Physician Driven Innovations — clinical AI built for the physician, not the system

Built by physicians, for physicians
Constraint, not governance
Your records stay yours
Descriptive, not prescriptive

We’ve already built the first one. A tool that helps OB/GYNs organize and defend the cases they’re tested on for board certification, built to every principle on this page. The rest is why it has to work this way.

01 What we’re for

You’ve been there. In that room.

A loved one you came to support. A personal problem you face. This room is where the heart of medicine lives. And every external force is vying for the physician’s attention, away from you. Attention is the most costly thing you pay a physician for. Your pain. Your body. Your story. It’s also the easiest thing to quietly take away.

02 The incentives

Show me the incentive and I’ll show you the outcome. Charlie Munger

Most tools arriving in medicine weren’t built for the people in the room. They were built for the systems around it.

Hospitals, insurers, software vendors: each has its own reason to want what happens between a doctor and a patient measured, scored, and acted on. None of them are villains. They’re just not the patient. And when you want to know where a technology ends up, the honest question isn’t what it promises. It’s whose incentives it quietly serves.

The obvious objection

“So you’re building a scribe?”

The scribe is the cautionary tale, not the model. It solved a real pain: documentation. But it was built by companies whose incentive is data, metrics, and a seat in every conversation a patient ever has. So relief came bundled with surveillance and dependence, and physicians traded a piece of their autonomy for the time saved, willingly, because the pain was that real.

A problem solved by creating the next one. It runs on the wrong incentives, with no limit built into the tool. So by the time rules get written, the harm is already done. We won’t solve one problem by creating the next.

Incentives tell you the outcome. So we set ours on day one. “Built by physicians, for physicians” isn’t a slogan. It’s the most aligned a tool can be: made by the one with the most skin in the game, and the only perspective that matters.

Why physicians

Society’s gold standard for a life held in another’s hands has always been the physician.

They train the longest and carry the final responsibility: the one a person is entrusted to when it matters most, and the name on the decision when it’s later questioned. Medicine is practiced as a team, but the moment it turns to litigation it becomes a solo endeavor — the physician, alone. It has always been that way. It doesn’t have to be. Yet nearly all of medicine’s weight still rests there, and when bad outcomes occur, the physician is the first to answer for them. That weight never moves.

The ones who carry the most should have the best tools.

PDI is built by a physician, on a simple conviction: preserve medicine from the top of that responsibility down. The people who have given the most to it deserve tools that help them reason more clearly, think further, and see how medicine is actually practiced. Not to replace their judgment. To strengthen it. Physicians deserve that, and so do the patients who depend on them.

03 How the room gets lost

Here’s how the room gets lost: not all at once, but a little at a time.

Picture the textbook running out, and a real person, you, who doesn’t quite fit it. What your doctor does next, the judgment made for you alone, is exactly what the pressure around them wears down.

That pressure has never existed at this scale before. AI makes it possible, and left alone, more powerful every year. No one set out to do harm. It’s just where the incentives point, unless something is built to point the other way.

Until now, this has been about the room, and what it stands to lose. From here, I’m speaking to the physician inside it.

04 How it’s built, not how it’s used

Everyone says AI in medicine needs governance, rules for how a tool is allowed to be used. We care about something earlier: how it’s built in the first place.

A rule can be softened, worked around, or quietly dropped when there’s enough money on the other side. Build the limit into the tool itself and it doesn’t depend on anyone’s good behavior.

“We promise not to look at your records” is a rule. “We built it so we can’t reach your records” is a fact.

The identifying details are stripped out inside your own environment, before anything leaves it. Where information does travel beyond it, it’s stripped of identity first, and what travels is covered by a HIPAA agreement as a second layer, so even a mistake is contained, not exposed. One approach asks you to trust us. The other is built so you don’t have to, on the assumption that anything that can be misused eventually will be.

We call it the gray zone: the moment the guidelines run out and it comes down to judgment and a real person.

Here’s what wears it down. When a tool measures every call against the average, then every time you tailor a decision you have to be ready to defend the deviation. So the safe move becomes the average one, every time. Practice at the edge starts to look like risk, and one by one, physicians retreat to the middle, until the standard itself narrows to whatever’s safest to write down.

Sometimes the guideline is exactly right, and a good doctor takes it. But when real-world care diverges from the page, that’s not always a mistake. Sometimes it’s medicine learning something new. The tools we build protect that. They don’t punish the doctor who practices at the edge.

Judgment that compounds

When you see it, you can’t unsee it.

Right now, every hard call in the gray zone is made alone. You reason through it with what you know, what you’ve seen, what you can find. And then you carry it by yourself. That isolation isn’t a feature of medicine. It’s a failure of its tools. Judgment has never been allowed to accumulate the way guidelines do. Every physician who ever faced your exact situation reasoned through it in their own silo, and all of that hard-won thinking evaporated instead of compounding.

It doesn’t have to work that way. Imagine the same gray-zone decision, made with the quiet backing of the physicians who stood where you’re standing: how they reasoned, what they weighed, what they chose when the guideline ran out. Not a rule telling you what to do. The collective judgment of your peers, surfaced when you need it, so the hardest calls are informed by more than one person’s memory.

