AI Tools Are Now Ahead of Referrals in How Patients Choose a Doctor. Here's What Owners Need to Decide.
Somewhere in the last year, AI stopped being a channel healthcare businesses could treat as secondary. Rater8's 2026 Patient Choice Report, based on nearly 1,000 US patients, found that among people actively searching for a doctor, 36% cited AI tools like ChatGPT and Claude as an influence on their choice. Google search came in at 34%. Doctor recommendations came in at 32%.
That's the actual headline for anyone running a clinic, practice, or wellness business right now: AI didn't just catch up to referrals. Among patients who searched, it's now ahead of them. It's worth naming the trajectory too. An earlier 2025 version of this same research had AI, referrals, and reviews sitting within a few points of each other. A year later, AI has pulled in front. This isn't AI holding steady as a peer channel. It's AI still gaining ground.
The number that matters more
Here's where it gets more interesting than the stat itself. The same wave of research showing AI's rising influence on where patients look also shows patients don't fully believe what they find there. A national survey from the Annenberg Public Policy Center at the University of Pennsylvania found that while nearly two-thirds of Americans who've seen an AI-generated health summary call it at least somewhat reliable, only 8% call it very reliable. Roughly two in three still click through to a source they already trust to double check it.
So AI earns attention at a scale healthcare businesses can no longer ignore. It does not earn trust at anywhere near the same rate. Those are two different jobs, and most practices are only doing one of them on purpose.
That changes what the actual competition is. You're not just competing to be the practice AI mentions. You're competing to be the practice that holds up once someone clicks through to check.
What's actually changed, and what hasn't
A few things are genuinely different now, worth naming plainly instead of folding into a general sense that everything changed.
Getting found got more complicated. It's no longer one channel to watch. Roughly seven in ten patients now use or are open to using AI tools somewhere in how they research a provider, and about a third go straight to tools like ChatGPT or Google's AI Overviews to do it, according to a 2025 patient research report. A newer 2026 survey adds another layer: when asked which part of a Google results page they trust most for doctor research, patients now rank the AI Overview section highest, at 37%, ahead of standard organic listings at 20%. That sits on top of Google search and review sites, not instead of them. The practices struggling most aren't behind on one channel. They've stopped tracking how many channels there now are.
Sounding like everyone else got easier to do by accident. When every practice has access to the same AI tools generating the same competently written paragraph, a lot of healthcare content has started to read like it came from the same source. Patients can feel the difference between a practice's real voice and a template, even when they can't say what's off about it. Marketers have started calling this the sameness problem, and it's a real cost, not a style issue. It's what makes a practice forgettable at the exact moment it needed to be memorable.
Physicians are cautious adopters, and that instinct almost certainly extends past the exam room. The American Medical Association's 2026 physician survey found 81% of doctors now use AI in their practice, up from 38% three years ago, but 85% want to be directly involved in decisions about how it's used, and 88% say they need real safety and validation evidence before trusting it further. That data is about clinical AI, not marketing content, and I want to be precise about that rather than stretch a survey to say something it didn't measure. But it's the clearest picture available of how physicians think about handing judgment to a system, and it would be strange if that same instinct disappeared the moment the subject changed from a diagnosis to a headline with the practice's name on it.
The five decisions this actually forces
Once you separate the attention problem from the trust problem, the list gets short and specific.
What you automate: first drafts, message variations across channels, scheduling and formatting, monitoring reviews and mentions so nothing sits unanswered for a week, and competitive research. This is the volume work, and it's where speed helps without putting anything at risk.
What stays human, without exception: anything touching a clinical claim, a treatment detail, or medical accuracy needs a clinician's review before publication. Responses to reviews, especially the negative ones, because that's where a practice either shows who it actually is or confirms someone's worst guess. The story of why the practice exists, in the founder's or clinician's own words, not a summary of it. Final sign-off on anything published under the practice's name.
What you publish as original evidence, not recycled explanation: most healthcare content answers questions a hundred other sites already answered. The content that earns trust, and gets cited when someone clicks through, is specific to your practice: real outcomes, real process, real photos of the actual space and the actual team. It's also the hardest material for a competitor to copy, because it isn't generic.
How your website, reviews, and content function as one system instead of five separate projects: a patient forms an opinion across all of these in one continuous experience, usually inside a few minutes. If any piece is inconsistent with the others, that's where trust breaks, not because one piece failed on its own.
Where you're actually visible right now, versus where you assume you are: most owners have never checked what a patient sees when they ask an AI tool about their practice. Ranking on Google is something you can verify yourself. Whether ChatGPT or Perplexity mention you by name usually isn't, unless someone goes and checks. That's the starting point for everything above it.
Why this is a design problem before it's a marketing one
I've spent my career thinking about how people enter a space, decide what they think of it, and choose to stay or leave. That used to mean a lobby, a front desk, a first conversation. Now a meaningful share of that same decision happens inside a chat window before a patient ever calls the office. The mechanism changed. The underlying problem, giving someone a reason to trust you fast and a reason to stay once they do, didn't.
AI decides who gets a chance. Your website, your reviews, your team, and how you actually treat people once they arrive decide who stays. Practices that treat those as one connected system, rather than five separate marketing tasks, are the ones that will handle this well regardless of which AI tool patients are using a year from now.
If you're not sure what a patient sees when they ask an AI tool about your practice, that's usually the first thing worth checking. My team and I do this kind of audit for clinics and practices regularly. Happy to take a look at yours.
Common questions
Do AI recommendations really influence which doctor or dentist a patient chooses?
Yes, and the gap is widening. A 2026 patient survey found AI tools were cited as an influence by 36% of patients actively searching for a doctor, ahead of Google search at 34% and doctor recommendations at 32%.
Do patients trust AI-generated health information?
Only partially. Most patients who encounter AI health summaries call them somewhat reliable, but very few call them highly reliable, and most still verify with a trusted source before acting on it.
Should a healthcare or wellness practice let AI write its marketing content?
For first drafts, formatting, and routine content, yes. Anything involving a clinical claim, treatment detail, or the practice's own voice needs a clinician's or owner's review and sign-off before it goes out.
What should stay human in healthcare marketing?
Clinical accuracy review, responses to reviews, the founder's or clinician's own story, and final approval on anything published under the practice's name.
Is traditional SEO still worth investing in with AI search growing?
Yes. AI mentions get attention, but patients still largely verify through a website and reviews before booking. Without a strong website and review presence, an AI mention doesn't convert into an appointment.
About the Author
Amir Zinati is a Vancouver-based communication strategist and Co-founder of ITANIZ. His work focuses on communication strategy and community: how a business builds trust with its clients, patients, and members, and keeps it, especially as AI changes how people find and judge a business before they ever reach out. He writes about communication, community, and brand strategy in an AI-driven market.
ITANIZ is a boutique marketing agency based in Vancouver, BC, led by Amir Zinati and Maryam Sheik, working directly with small and mid-sized businesses on strategy, branding, marketing campaigns, and content.
Comments