This blog is being dictated into my phone from my truck. By the time you read it, AI will have transcribed my voice memo, organized my brain dump into something coherent, and handed me back a draft I edited in fifteen minutes instead of writing from scratch in two hours. That’s use case ten of ten in this post, and I’m leading with it because it’s the most honest entry point into how I actually think about this technology.
For independent practice owners, the constraint is time. Not capital, not ideas, not even staff in the long run. Time. AI doesn’t replace clinical judgment, it doesn’t replace the relationships we build with our patients, and it doesn’t run the practice for me. But used deliberately, it’s the difference between building in two years versus five. Here’s the working stack as of right now —what I’m using, what I’m building, and what I’m staying away from.
Operations and Compliance
The most boring use case is also the most valuable. AI is doing a significant portion of the heavy lifting on our SOPs. When we’re writing a medication-related protocol — say, an IV nutrition or ozone protocol — I have AI pull the relevant state regulations, surface what’s actually required versus what’s customary, and draft language we can verify against the source documents. We still confirm everything. But I’m not paying my Director of Operations to spend twenty hours a week reading administrative code when AI can do the first pass in twenty minutes.
In parallel, we’re building a JC 101 binder — the cohesive onboarding resource every new hire walks into on day one. AI is helping us pull and organize that material so it’s reproducible. New staff member starts in McKinney next quarter? The same binder gets generated in two clicks instead of being rebuilt from scratch every time someone joins the team.
The next thing on the operations roadmap is an internal AI search across our own database, so a newer team member with a question doesn’t have to interrupt a senior team member or guess. They search our protocols, our SOPs, our internal guidance, and get a synthesized answer pulled from our own documents. That’s in build right now.
Clinical Support
Our EMR has a built-in ambient scribe feature. We can turn it on during a patient visit, and at the end of the encounter it produces a draft note that the physician edits and finalizes. The clinician is still doing the medicine, the documentation, and the thinking. The AI is just removing the tax of typing.
The other clinical use case is pre-visit planning for complex patients. When I’m preparing for a complicated case, I can feed lab reports into AI and have it produce a cohesive synthesis — a working frame for the conversation, not a treatment plan. I want to be very clear about this one: we are not using this as a patient-facing tool, and we are not letting AI generate clinical recommendations that go anywhere near a patient. We’re testing it on ourselves and our team first. If it doesn’t earn trust over the next year, it doesn’t get deployed.
Education and Content
Patient education is one of the hardest things to do well in an integrative practice. Our patients want to understand what’s happening in their own bodies, and that takes coordinated material —not one-off handouts produced whenever someone has a free afternoon. AI is letting us build areal educational calendar with handouts, slide decks, and topics that actually flow together over the course of a year.
The visuals are a separate breakthrough. I no longer need to commission a medical illustrator or worry about licensing on stock imagery. ChatGPT generates the diagrams we need for our handouts and slides. They’re accurate enough to work, original enough that we own them, and produced in minutes instead of weeks.
After our patient webinars, we run the recordings through AI for editing and SEO optimization before they post. Cleaner cuts, better titles, better metadata. More of the right people find them.
Build and Marketing
When I want a new website page or a redesigned section, I no longer try to describe the vision to my web designer in words while she tries to read my mind. AI generates a visual mockup. I hand the mockup over. She has a starting point. The iteration cycle goes from weeks to days, and we both stop guessing.
And then there’s the dictation work. I dictated my entire book that way. I dictated job descriptions that way. I’m dictating this blog that way. On a drive from Ankeny to Chesterfield, on a flight, anywhere I have voice and signal, I can feed the brain dump in and let AI structure it while I keep moving. For someone with ADHD running two companies, this isn’t a productivity hack. It’s a structural advantage.
What I’m Not Using It For
I’m not using AI to replace clinical judgment. I’m not using it patient-facing for medical advice.I’m not using it to write content that pretends to be human-thought when it isn’t. Every piece ofwriting that goes out under my name passes through me. Every protocol gets verified. Everyclinical synthesis gets read by a clinician.
The point of AI in an independent practice isn’t to outsource thinking. It’s to outsourceeverything that isn’t thinking, so the doctors can do more of the work only doctors can do.
That’s the deal I made with it. It’s working.
