What percentage of patients use AI to find a doctor?
Nearly half. rater8's 2026 survey of 992 patients found that 47% have used an AI tool to research or find a healthcare provider, an increase of 16 percentage points in a single year. Adoption is highest among patients aged 45 to 60, where 64% report having used AI this way, which undercuts the common assumption that this is a behavior belonging to younger patients.
Pew Research Center's separate and more conservative measure points the same direction. In a survey of 3,488 U.S. adults fielded in June 2026, Pew found that 34% had used an AI chatbot for at least one health-related reason, including figuring out symptoms, learning about treatments, and deciding whether to see a doctor at all. The two studies ask different questions of different populations, and both show the same rapid climb.
The verification layer: what happens between the search and the booking
Very few patients discover a physician on Instagram and book an appointment the same afternoon. What happens instead is verification.
A patient gets a name from a referral, an insurance directory, or an AI answer, and then goes looking for evidence that the name belongs to a real, current, competent human being. That check is where a social presence earns is vital, and it is also why follower counts describe almost nothing useful about whether the channel is working. Social media is a credibility tool. Is this human who they say they are? Are they an expert? Can they help me?
Pew Research Center, in a survey of 5,111 U.S. adults fielded in October 2025, found that 36% of Americans get health information from social media at least sometimes, rising to 52% among adults under 30. More telling for the way patients actually behave, 61% of the people who get health information from AI chatbots also get it from social media. That is the same patient, moving between both channels, inside the same research process.
The reach makes opting out difficult. Pew's 2025 survey of social media use, covering 5,022 U.S. adults, put YouTube at 84% of adults, Facebook at 71%, Instagram at 50%, and TikTok at 37%. Whatever a physician's patient population looks like, it already holds accounts somewhere a practice can be evaluated.
Absence communicates something in this environment. A profile that has not posted in fourteen months tells a patient that nobody is staying on top of the practice's reputation, and that impression travels to the practice as a whole.
Do patients check a doctor's social media before booking?
Many do, as part of a broader verification habit that now includes reviews, directories, and AI answers. The bar they apply has risen sharply. rater8 found that 55% of patients have canceled or avoided booking with a provider because of what they read online, up 15 percentage points from 40% in 2025, and 75% will not book with a provider rated below 4.0 stars.
Patients also read for evidence of responsiveness. Two-thirds of patients, 66%, now say a provider's response to reviews directly influences their trust, which is up 24 percentage points year over year. Behavior a stranger can see counts for more here than any claim a practice makes about its own values.
Why AI gets your practice wrong, and why patients believe it anyway
This is the finding that should concern every clinician reading.
Among the 465 patients in rater8's survey who had used AI to research providers, 66% said they had encountered incorrect provider information from an AI tool, including wrong addresses, phone numbers, insurance details, and office hours. And 60% of AI users said they trusted the summary they received without verifying it anywhere else.
🤯🤯🤯
A 2024 study published in BMC Health Services Research examined 449,282 physicians listed across five major national insurer directories and found that address information was consistent across those directories only 16.5% to 27.9% of the time, and phone numbers only 16.0% to 27.4% of the time. The underlying record of American medicine is already contradictory, and AI tools are now reading that record aloud to patients with total confidence.
The question is no longer whether you want to build an online presence. Your presence already exists, assembled without your involvement, from data you have never audited, and it is being narrated to patients who will not check it. Correcting and controlling that material is the most basic form of patient care in a digital environment.
Why does AI give patients incorrect information about doctors?
Because AI tools assemble answers from the same sources that have always shaped online reputation, including directories, review sites, practice websites, third-party listings, and social profiles. When those sources disagree with each other, the model has no way to adjudicate, so it selects confidently and passes the error along. Physicians whose name, credentials, specialty, locations, and contact details match everywhere are the physicians AI describes correctly.
The credibility vacuum: patients want a credentialed voice and are not getting one
Here is the part of the data that should encourage clinicians rather than alarm them.
Patients are in these channels, and they do not believe most of what they find there. Among the Americans who get health information from social media, only 7% rate that information as highly accurate, while 47% say it is not too accurate or not accurate at all, according to Pew.
At the same time, Americans are explicit about what they want from a health information source. Pew found that 75% say it is extremely or very important that the source has medical training, 73% say the same about transparency regarding conflicts of interest, and 72% about being easy to understand.
Every week that a credentialed clinician stays quiet, that space gets filled by someone with better lighting, a viral hook, and half your qualifications.
