Originally published by Dave Chapman on LinkedIn.
A few weeks ago, out of frustration, I built an AI agent for my own spine issue.
I called it Spine Harmony. It is not a doctor. It is not a diagnostic product. But it changed how I think about medicine - and what patients can bring into a consultation room.
Because the immediate revolution in healthcare is not what most people assume. It is not AI replacing doctors. It is AI upgrading patients.
For most of modern medicine, there has been a huge information gap between the patient and the system. Doctors have training, experience, diagnostic context, clinical judgment, and access to medical language most patients do not understand. Patients have memories, symptoms, PDFs, portal logins, half-remembered conversations, scan reports, WhatsApp messages, family history, and a vague sense that something changed around March but they cannot quite remember what.
That gap is not the doctor’s fault. It is structural. A specialist may have twenty minutes with you. They cannot reasonably reconstruct ten years of symptoms, treatments, imaging, medication changes, injuries, flare-ups, failed experiments, and previous opinions from a scattered pile of records. Even if they are excellent, the system is not designed for deep continuity.
AI changes that.
I have had a long-running L5/S1 problem: a couple of years of history, imaging, consultations, treatment options, injections, questions about disc replacement, and all the usual uncertainty that comes with chronic pain. Spine Harmony did not replace my doctors, and I would never use it as a standalone medical authority. But it did help me make sense of the diagnostic picture. It pulled together years of history, imaging, symptoms, and previous medical opinions, and helped confirm that the L5/S1 disc was indeed the pain generator. It also helped me reason through treatment options: because I had historically responded well to anti-inflammatories, an epidural steroid injection looked like a logical next step to discuss with my doctors.
But the real value was not one clever recommendation. It was continuity.
It organised the history. It tracked symptoms. It helped me prepare for consultations. It kept a pain diary after I went ahead with the epidural steroid injection. It remembered what had been said by which doctor and when. It helped turn a messy personal medical story into something structured enough to discuss properly with specialists.
In other words, it made me a better patient.
A patient with an AI agent can bring a doctor a clean timeline instead of a rambling memory. They can bring a structured symptom diary instead of “it hurts sometimes.” They can bring better questions. They can understand the words in their report before the appointment. They can compare what changed after a medication, an injection, an operation, a supplement, a new exercise routine, or a bad night of sleep.
It also changes the balance of the conversation. A better-prepared patient can ask why one option is preferred over another. They can notice when something in their history has been missed. They can follow up on inconsistencies. They can understand the vocabulary well enough to participate. The best doctors will welcome this; it makes the relationship more honest and the diagnosis more accurate. Doctors who relied on the information asymmetry to avoid hard questions will not. That is fine.
A doctor recently asked me how much I would sell Spine Harmony for. It was a flattering question, because it suggested the thing was useful. But I have no desire to commercialise it; I’d rather help others in similar situations.
So I have open-sourced the Spine Harmony agent spec under an MIT license: github.com/d2ma-tech/spine-harmony-agent. Here is an example of what a Spine Harmony report looks like.
It is not medical software. It is not a diagnostic product. It is not a replacement for clinical judgment. It is an agent specification and workspace pattern for managing a spine-health journey end to end: organising research and medical history, tracking pain and symptoms over time, preserving imaging and scan-report context, preparing for specialist consultations, comparing treatment options, and generating patient-facing reports.
Once I saw what Spine Harmony could do for one condition, it became hard to accept that the rest of my family’s health information should continue living in PDFs, lab reports, hospital discharge notes, scattered memories, portal accounts, and late-night messages.
So I started building the same pattern elsewhere.
Vitalis Health is where I am organising adult health, bloodwork, longevity markers, supplements, peptides, and general medical context. Little Wellness is for my kids’ health matters: post-operative notes, symptoms, medicines, sleep, questions for doctors, and the practical chaos of being a parent trying to remember exactly what happened at 2am. Perhaps I open-source those too, let me know if you’re interested.
None of this is glamorous. It is not a moonshot biotech company. It is mostly memory, organisation, summarisation, structured questions, and continuity. But that is precisely why it works.
So much of healthcare is not limited by the absence of data. It is limited by the inability to use the data we already have. We are surrounded by information now: blood tests, wearables, sleep trackers, glucose monitors, MRI scans, genetic data, food logs, medication histories, supplements, symptom trackers, hospital letters, and doctor notes.
The bottleneck is no longer collection - it is interpretation, continuity, and knowing what to ask next.
Good AI use is not about outsourcing judgment. It is about improving the quality of your own questions: what changed, what did not, what is supported by evidence, what is anecdotal, what should I stop pretending I understand?
One of the most striking recent stories was about Paul Conyngham, an Australian entrepreneur who used AI, scientists, and mRNA tooling to help create a personalized cancer vaccine for his dog, Rosie. A few years ago, that story sounded strange. Today, it reads like a preview of the next stage of medicine: patients arriving with research capability, memory, and coordination power that used to belong only to institutions.
The lesson is not that everyone should start designing medicine at home. The lesson is that motivated individuals are gaining access to research, reasoning, and coordination capabilities that used to sit far outside ordinary life.
Of course, there are real risks. Arguably, the most serious one is privacy. If you paste your MRI report, bloodwork, medication history, or child’s medical notes into a cloud AI system, you should assume you have disclosed that information to that provider. That may be acceptable to you, or it may not be. But people need to understand the trade-off.
And AI should not be used to decide whether to have surgery, stop medication, ignore symptoms, self-treat serious disease, or avoid qualified medical care.
The unlock is not “ask AI what to do instead of seeing a doctor.” The unlock is “use AI to become better prepared for the doctor.” Those are very different things.
We are moving into a world where a patient can read the latest research with help. Where a parent can maintain a structured medical timeline for a child. Where a person with a chronic condition can track symptoms over years with context. Where a doctor can receive a clearer history. Where a second opinion can start from better organized information. Where the gap between “I have a pile of records” and “I understand the shape of my own health story” becomes much smaller.
That is not anti-doctor. That is pro-medicine. The best doctors should want better-prepared patients. The best patients should want better conversations with doctors. AI can help both sides get there.
Medicine is becoming more personal, more data-rich, more continuous, and more participatory.
The doctor has not disappeared. But the patient has changed.
Medicine is about to get weird. And for the first time, that weirdness is on the patient’s side.
-
Dave Chapman is Chief Operating Officer of Quantum Leap Acquisition Corp, a NYSE-listed SPAC focused on AI, quantum computing, and blockchain technology. This article reflects his personal views and is not investment advice.