Aven Grid

A monitoring problem becomes a prioritisation problem.

The scenario brief described 18,742 alerts in a month, with 61% clearing themselves within five minutes. The challenge was to distinguish work that needed a person from noise that could wait.

I treated those figures as design inputs, not research findings. Working without access to operators, I focused the concept on triage and ownership. Repairing equipment and the after-hours response were outside the scope.

A queue organised around responsibility.

The overview answers three questions: does this need attention, how urgent is it, and who is responsible? Unassigned incidents move to the top. Opening a record preserves the operator’s place in the queue.

Ownership belongs with the action.

I explored eight layouts. The first four positioned ownership as a separate field; the next round tested whether the action row, heading and timeline could carry it more naturally.

The chosen direction removed the duplicate field. It distinguishes an unassigned incident, your own work and an incident owned by someone else. The primary action changes with that responsibility.

Designing the states beyond the ideal flow.

The interface also needs to work when information is loading, the connection disappears, access is read-only or an incident closes. I mapped these states using the same components as the prototype.

Reviewing the disconnected state exposed a mismatch: acknowledgement was blocked while reassignment was still available. Both require a connection. This is the kind of inconsistency the state review was designed to catch.

From the whole fleet to one site.

The final prototype connects four surfaces: the overview, queue, sites and incidents. A site with open work gets detail; quiet sites are accounted for without filling the page with inactive rows.