A patient sits across from you, nods along, thanks you, and leaves. Two hours later they are at their kitchen table typing your diagnosis into ChatGPT, along with their lab numbers and the plan you gave them, asking a chatbot whether you got it right. This is not a rare thing anymore. It is fast becoming the normal last step of a doctor visit, and most practices have no idea it is happening.
The scale is the part that surprises people. In Rock Health's 2025 Consumer Adoption of Digital Health Survey, which polled 8,000 U.S. adults, 32 percent said they had used an AI chatbot for health information. A year earlier that number was 16 percent. It doubled in twelve months. And these are not one time dabblers: 64 percent of those users go back weekly or more. When people find something that answers fast and never judges them, they keep using it.
The exam room is no longer the last word
For a long time the visit was the end of the conversation. Whatever the doctor said was the answer, because there was nowhere else to check without another appointment. That is over. A Baltimore Sun commentary this year called ChatGPT the new second opinion, and that is exactly what it has become: a free, instant, always available second doctor that every patient now carries in their pocket.
It shows up on both ends of the visit. Before they book, patients research their symptoms with AI, which we wrote about in how patients self diagnose with AI before the visit. After the visit, they run your plan back through it to make sure you did not miss anything. Rock Health found that 81 percent of AI health users took some action after their chat, and the most common ones were searching for more information and then talking to a provider. So AI is not always replacing you. Often it is grading you, and then deciding whether to trust what you said.
Why this should get your attention, not your eye roll
It is easy to wave this off as patients being difficult. Do not. There are two real reasons it matters for your practice, one risky and one full of opportunity.
The risk: AI sounds confident even when it is wrong, and patients act on that confidence. In one widely reported survey, nearly one in three Americans said they would skip or delay seeing a doctor if an AI tool told them their symptoms were low risk. Physicians have warned this can lead to missed diagnoses and delayed care. When a chatbot casually reassures someone who actually needed to come in, that is a patient you never see and a problem that grows.
The opportunity: a patient who checks their care with AI is an engaged patient. They want to understand. They are not lazy, they are anxious, and anxiety is where trust is either won or lost. If your explanation was clear and your practice looks credible everywhere the patient looks, the AI check confirms you and their trust deepens. If your explanation was rushed and you are nearly invisible online, a generic chatbot fills the gap and becomes the voice they believe. Same patient behavior, two very different outcomes, and you get to influence which one you get.
What you actually do about it
You are not going to talk patients out of using AI, and telling them to stop just makes you sound like the doctor who says do not Google it. The winning move is to make your practice the source that AI and the patient both keep landing on. Three things do that.
1. Explain so clearly there is nothing left to panic search
Most after visit AI checks are driven by a simple gap: the patient did not fully understand what you said, so they go find someone who will explain it slower. Close that gap in the room. Say what the diagnosis means in plain words, say what the plan is and why, and ask them to repeat it back. A patient who leaves genuinely understanding their care has far less reason to seek a second opinion from a machine at 11pm. Clarity is the cheapest defense you have, and it costs only a little more attention.
2. Publish the answers yourself, so you are the source AI cites
Here is the leverage point most practices miss. When a patient asks AI about a condition you treat, the tool answers using content it found online. If the clearest, most trustworthy content about that condition in your area is on your website, you become part of the answer. If your site is thin and jargon filled, some other source speaks for you.
This is why real, plain language content on your own site matters more now than it ever did, not less. It is the same reason your practice needs to be visible in AI answers at all, which we broke down in how to get your practice recommended by ChatGPT and in our look at why ChatGPT recommends hospitals and not your practice. Good healthcare SEO is no longer only about ranking on Google. It is about being the credible source that the AI second opinion pulls from, so the machine ends up agreeing with you.
3. Be reachable the second the question comes up
Patients turn to a chatbot partly because it answers instantly and your office does not. It is 8pm, they have a worry, and calling means voicemail until morning, so they ask AI instead. Every one of those moments is a small handoff of trust away from you. Shrink that gap. Answer questions fast, capture the after hours ones, and give patients a real way to reach a human sooner. This is a big part of why we built our AI receptionist: not to replace the human touch, but to make sure a patient with a question gets your practice on the line instead of settling for a stranger's chatbot.
Trust is now built before and after the visit, not only during it
The old model was simple: build trust in the exam room, and that was that. The new model is wider. Trust starts before the patient books, when they are researching you and their symptoms online, and it continues after they leave, when they are checking your advice against AI and reading how you show up everywhere else. The practices that win the next few years are the ones that look and sound trustworthy across that whole arc, not just in the fifteen minutes face to face.
That is largely a marketing job, and it is a human one. It means a website that explains things like a caring person wrote it, a social presence where patients can see your real doctors and hear a real voice, which we make the case for in why doctors should show their faces on social media, and reviews that read like real people were treated well. A patient who has already seen you be clear, human, and credible online walks into the AI second opinion on your side.
Where EtherealMinds fits
We work only with healthcare practices in the United States, and this shift is one of the biggest we are helping clients get ahead of. The care you give is good. The problem is that a chatbot with no idea who you are is now weighing in on it, for free, in your patient's kitchen, and it will win by default if you are invisible online.
So we make sure you are heard. Our websites turn your expertise into clear, patient friendly content that both people and AI tools trust and cite. Our healthcare SEO works to make you the source that shows up when patients research, instead of a hospital three towns over. Our social media management puts your real people and real voice in front of patients so the trust is already there. And it all runs as one connected patient acquisition system, so the search, the site, the social, and the follow up all say the same thing. Patients are going to ask AI about their care no matter what. We make sure the answer sounds a lot like you.
Would an AI second opinion confirm your practice, or replace it?
Book a free strategy call. We will look at what patients and AI tools actually find when they research a condition you treat and search your practice name, and show you where you are getting spoken over. Clear, honest, no jargon.
Book a free strategy call →