What kind of health content do patients trust most?
Educational content delivered by a credentialed clinician, by a wide margin. Pew's data shows that 65% of people who get health information directly from healthcare providers rate that information as highly accurate, compared with 48% for major health websites and 7% for social media. Clinical credibility is the scarcest asset in the entire information environment, and most of the people holding it are spending it nowhere a patient can see.
Identity-first: why posting more does not fix invisibility
At this point the usual advice is to post more, and that advice fails reliably. I have worked with more than 200 physicians, and the ones who stall are almost never the ones who lacked a posting schedule. They are the ones who never decided what they were actually there to say.
Pew found that 59% of social media health information users say the information they encounter is not too personalized or not at all personalized, and 47% say the same about AI chatbot answers. Patients are getting generic, and generic is not a content problem that volume can solve. A physician who has not defined their own point of view will produce content indistinguishable from the feed it lands in, which means the algorithm has no reason to surface it and a patient has no reason to remember it.
Our whole theory at Atria Social is that identity leads to your contribution, which leads to your visibility.
Identity is the honest answer to who you are as a clinician and why you chose this work.
Contribution is the specific thing you understand about your patients that your colleagues do not say out loud.
Visibility is what happens naturally once those two are clear, because you finally have something to be visible about.
That sequence also happens to be what the machines reward. A clinician with a defined focus produces consistent, specific, semantically coherent content across their site, their profiles, and their video, and consistency is exactly what AI systems use to decide who you are and what to recommend you for.
What should a physician post about on social media?
Start with the questions you answer in the exam room every week, delivered in plain language. Pew found that 72% of Americans rank being easy to understand among the most important qualities in a health information source, and 41% of AI chatbot users already describe AI health answers as easy to understand, which sets the expectation your content has to meet.
A single 90-minute recording session that works through fifteen common patient questions will generate a quarter of short-form content.
What physician visibility actually requires
Visibility is a system rather than a posting habit, and the work sorts into four layers.
The first layer is your entity. Your name, credentials, specialty, practice address, and phone number must be identical across your Google Business Profile, the major physician directories, your hospital or practice page, and LinkedIn. Given what the BMC study found about directory consistency, this is unglamorous work with more leverage than anything else on the list.
The second layer is your reputation. Review volume, recency, rating, and your visible responses now function as a filter that determines whether a patient ever evaluates anything else about you. rater8 found that 68% of patients would leave a review if their doctor asked them to, so the willingness is there and largely untapped.
The third layer is your own content. Comprehensive pages on your site and social media that answer real patient questions, with the answer placed in the first third of the page and question-style headings throughout, are what AI systems read and cite. Pro tip: Add physician and FAQ schema so machines can parse your specialty, locations, and services without guessing.
The fourth layer is your voice, on the platforms where patients verify (Instagram, X, LinkedIn, and other social media platforms). This is the layer that turns an accurate listing into a person a patient wants to see, and it is the only layer nobody else can build for you.
How long does it take for a physician's online presence to bring in patients?
Expect six to nine months before attribution data shows meaningful appointment volume, with earlier signals appearing in profile visits, site traffic, and the quality of questions patients bring to a first appointment.
It's a long game. But so many people are just sitting by not realizing that these tools are spreading misinformation about them without their involvement. The sooner you get on top of this, the sooner AI can be a referral source for you. It might take months, but you're lightyears ahead of your peers already.
The most reliable measurement methods here are also the simplest. Ask new patients how they found you during intake, use a channel-specific phone number, tag your links consistently, and re-run a list of real search queries across ChatGPT, Perplexity, Google AI Mode, and Claude every month to see whether your name has entered the answers. In the AI Referral Challenge from September 21-25, 2026, we provide an AI Findability Audit, which helps you do just that for FREE.
Conclusion
Patients are researching harder than they ever have, across more channels, with higher standards and less patience.
They are asking machines that cannot assess clinical skill and can only repeat what they find.
They have told researchers plainly that they want to hear from people with medical training, and they are not hearing from them.
Heart-led physicians and clinicians who do this work will get found by the patients they are best suited to help, and the ones who wait will keep watching better-marketed colleagues with a fraction of their experience take those patients instead.
Where to start, and where to start this week
The AI Referral Challenge runs September 21 to 25, 2026, and it is free to join.
