A maintenance request that moves fast without bypassing the controls

Maintenance is where most agencies feel the operating-model problem first, because it is the workflow with the most participants and the least structure. A tenant reports something. Somebody reads it, forms a view, rings a contractor, and moves on. The record of what happened is distributed across an inbox, a phone, and a memory.

That worked when the consequence of a slow response was an unhappy tenant. Under the current regime the consequence can be a statutory timescale you have already missed by the time anyone notices.

Speed and control are usually traded against each other

The instinct when response times are poor is to remove steps. Fewer approvals, faster dispatch, less admin. It works, and it works right up until the first case where the step you removed was the one that would have caught something.

The alternative is to make the controls cheap rather than optional. A check that runs automatically and surfaces only when it fails costs the property manager nothing on the ninety-nine requests where everything is fine.

Why the classification step is the important one

A tenant does not write “Category 1 hazard, damp and mould, first-floor bedroom”. They write that the wall in the back room has gone black again and the baby has a cough.

Turning the second into the first is a language problem, and it is the single highest-leverage automation in the whole workflow — because the classification is what determines which clock starts. Damp and mould reported in a property with a young occupant is not the same triage outcome as a dripping tap, and the difference has to be made at intake rather than three weeks later when somebody reviews the queue.

The evidence half, which nobody asks about until they need it

The reason to run maintenance through a system rather than a phone is not the dispatch. It is that eighteen months later, in a dispute or an enforcement enquiry, the question will be what you were told, when you were told it, what you concluded, and what you did.

Every mutation writes to a hash-chained audit log, which means the trail is not merely present but tamper-evident. That distinction matters exactly once, and it matters enormously.

The commercial tail, built where clients needed it

The core workflow above runs from intake through triage and hazard classification to contractor dispatch. What happens after dispatch — the money side of a job — differs enormously between operators, which is precisely why we treat it as bespoke rather than shipping one opinionated version of it to everybody.

Where a client’s operating model has required it, we have built that tail out: quote capture before instruction, appointment scheduling against contractor availability, and invoice retention tied back to the job so the cost trail sits alongside the evidence trail rather than in a separate finance system.

Approval thresholds and signatory limits are the same category of work. For groups where spend above a figure has to route to a named approver before a contractor is instructed, the threshold, the approver and the escalation path are configured to that group’s delegated authority — and a job that trips the threshold pauses rather than proceeding, with the reason recorded.

What the bespoke work usually is

In practice, three things. An intake route from wherever requests actually arrive for you — which is frequently not a portal. A classification tuned to the stock you manage, because a purpose-built block and a Victorian conversion do not fail in the same ways. And the reporting the other side of it, so a head of property management can see what is running late while there is still time to act.

If maintenance is the workflow costing you the most, it is usually the right one to phase first — the risk is concentrated there and the improvement is measurable within weeks. We are at hello@tekniti.ai.