How maintenance software works

Maintenance software is usually explained through its modules, and that doesn’t say much: they all use the same words. It’s understood much better by looking at what the system does on its own and what a person does.

That’s the real difference between a CMMS and a spreadsheet: it’s not that it stores information better, it’s that it generates work and demands.

What the system does on its own

Generates the checks. From the check routine tied to the asset, its model or an entire family, and from the defined frequency, it creates the work orders with their technician and date. It’s the difference between a plan that exists and one that actually happens.

Checks before generating. Whether that day is a holiday and whether the technician is available. Scheduling a check for August 15th isn’t planning.

Reacts to a reading. If the asset carries a counter —hours, cycles, kilometers— with a limit and a warning percentage, once a reading crosses the threshold the preventive order is generated automatically. The plan follows actual use, not the calendar.

Turns emails into requests. A mailbox the system empties and converts into incidents.

Warns about what’s about to expire. A daily check of documentation about to expire, and of dated materials and protective equipment.

Deducts material. When the order is closed, not in a later entry.

Notifies status changes to the right person, and only for the statuses you’ve chosen.

What a person still has to do

Worth saying just as clearly, because it’s where false expectations get created.

Decide what’s critical. Which assets stop production or service if they fail is a business decision.

Set the technical criteria. What gets checked and how often is set by the manufacturer, the standard or experience.

Prioritize. Who gets attended to first is decided by someone with the context in front of them.

Close the work. If an order could be closed without anyone confirming it was actually done, the history would stop holding up as evidence.

How work comes in

Through two channels, and it’s worth the system distinguishing between them because they get analyzed separately.

Planned work, generated by preventive maintenance.

What comes up, which enters as an incident with its type, priority, client, address and affected asset. From the backend, from the client access with a photo, or from the mailbox.

The order comes out of the incident and keeps its origin: that way you can always trace back from the completed work to the request that triggered it.

How it gets distributed

The calendar shows each technician’s load across the weeks, and lets you move work by dragging it. The most useful thing on that screen is what has no owner, because that’s what’s going to fall through.

And the technician can pick up unassigned work themselves, and sort their orders by route using GPS.

How it gets logged

On site, not in the office. From the app: a stopwatch inside the order itself, material consumed against the warehouse, a completed checklist —on preventive orders it won’t let you close without it—, photos and the client’s signature on screen.

Everything works offline: orders, assets and documents are stored on the device and actions go into a queue that empties once signal comes back. If one fails, it stays flagged with its reason instead of disappearing.

This is the point that decides whether the system actually works. If there’s friction here, logging gets pushed to the end of the day and stops being accurate.

What can be done with what’s recorded

With orders closed out with measured times and material logged, reports answer what used to be opinions: how much it costs to maintain each piece of equipment, which contract is profitable, what percentage of the plan has been executed, which assets concentrate the anomalies, how long it takes to respond and to resolve.

And there’s the evidence: legal maintenance —which isn’t a module, but this same mechanism with the check routine the standard requires— rests on the history, the checklists and the documentation with its dates.

And artificial intelligence

Bounded functions, each with its own switch: an assistant that answers questions about your assets and your history —also inside the app—, document reading that turns a supplier’s PDF into data, help drafting order text, and translation of your catalogs.

It proposes, it doesn’t execute: it doesn’t close orders or assign work. Covered in detail in AI in GMAO Cloud.

Who sees what

A cross-cutting aspect worth deciding early, because it runs through everything above: not everyone logs in the same way or sees the same thing.

In GMAO Cloud there are three distinct interfaces —not the same one with things hidden—: the web backend for whoever plans and administers, the app for whoever executes in the field, and client access for whoever receives the service.

Within each one, scope is controlled by profile, and every document carries its own visibility: whether the client, the technician or the supplier can see it. That turns a decision that repeats every week into a one-time configuration.

And a detail that shapes everything: if the software is billed per user, adding a temp worker, a subcontractor or a client becomes a financial decision, and the temptation to share accounts shows up. As soon as two people sign in with the same account, the history stops holding up as evidence. In GMAO Cloud licenses are unlimited across all three plans.

What to understand before buying

That the value isn’t in storing information, it’s in generating work and demanding. A spreadsheet stores just as well; what it doesn’t do is complain when March’s check never got done.

And that the system can only demand what someone has taught it: the check routine, the frequency and which assets are critical are set by a person. That initial configuration is the real work, and it’s the same no matter which tool you choose.

If you’d like to see it running on your own equipment, you can request a demo.