We open with a live kickoff on Monday, September 21 at 5:00 PM CT, where Amanda teaches Lesson 1 and take your questions. From Monday through Friday, a short self-paced lesson and a daily challenge arrive each day, so you can work through them between cases instead of blocking out an afternoon.
Across the five days we cover how AI finds and interprets information about clinicians, the positioning work that decides what you want tied to your name, and the specific properties to fix first across your website, your profiles, your reviews, and your content.
Clinical schedules are unpredictable, so finishing inside the five days is optional. You keep access to every lesson for one full year, and you can work through it whenever your schedule allows.
There is also a VIP upgrade at $197, which adds a private 60-minute strategy review with me in the two to three weeks after we finish, access to the AI Referral Challenge GPT built to help you apply the work, and a group coaching call the following week. VIP moves to $297 once the challenge begins on September 21.
Join the AI Referral Challenge
If you would rather start smaller, the Clinician Brand Archetype Quiz takes six questions and gives you your archetype, which is the identity work every piece of visibility rests on.
Take the Clinician Brand Archetype Quiz
Frequently asked questions
How many patients use AI to research doctors? 47% of patients report having used an AI tool to research or find a healthcare provider, according to rater8's 2026 Patient Choice Report of 992 U.S. patients, up from 31% in 2025. Usage peaks among patients aged 45 to 60 at 64%.
Is social media worth it for physicians who already have full schedules? Yes, for two reasons that have nothing to do with lead volume. A current, credible presence protects the patients already being referred to you, since those patients verify you online before booking. It also feeds the AI systems that now influence more than a third of provider decisions, because those systems read your profiles as evidence of who you are.
What is the most important first step for a physician with no online presence? Fix your entity data before you post anything. Make your name, credentials, specialty, address, and phone number identical across your Google Business Profile, the major directories, your practice or hospital page, and LinkedIn. Contradictory listings are the reason AI describes you incorrectly, and no amount of content fixes that.
Do physicians need to be on every platform? No, and attempting it is the fastest route to abandoning all of them. Choose one primary platform based on where your patient population actually is, using adoption data rather than assumption, and maintain a complete and accurate profile everywhere else. Pew's 2025 data shows YouTube and Facebook are the only platforms a majority of adults in every age group use.
Can a physician post educational content without running into compliance problems? Yes. General health education, explanations of conditions and procedures, provider credentials, facility content, and consented staff features require no patient authorization at all. The exposure comes from patient-identifying content, from replies that confirm someone received care, and from photographs that capture incidental information in the background.
What should a physician measure to know whether this is working? Track whether your name appears when real patient questions are asked of ChatGPT, Perplexity, Google AI Mode, and Claude, and re-run that query list monthly. Track review volume and recency. Ask every new patient how they found you at intake, because self-reported attribution survives any change in tracking technology.
SOURCES CITED IN THIS POST
- rater8, 2026 Patient Choice Report: How AI, Online Reviews, and Rising Standards Have Redrawn the Path to Care (published June 3, 2026; n=992 U.S. adult patients, with subsamples of 714 searchers, 463 switchers, and 465 AI users). https://rater8.com/2026-patient-choice-report/
- Pew Research Center, "Health information from social media and AI rated more convenient than accurate," April 7, 2026 (n=5,111 U.S. adults, fielded October 20 to 26, 2025). https://www.pewresearch.org/science/2026/04/07/users-of-social-media-and-ai-chatbots-for-health-information-are-more-likely-to-say-they-are-convenient-than-accurate/
- Pew Research Center, "What do Americans want from their health information sources?", April 7, 2026 (same sample). https://www.pewresearch.org/science/2026/04/07/what-do-americans-want-from-their-health-information-sources/
- Pew Research Center, "From Diagnoses to Treatments: Why Americans Use AI Chatbots for Health," August 25, 2026 (n=3,488 U.S. adults, fielded June 22 to 28, 2026). https://www.pewresearch.org/science/2026/08/25/from-diagnoses-to-treatments-why-americans-use-ai-chatbots-for-health/
- Pew Research Center, "Americans' Social Media Use 2025," November 20, 2025 (n=5,022 U.S. adults, fielded February 5 to June 18, 2025). https://www.pewresearch.org/internet/2025/11/20/americans-social-media-use-2025/
- Adler-Milstein et al., "Characterizing physician directory data quality: variation by specialty, state, and insurer," BMC Health Services Research, July 18, 2024 (449,282 physicians across five national insurer directories). https://link.springer.com/article/10.1186/s12913-024-11269-5
