Sites that run both end up with something worth describing on its own terms, because the day-to-day of it is not the theory. There is one asset list, one place work is recorded, and two different reasons a job appears: an interval elapsed, or a reading crossed a line. Everything downstream of that, the scheduling, the technician's visit, the record, is identical. Getting the hybrid right is mostly about resisting the pull toward two separate systems, which is where most of them end up.
One list, one record, two triggers
The monitored assets are not a different kind of asset. They sit on the same list, they have the same history, and their jobs arrive in the same queue. What differs is only what caused the job to exist. Sites that let the monitoring vendor's platform become a second system find that the monitored assets develop a separate history, and the whole point of keeping a record was to have one place to look.
Keep an interval on monitored assets anyway
Monitoring fails quietly: a sensor drifts, a gateway stops reporting, a threshold was never set. If the interval was removed when monitoring was added, that asset now has no maintenance at all and nothing will say so. A long backstop interval on monitored assets costs very little and converts a silent failure into a visit that finds the monitoring is down.
The alert has to become a job, or it becomes noise
An alert that a person looks at and closes is not maintenance. The hybrid works when an alert creates a job in the same queue as everything else, with the same route, the same scheduling and the same record of what was found. Sites where alerts live only in a monitoring dashboard reliably reach a state where nobody reads the dashboard.
Questions people ask about predictive preventive maintenance
How many assets should be monitored?
Fewer than a first enthusiasm suggests. Start with the handful whose failure would be genuinely expensive and which give warning, prove the route from alert to job on those, then extend.
Can the monitoring platform be the system of record?
It can, and then it has to hold every unmonitored asset too, which most are not built for. Keeping the maintenance record as the system of record and letting monitoring feed it is the arrangement that scales.
Does a hybrid need more people?
It needs one person who can interpret readings, in-house or bought as a service. The scheduling and recording effort is essentially unchanged, because it is the same programme with an extra trigger.