Smart Alerts

The complaint that reaches you first is already too late

A tenant tells the receptionist. The receptionist mentions it to facilities. Facilities emails the account manager. The account manager forwards it to the supervisor. Somewhere in there it gets fixed, or it doesn’t, and nobody can say precisely when anything happened.

By the time a problem has travelled that path, two things are true: it has been visible to building users for a while, and the record of it is a chain of forwarded messages nobody will be able to reconstruct in a month.

One tap, straight to the people who can act

Your team raises an alert directly — no phone tree, no email chain. The reason is categorised, because a restroom out of supplies and a spill in reception need different responses and shouldn’t arrive as identical free text.

It reaches the assigned team immediately by email, SMS and push at the same time. Not one channel with fallbacks, all three — the failure mode we’re designing against is a single notification arriving on a phone that’s on silent.

The window belongs to the contract

Response time is defined by your service level and measured against every alert. It is not a number we put on a website and hope nobody times.

That’s a deliberate structure. A clinic and a distribution centre have genuinely different requirements, and a single advertised response time either overpromises for one or overprices the other. Yours is agreed, written down, and reported against.

An alert approaching its window escalates automatically. That matters more than it sounds: on a difficult night, the thing that gets forgotten is not the problem itself, it’s the decision to escalate it. Removing that decision from human judgment is the point.

Only you can close it

An alert is resolved when the person who raised it confirms the response was adequate.

We cannot mark our own alerts solved, and that constraint is the one that makes the whole mechanism worth anything. Any vendor that can close its own tickets will report excellent response statistics — the numbers measure the vendor’s opinion of itself. Requiring your confirmation means the metric measures whether the problem actually went away.

Alerts and their response times feed the account’s operational record and its Protex Verified score, alongside corrective actions. A pattern of alerts from the same area is treated as a recurring issue — the fix is upstream, in the scope or the stock levels, not in responding faster to the same failure every week.