Dr. Akrum Al-Zubaidi ("Dr. Aki”) didn't set out to become a tech founder. He was an interventional pulmonologist at a top institution, watching patients with early signs of cancer fall through the cracks. (Not because anyone was negligent, but because the system had no way to catch what it was dropping). He got pissed off. He dumped his savings in, took an at-risk loan, emailed 40 billionaires every Saturday morning until Mark Cuban replied, and built a prototype in his own clinic.
What a story.
That was ten years ago. Today, Eon Health has processed over 500 million patient records and is helping two million patients a year get the follow-up care they were never going to get otherwise. One Florida hospital caught 75% of their lung cancer cases at early stages using Eon, nearly three times the national average.
Here are three things I took away from our conversation.
The System = The Problem
Seventy percent of documented abnormalities never get the appropriate follow-up, and it's because the system buries patients.
Dr. Aki's example is one most people can relate to: You go to the ER after a car accident, they scan you for broken ribs, and in the process they catch a shadow on your lung or an enlarged vessel. You leave. You don't have a primary care doctor. The ER isn't thinking about a six-month follow-up. Your doctor is drowning in their own documentation. A year passes. Then three. Then you show back up with late-stage cancer and no curative options.
It happened to my own father — something found on a scan that nobody connected the dots on until it became a crisis. Dr. Aki's whole company exists because that story, unfortunately, is not rare. And the fix isn't asking doctors to work harder, but instead building a system smart enough to catch what humans inevitably miss at scale.
Not All AI Is the Same. In Healthcare, That Distinction Can Cost Lives
When most people talk about AI right now, they mean large language models, probabilistic engines that guess the most likely next output. They're powerful but variable. Run the same record through an LLM a thousand times and you might get a thousand slightly different answers.
In a high-stakes clinical workflow, that's dangerous.
What Eon built over ten years is a combination: a deterministic rules-based engine (same input, always the same output) layered with a probabilistic model, cross-checked against a massive human-validated dataset. When both models agree, and their agreement clears a confidence threshold that's twice as accurate as a human validator, the workflow executes automatically. This means no bottleneck and no validation burden on already-burned-out staff.
Dr. Aki's broader critique of the AI-in-healthcare hype cycle is worth sitting with: A lot of companies are building co-pilots and agents that flag things for humans to review, which just shifts the bottleneck instead of solving it. The right question isn't how powerful the algorithm is, but how you're handling its errors.
The "Wow" Moment in Healthcare AI Is Still Five Years Out
Dr. Aki's take on where AI in healthcare is headed is honest: Healthcare is in a confused pause, half-convinced that frontier models will do everything and hesitant to invest in narrower, proven tools in the meantime. That cycle will last 12 to 24 months. Some health systems will pick the wrong vendor, fail, and learn the hard way.
After that? Real momentum. But the thing that genuinely excites him isn't automation. Getting humans out of the loop, he says, is table stakes. The actual wow — the thing that keeps him going — is what happens when you aggregate all that data, organize it, put AI on top of it, and let it find signals that even the best human clinicians can't see. Patterns across millions of records that no individual doctor could ever hold in their head. A patient who just got a scan automatically matched to a clinical trial their doctor never would have known existed.
That seems like the version of healthcare AI worth building toward. Based on what Eon has already done in ten years, Dr. Aki seems like exactly the kind of person who will be there when it arrives.
Listen to the full episode of Actually Intelligent to hear more from Dr. Aki Al-Zubaidi on the Armenian guy at the club who inspired his "what's the maximum?" approach to life, why he thinks LLMs are a net positive for patients despite the real risks, and his simple sign-off that pretty much says everything: Do the right thing.
LEAVE A COMMENT