More than 40 million people ask ChatGPT a health question every single day. That is roughly the population of California, every day, asking an AI about their symptoms, their diagnosis, and what to do next.
Now OpenAI has announced an integration with Epic, the EHR system behind the records of hundreds of millions of patients. Most of the conversation so far has been about privacy, compliance, and how the technology fits into clinical care. Those are real questions. But there is another piece that specialty practices need to pay attention to, because this changes more than how patients understand their care. It changes how they find you.
If a patient asks ChatGPT who to see for the exact problem you treat, will your practice show up? If not, why not, and what can you do about it?
I walked through this in the video below, and the rest of this post covers the same ground in writing, including the checklist I use to audit a practice's online presence.
What did OpenAI and Epic actually announce?
There are two related developments, and they matter for different reasons.
The first is for providers. The Epic integration lets healthcare providers bring information from a patient's Epic chart into ChatGPT for healthcare. A provider can use it to organize a patient's last five years of lab history, or to help a patient prepare the questions they should ask at their appointment.
The second is for patients, and it is the one that changes your marketing. Patients can connect supported medical records directly to their own ChatGPT and use their own health data as the context for their questions.
People have been able to ask ChatGPT "what are uterine fibroids?" for years. That is a general question with a general answer. When a patient's actual health information is part of the conversation, the question becomes personal. She is no longer asking what fibroids are. She is asking what a finding in her own record means, whether it could explain her symptoms, and what her options are next.
"AI is moving patients from general information toward decisions that sit much closer to actual care."
How is the patient journey changing?
For years, the patient journey was predictable. Something happens. The patient opens Google, researches the problem, reads a dozen different answers, learns about treatment options, and eventually searches for "a doctor near me."
For a growing number of patients, that journey now starts somewhere else entirely, and it happens in one place. Here is what it looks like:
- A patient asks ChatGPT to help her understand a diagnosis of fibroids.
- She asks whether surgery is her only option.
- She learns about uterine fibroid embolization for the first time.
- At that moment she stops researching fibroids and starts looking for someone who performs UFE.
Diagnosis, treatment discovery, and specialist search used to be three separate trips through a search engine. Now they can happen, and already do happen, inside a single conversation with an AI. AI is becoming part of how patients discover care, not just how they learn about a condition. Practices can no longer assume every patient starts with Google.
How can you check whether ChatGPT recommends your practice?
You can run this test right now, and it takes about five minutes.
Pick the one thing you want to be known for
Not your full service list. The condition or procedure that matters most to your practice: peripheral artery disease, UFE, knee embolization, whatever it is for you.
Search the way a patient would
Open ChatGPT and ask who in your area treats that condition or performs that procedure. Use plain language, the way a patient talks, not the way your service page is titled.
Look at what comes back
Do you appear? If you do, note how you are described and whether it matches what you actually do. If you do not, note who does appear, and then look hard at what exists online about you.
Why doesn't my practice show up?
If you did not appear, the answer is almost always in what the internet can see about you. Ask yourself honestly:
- When was your blog last updated?
- Does your website clearly explain what you specialize in?
- Is your physician's LinkedIn photo from twenty years ago?
- Has your Facebook page been silent since January of last year?
- Do you have one basic service page for a procedure your physician has performed hundreds of times?
That is the version of your practice the internet can see. Your real-world expertise may be excellent, but if the online version of your practice is thin, outdated, or inconsistent, that expertise becomes much harder to discover. I wrote more about the most common causes in why most medical practices are invisible online.
I see this all the time in specialty healthcare. A physician can have years of experience treating a condition and still have almost nothing online that conveys that depth. Meanwhile, another practice publishes current, useful information explaining the condition, the treatment options, who is a candidate, what recovery looks like, and the questions patients ask most.
From the patient's perspective, one of those practices is much easier to understand than the other. The difference is not the quality of the care. The difference is that one practice has made its expertise visible and the other has not. As AI becomes a primary layer of how patients research, that gap gets wider.
