Preventive vs predictive fleet maintenance.

In passenger transport fleets, an unplanned breakdown is more than an inconvenience: it is lost revenue, a contractual fine, and a hit to your reputation. The difference between preventive and predictive maintenance is the moment of action — and the information available at that moment.

Preventive maintenance

Parts replaced or inspected at fixed intervals: every 10,000 km, every 6 months, and so on. It is a calendar, not a diagnosis — and it works well when the fleet is homogeneous and usage is predictable.

Limitation: a vehicle may need attention before the interval, or after, depending on how it is actually used. Preventive maintenance does not see that.

Even so, it stays the backbone of the operation. Skip preventive work and your history becomes unreliable — and weak history supports nothing further down the line.

Predictive maintenance

Replacement based on the vehicle's actual behaviour: risk score, failure history, wear indicators. It anticipates breakdowns. It works when the fleet is heterogeneous, usage varies, and the cost of downtime is high.

Advantage: you invest where the risk is highest, not where the calendar dictates.

Guesswork is not the point here. Someone approves the work order: the machine ranks the queue, the person in charge decides what goes into the workshop.

When to use each

The question is not which one is better. It is which one fits your operation today:

In real operations the two live side by side: preventive for everything required by regulation or policy, predictive for what actually takes a vehicle off the road.

How to build the data base

Predictive maintenance never starts without records. The bare minimum: work orders with date, odometer reading, part replaced, cost, and reason for the visit — including preventive visits.

This is data analysis in the plain sense: shaping history until it answers "where should I look first". The Google Cloud guide to data analytics describes that step from raw log to decision well.

If your records live in scattered spreadsheets, the first win is not a tool — it is filling in the same form every time. After that, the data needs to show up in one place, which is exactly what a fleet maintenance dashboard is for.

How AI helps prioritize

Pirbrain Fleet calculates a risk score per vehicle (0–100) combining maintenance history, item type, and days since the last replacement. The result is a prioritized list: where to invest before the breakdown.

It is not magic: it is organized data, clear rules, and a method the team reviews. The assistant runs the calculation; the person responsible approves the work order.

Frequently asked questions

Does a small fleet need predictive maintenance?

Not necessarily. Well-run preventive maintenance is enough when usage is predictable and downtime is cheap. Predictive work becomes worth it when downtime gets expensive, when vehicles differ widely from one another, or when the same failure keeps repeating with no clear reason.

How much history do I need?

There is no magic period: you need to cover at least one full cycle of the items that fail most in your fleet. If you are recording nothing yet, start now — in a few months that record is what separates decisions from guesses.

Does the score replace the mechanic?

No. The score only orders the attention queue and makes the priority visible to the whole team. The diagnosis, and the decision to open a work order, stay with people.

What if the team does not log the jobs?

No records means no score, and no score means no predictive maintenance — just opinions with a spreadsheet open. Start with a single work-order form filled in at the time of service, and only then think about software.

Once the data lives in one place, you can see the whole picture: read spreadsheet or dashboard: how to decide and pick the format your team will actually use.