The four integrations worth building, and the rest

Vendor integration pages are long because logos are cheap to add. In practice four connections change how maintenance work runs, and everything else is a file somebody exports once a month and nobody misses.

1. Sign-in, joined to your directory

Not exciting, and the one that quietly matters most. When somebody leaves the company they lose access to your maintenance record the same afternoon, without anybody remembering to do it. It also removes a password from the list of reasons people avoid the software.

2. Alarms and readings that raise work

A building management system, a telematics box or a temperature sensor that turns a threshold into a job with the asset already attached. This is the one that changes response times, because it removes the step where a human notices.

Build it narrow. Three alarm types that always mean work beats every alarm feeding a queue nobody reads.

3. Cost out to finance

Maintenance spend by asset and cost centre, monthly, in the shape the finance system already expects. This is what stops the annual argument about what the buildings actually cost to run.

4. Requests in from where people already are

A tenant portal, a service desk, a shared mailbox or a chat channel. People will not move to your tool to report a leak. Meet them where they are and put the result in one place.

What can stay a spreadsheet

• •Purchase orders, unless parts are genuinely holding up jobs
• •HR and payroll, unless somebody bills labour by the hour
• •Space planning, unless the floor plan changes monthly
• •Anything reported quarterly, which is a report and not an integration

One rule before you build any of them

Name the decision that changes when the data arrives. If nobody acts differently, you have built a pipe that moves numbers between two systems and costs somebody an afternoon every time either one changes.