Spend five minutes reading about healthcare AI and you’ll come away convinced medicine is about to be automated out of existence. AI will diagnose. AI will answer patients’ questions between visits. AI will draft the note, monitor the glucose sensor, and greet the patient at check-in. Soon, we’re told, AI will simply become the interface between patients and the healthcare system.
Much of that might be true. But it’s also missing the more interesting story.
Every significant trend creates a countertrend. When the herd stampedes in one direction, the money, and the differentiation, often sits in the other. Ask anyone who still sells vinyl records. E-books were supposed to kill the printed book. Instead, people rediscovered that there’s something to holding an actual book, turning an actual page, and not staring at a screen for once. The technology didn’t erase the older experience. It reminded people why they liked it.
Healthcare is walking the same path, and most physician leaders haven’t noticed yet.
What I’m Not Talking About
Let’s be precise, because this gets muddled fast. I’m not talking about AI as a diagnostic tool, or as decision support sitting quietly behind a physician’s shoulder. I’m not talking about AI running your scheduling, your registration, or your check-in kiosk. Those are worth their own conversation, and I’ve had it elsewhere.
I’m talking about AI stepping into the clinical relationship itself: taking the history, educating the patient, fielding routine questions, explaining the treatment plan, following up after the visit, potentially becoming, for a growing number of patients, the primary voice of the practice.
For plenty of organizations, that’s the whole point. Lower cost, higher throughput, more patients per FTE. Those are their business goals; I’m not here to pretend otherwise. But they raise a question worth sitting with: if the relationship gets mediated through a chatbot, what exactly are you still selling?
Two Models, Not One
Some organizations already know their answer. They’ll keep marching toward scale, standardization, and technology-mediated everything. AI becomes one more tool in the industrialization of medicine that’s been underway for two decades. Whether you think that’s progress or a slow-motion mistake is beside the point; it’s happening, and arguing about it won’t stop it.
The more useful question is what happens on the other side of that trade.
My prediction: AI doesn’t produce one model of medical practice. It produces two. The first keeps industrializing. The second, direct primary care, concierge medicine, membership models, and relationship-driven specialty practices, for example, moves in the opposite direction, on purpose.
At first glance, that second group looks like it’s fighting AI. I think that’s backwards.
The Force Multiplier, Not the Replacement
The physicians who benefit most from AI may turn out to be the ones most committed to the old-fashioned business of knowing their patients.
That only sounds counterintuitive if you assume AI’s job is to replace the physician. Assume instead that its job is to amplify one. Let AI handle the documentation, chase the lab values, draft the after-visit summary, and manage the routine messages that currently eat a physician’s day one inbox item at a time. Every hour it saves is an hour that can go back into the conversation, the exam-room judgment call, and the reassurance patients actually remember, which is the part no language model, however fluent, can manufacture.
Used that way, AI isn’t shrinking the physician’s role in the relationship. It’s clearing the floor for it.
That distinction matters because patients were never just shopping for information. They’re shopping for judgment, for confidence, and, far more than either, for trust. AI can support all three. It cannot substitute for any of them, no matter how good the next model release is.
The same logic runs through specialty practices. As technical information gets commoditized by AI, and it will be, one of the few things left to differentiate a specialist’s practice is availability, continuity, and a relationship that survives past the procedure. That stops being a nice-to-have. It becomes the product.
Where AI Belongs
Most of the current debate asks whether physicians should embrace AI or resist it, as if those were the only two options. They aren’t. The real question is where AI belongs in your practice.
Deploy it to replace the relationship, and you’ve bought yourself a faster version of the same industrialized medicine everyone already complains about. Deploy it to protect your time for the relationship, and you’ve built something closer to the opposite: a practice where the technology is the reason you can still afford to know your patients’ names.
The irony, if there is one, is that the tool most people assume will finish off the physician-patient relationship may be exactly what saves the practices built around it, and hands them a competitive advantage the industrialized side of medicine can’t buy back.
If you’re thinking about where AI fits into your group’s strategy, as a threat to manage or an advantage to build, I’m glad to talk it through. You can reach me at markweiss@weisspc.com.