That’s the Gray Zone Intelligence Network. And here’s why it’s never existed: the incentive was never there, and the technology wasn’t ready. Every system capable of aggregating how medicine is actually practiced was built by someone whose interest was in watching physicians, not backing them, so physicians were never going to feed it honestly, and they were right not to. The thing that’s missing isn’t the data. It’s an aggregation physicians can trust because physicians govern it.

I’m not going to tell you this network exists in full today. It doesn’t. What exists is the first product: the proof that clinical judgment can be collected, protected, and returned to the physician without a single one of the usual compromises. No payer can reach it. No employer can score you with it. No company with deep enough pockets can buy its way to your record, because the constraints aren’t policies that can be sold off. They’re built into how the thing works. The proof of concept is the groundwork. The network is what it’s the foundation for.

Others will try to emulate this. Some already are. But the one thing they can’t replicate is the reason it works: it belongs to the physicians who build it, and sovereignty isn’t something you can bolt on after you’ve already built your business on extraction. You either constrain the architecture from the first line of code, or you don’t. We did.

This is what it means to not be alone in the gray zone anymore.

Standing behind the call

Every physician learns to defend a decision after it’s made.

To the oral boards. To a lawyer. To a patient asking why. You explain what you did, on what evidence, on what grounds. Even when the grounds were judgment the guideline never covered, and the evidence was hard to put into words.

Medicine may be the only place a good-faith decision can be made to look indefensible.

The tools we build move that defense to the front of the decision, not the end of it. You can see how physicians who faced the same situation reasoned through it, and build your own rationale while the patient is still in front of you, not years later under oath. A documented reason, every time the guideline doesn’t fit. Not to cover yourself. To practice with the confidence that you can stand behind what you chose. That confidence is the constraint PDI was founded on.

Guidelines will become expansive, while clinical judgment will become scarce.

06 How a good decision gets made

A good decision for one person draws on more than a guideline.

It draws on four things at once: what the research says, what happens across whole populations, what experienced colleagues actually do, and what you’ve learned from your own patients over the years. Your judgment is what weighs all four. A tool’s job is to put them in front of you, clearly and honestly, then get out of the way. It can show you the guidelines, show you the choices other physicians made, and then the decision rests with you. Nonpunitive. Just presence.

07 What we promise

Four promises about whose side the tool is on.

It stays yours.

Your records and reasoning stay in your hands. No one else holds the key.

It never decides for you.

It informs your judgment; you make the call.

It never watches you.

Nothing you do becomes a score that grades you from the outside.

There’s nothing to turn against you.

The record you keep lives where only you control it, so there’s no outside copy waiting to be used as a weapon against the judgment that filled it.

And the simplest promise of all: if a tool can’t be built to keep these, we don’t build it.

08 How it works, plainly

It stays with the doctor. Only the lessons travel. And what’s learned comes back to them first.

It stays with the doctor

A doctor’s records live in their own locked vault. They hold the key. Nothing leaves without them.

Only the lessons travel

When something is worth sharing, the identifying details are stripped away inside your environment first. Only the pattern leaves, not the patient. And where that pattern is processed, it travels under a HIPAA agreement, so the protection holds even if something is missed.

What’s learned comes back first

Whatever the wider network learns, the doctor who contributed gets it back before anyone else. That’s the whole test: when what’s learned returns to the doctor, it’s serving them. When value gets pulled away instead, that’s the thing we’re built against.

It only works if it reflects how medicine is really practiced: every kind of case, not a hand-picked few. Otherwise the lessons would be as skewed as the evidence we already have.

How the lessons travel without the records ever leaving: the full architecture, on the PDI Med Substack

09 The first thing we built

We started with one real problem, built the right way.

To sit for the OB/GYN oral boards, you have to collect and defend a list of the real cases you’ve managed. It’s tedious, high-stakes, and exactly the kind of place where judgment matters and the tools have always been bad. So that’s where I started: a tool that helps a doctor organize and stand behind their own clinical judgment, built to every promise above. Proof it can be done this way.

I started with OB/GYN because it’s mine: the pain I can solve right now. But it’s where we start, not where we end. The architecture doesn’t care what specialty you practice, and the principles behind it reach far past any one of them.

PDIMed · Case Logs · Oral Board Prep

Your cases. Your vault. Built to stay yours.

Works in any browser · Independent of any hospital system · Yours to keep

pdi-med.com

10 The invitation

The best tools come from real frustration.

If you’ve stood in the gray zone — made the call the guideline couldn’t make, felt the pull to match the average just to stay defensible, watched good judgment get treated as a liability — then you already know this is real. You don’t need me to convince you. You live it.

This is being built for you. The physician who carries the decision, and answers for it. Not the hospital, not the payer, not the layer that grades the call from the outside without ever standing behind it.

Latest from the writing · June 25, 2026 A Hemoglobin of 10.5 Read it

The whole argument, and where it goes next, is on the PDI Med Substack.

Start there

The space between a physician and a patient is sacred. Almost every tool built for medicine takes from it. We build the ones that give back.