NADIR / Platform / 02 of 04

A queue, not a dashboard.

Detection is worth nothing if the work does not clear. Operate turns a tier into a named owner, a booked bay and a running clock — and shows you the places where the queue stops moving.

The list

Sorted by consequence, not arrival.

The queue is ordered by tier and then by age, which means the oldest row is frequently not the top row. That is intentional. A CAUTION event that has been open sixty-one hours matters less than a CRITICAL one opened twenty-two hours ago, and a queue that sorts by timestamp quietly optimises for the wrong thing.

Opening a row produces a work order rather than a chart. It names the vehicle and the repair order it is anchored to, what was measured and against which tolerance, what that error projects to at range, the action required, the hours remaining against the deadline, and the identifier of the evidence bundle backing all of it. It is addressed to a person, and it says so.

The work queue Tier first, then age. Unassigned owners are called out because an unowned flag is the most common way a queue silently stops clearing.

A work order Measurement, tolerance, projected consequence, deadline and evidence identifier. Advisory by design — it is not an ECU command.

Constraint

Capacity, not detection, is the bottleneck.

Once a fleet is monitored the limiting factor stops being whether drift is found and becomes whether a bay is free before the deadline. Eighty-four open flags against fifty-four bookable slots over two weeks is a scheduling problem, and pretending otherwise produces a product that generates work nobody can absorb.

The schedule is shown with its weekends and its full days, because the shape matters: capacity is tight in week one and opens in week two, which is precisely the wrong way round for a 168-hour deadline. Operate routes against real availability rather than proximity, which is why the third bay — the quiet one — takes the booking.

Bay schedule Fourteen days across three bays, weekends closed and full days hatched. The booking lands where there is room, not where it is nearest.

The console, at flag Tier escalation on a tolerance breach, with the subject vehicle picked out of the feed it arrived in.

Ownership

Who owns it, and from which hour.

An event without an owner does not clear. The ladder assigns one explicitly at each threshold: nothing at all while a vehicle is merely monitored, a depot lead at twenty-four hours, fleet maintenance at seventy-two, fleet safety at a hundred and twenty, and a report to the safety board past a hundred and sixty-eight.

The last rung is worth being precise about. OVERDUE is a reporting state, not an intervention — nothing in the platform escalates a vehicle out of service, grounds an asset, or contacts a driver. The system's final action is to make an unresolved item impossible to lose track of, and then to stop.

Escalation ladder Four owner changes across a 168-hour window. The top rung is a report, not an action — the platform has no authority to remove a vehicle from service.

Operate is deliberately unglamorous. It is a queue, a schedule and an ownership rule.

The reason it exists is that every fleet we spoke to could already list its problem vehicles and still could not tell us who was going to fix them, or by when.

Platform

The rest of the category.

01MonitorConnect the feeds, score in shadow 02OperateA queue, an owner, a clock 03ProveSigned evidence anyone can check 04ResolveClosing the loop, and its limits

Want this pointed at your fleet?

We are onboarding fleets and repair networks a few at a time. Tell us what you run and we will reach out.

Join the waitlist