The most expensive mistake in healthcare AI is not a bad model. It is building the right tool for the wrong problem — confidently, for months. We use a fixed four-week discovery to make that mistake cheap to avoid, and to leave you with a clear decision whether or not you build with us.
Why four weeks, and why fixed
Open-ended scoping rewards the wrong things: more meetings, more scope, more cost, and a slow drift away from the original problem. A fixed window forces the discipline of deciding what actually matters. The fee is fixed too, so the incentive is clarity, not billable hours.
Week one — framing the problem
We start with the people, not the technology. Who feels the problem? What does a bad day look like for them? We map the workflow as it really runs, including the workarounds nobody documents, and we agree on what a good outcome would actually change.
Week two — clinical and data reality
A solution that cannot be deployed safely or supplied with data is not a solution. This week we assess the clinical-safety implications, the data that exists and how it can lawfully be used, and the systems any tool would need to live inside — SystmOne, EMIS, ePMA and the rest.
Half of the ideas that look brilliant on a whiteboard meet a data or safety reality here. That is the point.
Weeks three and four — design and a costed plan
With the problem framed and the constraints understood, we design the solution and cost it honestly. You leave the four weeks with:
- A clear definition of the problem and who owns it
- A clinical-safety and data assessment
- A solution design, sized to the real workflow
- A costed, staged plan — discovery, pilot, production
The deliverable is a decision
The output of a discovery is not a deck and it is not a commitment to build. It is a decision you can defend — to your board, your clinicians and yourself. Sometimes that decision is not yet, or not this. We would rather you reach it in four weeks than in four months.