If you want to be known for UFE, your digital presence has to show more than a single line saying "we offer uterine fibroid embolization." There should be enough substance online that a patient, or an AI reading on her behalf, can tell this is an area your practice genuinely knows.
Where should a practice start?
This is where we start with our clients, and it is not where most people expect. We do not begin by deciding whether a practice needs more blog posts or a different social platform. We start with what patients are actually trying to understand:
- What questions come up repeatedly during consultations?
- Where do patients get confused?
- What holds them back from booking?
- What do they need to know before they feel comfortable scheduling a consultation?
Then we compare those questions against what the practice already has online. The gaps usually jump out. Maybe you have a UFE page but nothing about candidacy or recovery. Maybe you explain the procedure well but never compare it with the other options. Maybe the answer your patients care about most only exists in conversations that happen inside the clinic.
Those gaps are the work. A patient's research moves from understanding the problem, to comparing options, to deciding who to trust. Your content has to support all three stages, rather than answering one question and sending the patient somewhere else to find the rest.
Do you need to create a lot of new content?
Here is the good news: most specialty practices are not starting from zero. The valuable information already exists. It comes out every time a physician explains a procedure, corrects a misconception, or answers a question they have heard a hundred times before.
The challenge is getting that expertise out of the exam room and into places patients can find it. One recorded conversation with a physician can become a long-form video, an article, several short clips, and an FAQ section. AI can make that process much faster, but it does not replace the source.
"AI is a photocopier. It can reproduce something quickly, but you still need something valuable to put on the glass."
For a specialty practice, the thing on the glass is the physician's real experience. That is also why AI-generated filler rarely helps: a copy of nothing is still nothing.
What should you actually do next?
The goal is not to write content for an algorithm. It is to make your practice genuinely useful and easy to understand online. In practice, that comes down to five things:
- Answer the questions patients actually ask, in their words, not your service-page headings. If you want the mechanics of structuring those answers so AI can cite them, the SOAP framework for AI search walks through it.
- Keep your information current. Stale pages and silent social profiles tell patients and AI tools the same thing.
- Make your specialty obvious. A visitor should know within seconds what you are known for.
- Show the people behind the expertise. Names, credentials, faces, and experience.
- Build enough depth that someone can keep learning on your site without hitting a dead end.
Google still matters. ChatGPT matters. Whatever comes next will matter too. The platforms will keep changing, but patients will always need to understand what you do and why your practice is the right fit for the problem they are trying to solve.
Frequently asked questions
What does the OpenAI and Epic integration do?
It works in two ways. Healthcare providers can bring information from a patient's Epic chart into ChatGPT for healthcare, for tasks like organizing lab history or preparing for an appointment. Separately, patients can connect supported medical records to their own ChatGPT and use their own health data as context when they ask questions.
Why does the Epic integration matter for healthcare marketing?
When patients ask AI about their own records, their questions become personal and move closer to a care decision. The path from diagnosis to treatment discovery to finding a specialist can now happen inside one AI conversation, so a practice that AI tools cannot find or understand can be left out of that decision entirely.
How do I find out if ChatGPT recommends my practice?
Open ChatGPT and ask, the way a patient would, who in your area treats the condition or performs the procedure you most want to be known for. Then look at what comes back. If your practice does not appear, review what currently exists online about you: your website, your blog, and your social profiles.
Does this mean Google no longer matters for my practice?
No. Google still matters, and so do AI tools like ChatGPT. What is changing is that Google is no longer the default starting point for every patient. The work that helps you in both places is the same: clear, current, in-depth information about what your practice does.
Can AI write my practice's content for me?
AI can make content production much faster, but it cannot supply the expertise. The most valuable material comes from your physicians' real experience, the explanations and answers they give patients every day. AI works best repurposing that source into videos, articles, clips, and FAQs, not inventing it.