Experience

No pricing model, no cost tracking, and a per-generation cost that rises with each model release instead of falling. The first months went into instrumentation, because you cannot price what you cannot see.

Credits came out of that. A unit designers can budget in, metered on work rather than on time. They use the tool in bursts around deadlines, so a seat price or a daily cap punishes exactly the professional you want to keep.

A dispatcher mid-incident was working across six systems that did not share data. Dispatch in one, the caller’s history in another, which vehicle was where in a third.

We built one OS across dispatch, intake, records, personnel and assets, with incident detection running on edge compute at the camera instead of a cloud round trip.

Customers told us their existing systems took new dispatchers one to two years to ramp on. The workforce is ageing, a typical operator is around fifty, and the people replacing them are less technical than the software assumes.

We built for recognition rather than training. Consumer patterns, few modes, nothing behind a shortcut you had to be taught. A full team was running on it inside four days.

A co-founder ran the sale. I built the demo and sat in on the calls and the on-sites.

Instrumenting it meant we could see where people stalled, what they asked about twice, and what they went looking for and did not find. That, plus watching real dispatch floors, is where most of the roadmap came from.

I wrote the spec for the edge detection and built the environments we tested candidate models in. Someone else wrote the production pipeline.

Same split as the model matrix at Vizcom. I would rather own what counts as good than own the implementation.

Pod cost set how many pods existed, and how many pods existed set everything else.

Most of the saving sat upstream of the bill of materials. Certifying to UL from the start instead of retrofitting it. Designing for the tens of units we were actually building rather than the hundreds in the plan. Cutting internal and external panelling specified well past what the enclosure needed. The largest single line was the body scanner, where a spinning rotating rig was replaced with a fixed solid-state one. Back of house dispensed medication with nobody in the room.

A CarePod is a clinic with no clinician in it. Every capability question is really a question about what someone will do alone in a box with nobody to ask.

I defined the experience from onboarding through repeat visits and decided which physical capabilities earned the space and the cost. Several we scoped and cut. I was the main line to the CEO on concepts.

Each pod was capital and lead time before it was a member. Five or six reached the field.

The member number people quote is the clinic’s, not the pods’. What the pods controlled was the ceiling, which is why the work went at unit cost.