I have seen this pattern more times than I can count. A company invests €150,000 in a CMMS. Six months later, the maintenance team is still using WhatsApp to report breakdowns. The CMMS has perfect compliance reports. The equipment reliability is falling.
Nobody is lying. The CMMS reports are accurate. The equipment really is breaking down more. Both things are true simultaneously — which is exactly the problem.
The CMMS was built for a different audience
Most Computerised Maintenance Management Systems were designed with IT departments, procurement teams, and compliance auditors in mind. They are excellent at producing reports, tracking work order histories, and satisfying audit requirements.
What they were not designed for is the technician standing in front of a machine at 2am who needs to know whether this vibration pattern has happened before, what was done about it, and whether it worked.
"CMMS: 97% compliance. Equipment reliability falling. Avarias a aumentar."
That caption describes a real situation I encountered at a packaging plant. The maintenance manager was proud of his CMMS compliance rate. The operations manager was pulling his hair out over unplanned downtime. Both were right. The CMMS was being used exactly as designed — to log and track. It was just not being used to prevent anything.
The three gaps that CMMS doesn't close
Gap 1 — The knowledge gap
When an experienced technician retires, they take decades of tacit knowledge with them. They know that Machine 7 always vibrates differently when the ambient temperature drops below 12°C. They know that the bearing on Line 3 needs to be checked every 340 hours, not every 500 hours as the manual says, because of the way that line is loaded.
A standard CMMS stores work orders. It does not capture this kind of contextual, experiential knowledge. It cannot answer the question "what does this symptom usually mean on this specific machine in this specific plant?"
Gap 2 — The language gap
CMMS interfaces were built by software engineers for software engineers. The logic is hierarchical: asset → work order → task → spare part. Technicians think differently. They think in symptoms, machines, and outcomes. They think in the language of the floor, not the language of a database schema.
The result is that data entry becomes a compliance exercise rather than a knowledge-building act. Technicians log the minimum required to close the work order. The system fills up with technically complete but practically useless records.
Gap 3 — The connection gap
Maintenance does not exist in isolation. An equipment failure affects OEE. OEE affects throughput. Throughput affects customer delivery. Customer delivery affects revenue. A CMMS tracks the equipment failure. It does not show the cascade of consequences — or the cascade of leading indicators that could have predicted it.
What a truly operational maintenance system looks like
At Logoplaste, across 68 plants in 17 countries, we learned that maintenance excellence is not a software problem. It is a knowledge management problem, a leadership problem, and a system design problem — in that order. The software comes last.
A truly operational maintenance system has four characteristics that most CMMS implementations lack:
- It speaks the language of the floor. Technicians can report symptoms in plain language. The system translates, not the other way around.
- It captures contextual knowledge. Not just what was done, but what was found, what was suspected, what was ruled out, and what the experienced technician thinks is happening.
- It connects maintenance to performance. Every work order is visible in the context of OEE, downtime trends, and production impact — not in isolation.
- It learns. When the same failure mode appears across multiple plants, the system recognises the pattern and triggers a preventive response across the network.
Where AI changes the equation
The Capabilium TPM Intelligent System was built specifically to address these gaps. It is not a CMMS replacement in the traditional sense. It is a layer of operational intelligence built on top of whatever data infrastructure already exists.
A technician can describe what they are seeing in plain language — "the pump on Line 2 is making a noise it doesn't usually make and the pressure is slightly lower than normal" — and the system can cross-reference that description against maintenance history, similar symptoms at other plants, known failure modes for that equipment type, and current production context.
This is not about replacing human judgement. It is about giving human judgement the information it needs to be right more often.
"The goal is not to automate maintenance decisions. The goal is to make maintenance knowledge available to everyone who needs it, at the moment they need it."
The compliance trap
One of the most dangerous things about a well-implemented CMMS is that it creates a convincing illusion of control. Compliance rates are high. Work orders are closed on time. Audit trails are clean. Everything looks like it is working.
Until the equipment breaks down unexpectedly. And then someone goes back through the CMMS records and finds all the signs that were there — logged correctly, closed correctly, never connected to each other.
The CMMS did exactly what it was designed to do. It just was not designed to prevent the breakdown you are looking at.
The question worth asking in your organisation is not whether your CMMS is being used correctly. It is whether your maintenance system is making your equipment more reliable. If the answer to the second question is no — regardless of what the answer to the first question is — you have a system design problem, not a compliance problem.
Vítor Vila Verde is the founder of Capabilium Partners, a Lisbon-based operational intelligence consultancy. He spent 20+ years in industrial transformation, most recently as VP Strategy Deployment & Operational Excellence at Logoplaste, overseeing 68 plants across 17 countries.