Every doctor’s consultation is a hidden adversarial game.

The doctor has already been paid. They want it over quickly. The patient has already paid (or implicitly agreed to). They want to extract maximum time. Both deploy strategies to get what they want - doctors rush and run behind schedule, patients show up with lists of secondary concerns to “get their money’s worth.” The fixed fee creates a dynamic where neither party’s interests around duration are aligned.

This is the default equilibrium. Nobody deviates because the entire infrastructure assumes it - premises, staff, Medicare item numbers, billing software, patient expectations. You can’t fix it by switching one clinic to per-minute billing. The switching cost is prohibitive when everything else stays the same. You have to build the entire system from scratch around the new billing model - no premises, no staff, no Medicare, video-only, AI everything.

The Intervention

Bill by the minute.

This sounds like a small operational tweak. It isn’t. It’s a single parameter change that cascades into a fundamentally different incentive structure.

The adversarial dynamic around duration dissolves. Not because both parties suddenly want the same thing, but because the patient gains a natural off-switch: they stay as long as the consultation is worth more to them than the per-minute cost, and they leave when it isn’t. The doctor has no reason to fight this - they’re compensated for however long it takes. There’s nothing to resent in either direction. A five-minute consult and a twenty-minute consult are both fine.

This is the important part. Everything else follows from it.

Second-Order Effects

Preparation becomes rational for the patient. A patient who shows up with “I have this thing, it’s been there a while, I tried some cream” burns 5 paid minutes transmitting information they could have organised beforehand. A patient with clear photos and a structured history might need 3 minutes instead of 12. The incentive to prepare well sits with the party whose time is cheaper. This isn’t a feature anyone needs to sell - it falls directly out of the pricing.

Patients genuinely want to use AI prep tools. If an LLM can turn disorganised medical history into a structured summary, that directly reduces the bill. The clinic can embed this as a prep step (open question on how far to take embedded medical AI - patients are already doing it independently with ChatGPT, so building it in at least gives quality control over the prompting).

The doctor becomes indifferent to how the patient uses time. If someone wants to spend 10 minutes having you read through disorganised biopsy results they didn’t summarise - fine. They’re paying. No resentment. The meter is running, everyone knows it, the patient made their choice. Under a fixed fee, this exact scenario breeds quiet frustration on the doctor’s side. Per-minute billing eliminates it.

An honest caveat: the doctor influences the pace of a consultation more than the patient does. A doctor could pad time with unnecessary follow-ups in ways the patient can’t easily detect. This is a real information asymmetry. The counterweight is reputational - a per-minute doctor who consistently runs long will lose patients to one who doesn’t. But it’s worth naming rather than pretending the incentive alignment is perfect.

Prefunded accounts reduce friction. $500 deposit on signup. Eliminates payment chasing and the awkward billing conversation. The consult ends cleanly.

No Medicare, no bulk billing, no rebates. Fully private. This removes an entire layer of administrative overhead and perverse incentives, at the cost of narrowing the patient pool to those willing to pay out-of-pocket. Patient pays doctor for time. That’s it.

Zero Staff

Once the billing model removes the need for complex scheduling and invoicing, most clinic infrastructure becomes unnecessary.

AI agents answer the phone. AI scribe generates clinical notes and referral letters from consult audio. Scheduling, billing, and patient comms are automated. No premises - video only. One specialist runs the entire operation.

The Platform Observation

None of this is specialty-specific. Scheduling, per-minute billing, AI intake, AI scribe, AI receptionist, Stripe integration, video with a visible timer - that’s a platform. Any specialist could run on it.

Three avenues:

The Clinic. Build and run one clinic. Prove the model, generate revenue. Solo project.

The Platform. Package it as SaaS. Turnkey clinic-in-a-box for other doctors. You become a software company.

The Chain. Don’t sell the platform. Use it as a moat, recruit specialists to run clinics on it. Capture consultation revenue, not subscription fees. Scale without selling the advantage.

Avenue 1 is a solo project. Avenues 2 and 3 are where a co-founder gets interesting.

Tech Constraints

  • Files not apps. Data portability over vendor lock-in. Every component hot-swappable.
  • Stripe usage-based billing. The billing model is the core differentiator - it needs to be solid.
  • Own the data. Encrypted backups with verified restores. AU data residency for patient records. No trusting a SaaS provider with data portability.
  • Self-host where it matters, cloud where it doesn’t. Patient records: AU-hosted, non-negotiable. Marketing site: Cloudflare, irrelevant.
  • Custom-built where it’s core, pragmatic everywhere else. The patient experience and billing engine are non-negotiable. Everything else should be the simplest thing that works.
  • Buildable by one person. If it needs a team to maintain, the architecture is wrong. A system one person can build is a system one person can understand end-to-end. That understanding is what lets you move fast.

The Bet

Per-minute billing isn’t just a pricing model. It’s the architectural decision that makes everything downstream simpler, more honest, and more aligned. The patient self-regulates. The doctor works without resentment. AI handles everything that isn’t the core clinical interaction. The whole thing runs on infrastructure one or two people can build and operate.

If that’s the kind of thing you’d want to build, let’s talk.