That is the whole bet. I build the accountable layer of healthcare AI — the named clinician, the owned record, the signed outcome — across a portfolio of products that are already live.
Everyone is automating the work. We own the part that has to answer for it.
Every vendor — OpenAI, Anthropic, Google, xAI — now ships the model, the agent, and the cloud computer as one bundled commodity. In healthcare that means the draft is about to be free: the note, the prior-auth letter, the billing code, the care plan. The scarce thing is no longer the ability to produce them.
When the draft is free, the signature is the business.
We build on those. Only those.
MD, twenty years at the boundary between surgeons and technology. I built BrainLAB’s orthopedic navigation vertical from zero to $250M across two continents; five patents in image-guided navigation. Part of three healthcare acquisitions — InVivoLink (HCA Healthcare), PumpOne (Anytime Fitness), Disior (Paragon 28). I’m a translator by training — MD-trained, not a practicing surgeon — fluent with surgeons, health systems, and reimbursement.
The person who used to advise the founders now builds the things — the clinical judgment and the production software in the same week.
Different doors. One engine: draft with AI, sign with a human, own the record.
A model vendor cannot be the accountable clinician. A claims giant cannot sit with your mother. The defensible layer is the one no amount of capital or compute can buy: a named human’s accountability, the owned outcome ledger, and the network of both. Competitors ship the automation in a weekend now — and that is exactly why the automation was never the moat.
The medical record bills. Ours proves — and answers to someone.
SurgeonValue catches the codes an EHR template skips. ClinicalSwipe turns AI output into signed, billable physician attestations. co-op.care is running with its first families in Boulder — a $59/month membership and a $199 physician assessment that documents care already eligible under §213(d).
Honest state: early, and deliberately small before it is replicable. No exits yet on this stack — this is the build phase. The three exits above are prior chapters, not this one.
CMS FY2027 rules · KFF Health Information & Trust, 2026 · public filings.
This is the whole picture; each product has its own deeper read. If you back the accountable layer of healthcare AI — as an investor, an operator who wants to build it, or a partner with distribution — that is the conversation I’m looking for. Sharpest-question-first. No pitch, no agenda beyond finding out if we see it the same way.
Find a time to talk →Or reach me directly: Blaine Warkentine, MD · orthoblaino@gmail.com · Boulder, CO