Case study · Field inspections
A property operations client standardised field inspections on an established industry platform, supplied through their management system. The platform produced good forms and clean reports. What the account could reach did not carry the part the operation actually ran on: which findings need a manager, who is chasing them, and whether anything was left undone.
The problem
Before the change, inspections ran on tools built for this operation: its properties, its standards, its exception handling. The client then chose an industry platform for field inspections, which is a reasonable thing for an operator to want. Staff get a maintained mobile application, a supported form builder and reports that look the same every time.
The gap was not in the forms. It was in what happens after a form is finished. The integration available on that account did not expose inspection completion events or an inspection interface to read from, so nothing downstream could see that a report existed, let alone that one line in it needed a manager that day. Exception routing, management alerts and a running task list all lived on the other side of that boundary.
The vendor sells the access that would close it on their top tier, published at four hundred dollars a month. That is the list price of the tier, not this client's bill. For an operation this size it buys one capability they already had, in software already written and already in use.
What was built
Completed inspections are exported as reports and land in a shared drive folder. A job runs every five minutes, converts each new report to text, reads it, and records what it found. Files are identified by their own storage identifier, so the same report processed twice produces one record rather than two.
A correction already made is logged as a defect that was cured. A failed or escalated item requires management review. Consumables marked low or depleted raise restock follow-up, with depleted ranked higher, and an item nobody checked is recorded without blocking anything. One consolidated notice goes out with the report attached, rather than a message per finding.
Each report is filed by property, then inspection type, then year, then month, derived from the report itself. Nobody decides where it goes, so nobody has to remember the convention, and the archive stays navigable by someone who was not there.
Findings become assignable work in a separate application behind Workspace sign-in: dashboards, assignment and reassignment, dated activity history, completion and deferral, and an installable mobile view for the people doing the work. Review is targeted within twenty-four hours of the inspection.
When it went wrong
One report hit a transient conversion error from the storage provider and was not read on its scheduled pass. The next run picked it up and processed it normally. Nothing was lost, and nothing was reported as handled when it was not.
The repair was in what gets escalated. A transient conversion failure now waits for a second consecutive failure on the same file before raising an operations alert, because one retry is the normal shape of that error and paging a person for it trains them to ignore the channel. Every other failure alerts immediately. The distinction is the point: an alert that fires on noise is worse than no alert, and a quiet failure is worse than both.
Why it stays hybrid
The field team uses the platform the client chose, with its own mobile application and its own support. The operation keeps the exception log, the review rules and the task list that were built around how this business actually runs. Neither one had to be abandoned to get the other.
The application this replaced is still in place as the way back. A migration that leaves no route back is a decision disguised as a project, and the point at which you discover the new arrangement does not fit is the point at which you most need the old one working.
What is not claimed
The client, the platform, the management system that supplies it and the properties are not named, and no inspection content, photograph or finding appears here. What is described is the arrangement, not the operation's records.
The four hundred dollars is the vendor's published price for the tier that carries their interface, read from their pricing page on 22 September 2026. It is not this client's invoice, not a negotiated figure, and not a savings claim: what it establishes is that the capability was available for money and was built instead.
The vendor's own task records are not written to. Follow-up lives in the operation's task register, which can hold a link to a task raised in the vendor's system by hand. The property record of authority is unchanged.
Let’s talk