Is AI Showing Patients Wrong Info About Your Practice?

Here is a scenario that is playing out right now, out of sight, in every city in the country. A woman has a nagging shoulder that finally scared her enough to do something about it. She does not open Google and scroll through ten blue links like she did two years ago. She opens ChatGPT, or she reads the AI box that now sits on top of Google, and she types "good orthopedist near me who takes my insurance." She gets a clean, confident paragraph back. It names a couple of practices, lists hours, drops a phone number. She calls the number. It rings out to a fax line you disconnected a year ago. She does not try again. She goes back and calls the next name on the list.

You never saw that patient. There was no missed call in your log, no form, no trace. The AI simply handed her old, wrong information about your practice, she believed it, and she moved on. Multiply that by the number of people now researching care this way and you start to see the size of the problem.

How big is this, really

Bigger than most practice owners think, and growing fast. A 2026 survey from the healthcare reputation firm rater8, reported by TechTarget, found that 47 percent of patients used AI to search for a provider by mid 2026, up from 31 percent at the end of 2025. That is a jump of 16 points in nine months. Separate reporting on the same trend found that reliance on AI to actually choose a provider more than doubled in about a year, from 17 percent to 36 percent.

And do not assume this is a young person thing you can ignore. In the rater8 data, adoption skewed older. Among people aged 45 to 60, 64 percent said they had used AI to find a doctor, far more than the 28 percent of 18 to 29 year olds. The patients with the chronic conditions, the commercial insurance, and the higher lifetime value are the ones leaning hardest on these tools.

Now the part that should make you sit up. Of the respondents who had used AI to research a provider, two thirds said they encountered incorrect provider information, and the specific errors named were wrong office addresses, phone numbers, insurance details, and office hours. Worse, 60 percent said they trust AI summaries without verifying them. So the wrong answer is not a small problem the patient catches and corrects. Most of them read it, believe it, and act on it.

Why AI gets your practice wrong

It helps to understand that these tools are not looking you up in some official registry. They are stitching an answer together from whatever they can find about you across the open web, then stating it with total confidence whether it is right or not. Your practice does not have one clean record online. It has dozens of scattered mentions, and a lot of them are stale:

• An old address on a health directory from before you moved.
• Holiday or summer hours you changed on the door but never on Google.
• A phone number tied to a location you closed, still floating on an insurance site.
• An insurance list from three plans ago on a page nobody updated.
• A former provider still listed as if they work for you.

When the sources disagree, the model has no way to know which version is current. So it picks one, or it blends them, and it never tells the patient "I am not sure." This is the same reason your details can be wrong in the AI even when they are right on your own website: the AI weighed a louder, older source more heavily. The messier and more contradictory your footprint, the more room the machine has to invent. Consistency is not a nice to have here. It is the whole game. We dug into why your practice info has to match everywhere online, and it matters more now than when we wrote it.

There is a related trust shift worth knowing about. In the same research, patients said they now trust the AI Overview at the top of Google (37 percent) more than the traditional organic links below it (20 percent). The reason many practices got away with a messy online footprint for years was that patients would click around and piece together the truth. That safety net is going away. People read the summary and stop.

The hidden cost: patients you never see leave

What makes this so dangerous is that it is invisible. A slow website at least shows up in your analytics. A bad review you can read and respond to. But a patient who got wrong hours from an AI and decided you were closed never becomes a data point anywhere in your practice. There is no missed call, because they never dialed a working number. There is no cancelled appointment, because they never booked. They simply chose someone else based on a fact about you that was not even true.

This connects to something we see constantly: practices spending real money on ads and reviews to get found, while the plumbing that turns "found" into "booked" leaks the whole way through. If you want the fuller picture of how patients rule you out before they ever contact you, we covered it in why patients rule you out before they call. Wrong AI information is now one of the biggest leaks on that list, and almost nobody is checking it.

