Be useful first. Earn the right to predict.

Value starts with saving clinicians time and making every promise visible. Ranking and prediction come after the ledger is in real use and the checks pass — and even then they run silently until they earn the surface.

1. 01

   Be useful

   Capture the visit, then track what was promised

   The clinician's app takes the visit down - typed, dictated or from a document - and turns it into a structured note. What was promised out of that visit becomes a tracked item with an owner, a deadline and a clock running on it, so a promise that is slipping can be seen before it is missed. It runs on the machine in front of the clinician and keeps working without an internet connection.

2. 02

   Start measuring

   Find out what a miss actually costs

   Once promises are tracked, count how often they slip, where, and what happened next. Measurement has to exist before any ranking can be checked against reality.

3. 03

   Turn on prediction

   OLRHM + TRIS

   The ranking engine runs where people can agree what counts as a miss - and stays off the surface until accuracy, workload, fairness and regulatory checks pass.

4. 04

   Scale up

   Across a whole hospital

   Shared infrastructure and connections to hospital systems - once the core, the evidence, and the controls can survive someone checking them properly.

Fund the proof, not the costume.

No amount is named here. The briefing sets out what each stage costs and what has to be true before the next one starts.

1. 01

   Foundation

   The shared safety core, traceable records, and written requirements

2. 02

   Protection

   Patent scope, and finding out whether regulators treat this as a medical device

3. 03

   Evidence

   Getting experts to agree what counts as a miss, gathering data, independent checking

4. 04

   Continuity

   More engineers, someone owning quality, and a second site to test at