A quick story from the trenches

A family medicine group called us this spring frustrated that new patient numbers had softened even though their ads were running fine and their reviews were strong. We did the simple thing first: we asked three different AI tools basic questions about their practice, the way a patient would. Two of them listed a phone number from a satellite office the group had closed eighteen months earlier. One confidently said they were "not currently accepting new patients," which was the opposite of the truth, apparently pulled from an old note on a directory page. And every one of them listed Friday hours that were two years out of date. The practice had no idea. They had been turning away patients who never called, because the AI had already told those patients not to bother. We did not run a single new ad to fix it. We cleaned up the sources, and the phone started ringing with the people the AI had been misdirecting.

How to make sure AI has your practice right

You cannot log into ChatGPT and edit what it says about you. But you can fix what it reads, and that is most of the battle. These tools are only as good as their sources, so control the sources.

1. Lock down your Google Business Profile. This is the single most influential source for local practice facts, for both regular search and AI answers. Your name, address, phone, hours, and services need to be current and complete, and your holiday hours actually set. If you do nothing else, do this. We walk through it in our guide to the Google Business Profile for a medical practice, and it is the foundation of the healthcare SEO work we do.

2. Make every listing match, to the character. Same name, same address format, same phone, everywhere. Old directories, insurance portals, your Facebook page, your website footer. When the sources all agree, the AI has nothing wrong to repeat. When they fight, it guesses. This is tedious, unglamorous work, and it is exactly the work that moves the needle. More on why in whether your business listings need to match.

3. Put the facts in plain text on your website, not in an image. If your hours, insurance list, locations, and services live inside a designed graphic or a PDF, an AI cannot read a word of them. Type them as real text on the page. Your website is one of the main sources these tools use to describe you, so it needs to feed them clean, current, readable facts. A fast, clear site does double duty now: it teaches the AI the truth, and it catches the patient the AI sends.

4. Answer real patient questions in plain language. The practices that AI quotes correctly tend to be the ones that clearly state, in words, what they do, who they serve, what insurance they take, and where they are. This is the heart of optimizing for AI search, and it overlaps heavily with good old fashioned local SEO. We laid out the approach in our guide to Google AI search for medical practices and looked at the deeper question of why ChatGPT recommends hospitals and not your practice.

5. Check it yourself, on a schedule. Once a month, open ChatGPT, Google's AI results, and one more tool, and ask the questions a patient would: are you accepting new patients, what are your hours, what is your number, do you take a common local plan. Whatever comes back wrong is a leak with your name on it. This is the cheapest audit you can run, and almost no practice runs it.

Do not panic, but do not ignore it either

The point here is not that AI is out to get your practice. It is that a new, fast growing front door has opened, most patients now trust what it tells them, and it is pulling from a version of your practice that you have never audited. The good news is that the fix is not exotic. It is the same discipline that has always won local healthcare marketing: be accurate, be consistent, be everywhere the same, and make your real information easy to find. Do that, and the AI becomes a channel that sends you patients instead of turning them away.

The practices that get ahead of this now will look, in a year, like they had a head start. The ones that ignore it will keep wondering why the schedule feels light while the ads look fine, never realizing the answer is a wrong phone number in a paragraph they will never read.

How EtherealMinds handles this

We work only with healthcare practices in the United States, and this is squarely the kind of invisible leak we hunt for. When we build a practice's patient acquisition system, we start by auditing what the internet, and now the AI, actually says about you: your Google profile, your listings, your website, the whole scattered footprint. We make it match, we make it current, and we put your real facts in plain, readable text so both search engines and AI tools describe you correctly. Then the website we build catches the patient the AI sends, and an AI receptionist makes sure that when they do call the right number, someone answers on the first ring, day or night. Being described correctly gets them to reach out. Answering fast turns them into a booking.

You can start free today. Ask an AI the basic questions a patient would ask about your practice, and see what it says. If any of it makes you wince, that is the version thousands of people may already believe.