Predictive Maintenance

Use Case

Why do you need to start using Predictive Maintenance to keep the vehicles of your fleet on the road

Predictive Maintenance is all about understanding the health status of vehicles to determine if, when and which components need to be repaired.

In order to be able to implement a system with predictive alerts,

Maintenance Teams need to have access to clear dashboards to visualize vehicle data and sensor readings that measure several parameters, such as wear and tear or temperature.

The traditional approach to fleet management is to take vehicles away from services for regular maintenance interventions with intervals based on mileage, which is often not always necessary nor cost-effective.

On the other hand, AI-enabled Predictive Maintenance allows Maintenance Managers to make repair choices based on the vehicle’s current and expected future condition, rather than on pre- scheduled time intervals.

Whether it’s an oil leak or an accident, all fleets will inevitably experience downtime, and even with well-managed preventive maintenance programs in place, there is a loss of revenue when drivers aren’t on the road transporting people or goods.

Some of these costs are difficult to accurately budget for and are avoidable with today’s technology, the ones when a driver and vehicle are sidelined and the company incurs in unexpected costs due to unplanned vehicle downtime.

Reducing downtime is therefore key to saving money, preventing revenue loss and keeping the vehicles on the road making money, being a critical aspect to ensure a fleet runs efficiently and contributes to optimising the productivity of the entire company.

Reducing downtime is therefore key to saving money


There are literally thousands of fault codes that can be broadcasted over the several networks and each one could indicate up to 25 different failure modes, making the task of filtering them by severity and sorting them by order of importance almost impossible for the Maintenance Teams without the help of technology and automation. 

Artificial Intelligence is key to joining the dots and putting these data points together in order to create actionable insights with the objective of issuing predictive alerts to warn about imminent failures of key vehicle components or systems. 

Without this ability to look into the future Maintenance Teams are limited to the old ways of preventive maintenance, losing efficiency and not being able to avoid high rates of unplanned vehicle downtime arising from situations that could have been easily avoided. 

A company can save an average of 15% in costs related to parts on brake pads alone if replaced at the right time rather than based on its mileage, or increase its overall productivity up to 10% by minimising the time a vehicle is sitting in the workshop for days waiting to be diagnosed and repaired instead of just a few hours when the problem is identified before it becomes critical.

On top of that, vehicle downtime is expensive and the unpredictable failure of a key component can compromise delivery deadlines or journeys, putting public or cargo transport fleet operations out of action for several days.

Drilling down on the real cost of vehicle downtime we found out that, for example, one of our customers from a cargo transport company saved more than


in one year by acting on the predictive alerts related to our air compression model.

With Predictive Alerts resulting from the accumulated expertise of our automotive analysts combined with state of the art technologies like Artificial Intelligence, it is possible to anticipate unpleasant surprises and implement enhanced maintenance plans in all sorts of fleets.

Case Study

The case study below refers to the same long-haul truck company of 750 vehicles and their data in two consecutive years. The air compressor machine learning model used by Stratio initially successfully detected around 50% of all issues without false alerts, and as the model kept learning its accuracy also improved. The effects of this new approach are reflected in the difference between both years – when machine learning was implemented to support maintenance teams.

Year 1 (no predictive model analysing air compressors State of Life)

Total Cost of Planned Repair = 1.300 x 25 = € 32.500
Total Cost of Unplanned Repair = 3.950 x 79 = € 312.050 Total Repair Cost = 32.500 + 312.050 =

= € 344.550

Year 2 (with predictive model analysing air compressors State of Life)

Total Cost of Planned Repair = 1.300 x 68 = € 88.400
Total Cost of Unplanned Repair = 3.950 x 49 = € 193.550 Total Repair Cost = 88.400 +193.550 =

= € 281.950

Total Savings = 344.55 – 281.950 =

= € 62.600

In regards to the example shown above, a few important notes:

• The figures were rounded up to improve readability.
• The value for labour / hour is the average for continental europe.
• The profit is an estimate of the gains from around 50% accuracy in fault detection.
• Each company can apply their specific information and start using this formula to assess unplanned downtime costs with more precision.

In Year 1 the fleet had 104 vehicles with air compressor malfunctions, 79 of those were road-side breakdowns, causing unplanned downtime. The total repair cost during this year for air compressors was € 344.550

During Year 2 machine learning models detected in time, and prevented, 68 road-side breakdowns from a total of 117 events. This allowed the operator to save € 62.600

The Solution

Today, fleets can take advantage of a technology that provides insight and real-time data. This solution can help improve vehicle maintenance, service scheduling and eliminate unplanned uptime by identifying issues before they become a problem.

This technology is called Predictive Maintenance and Stratio, the AI- Powered Fleet Automation Platform, is the leading solution on the market, predicting malfunctions and notifying Fleet Managers of potential failures well in advance.

The type of data Stratio can capture offers maintenance technicians full diagnostic information in advance to become more efficient when the vehicle enters the workshop for repairs. The high quality data captured by Stratio allows  Artificial Intelligence models algorithms based on machine learning to automatically learn from past data without the need of human intervention.

Our Tools

Our Research team is constantly developing new models to increase the scope of the Stratio Platform ability to monitor increasingly more key components and eliminate unplanned downtime in the fleet of our customers. This is the current list of components covered by the Stratio Platform with Artificial Intelligence models to predict failures:


  • Brake pads
  • Starter battery
  • Available engine torque
  • Battery pack (Electric Vehicles)
  • Turbo  pressure
  • Cylinder compression
  • Engine oil compression
  • Air compressor
  • Wheel speed sensors
  • Air leaks

We aim at being able to ultimately cover all the systems of all types of vehicles to ensure a future with no surprises and no disruptions.

Top Benefits

Reduce unplanned downtime and risks

Predict vehicle failure and initiate action to avoid costly breakdowns

Reduce maintenance costs

Avoid over-maintaining assets by enabling scheduled maintenance before failure based on early warnings.

Improve vehicle utilization

Enable more efficient use of existing assets when you can predict maintenance issues and reduce failures on the job.

Extend asset life

Identify operational performance factors and improve maintenance practices and reliability.

Increase production output

Take full advantage of having the vehicles on the road transporting people and goods.

Why Stratio

We believe that in the future no fleet business will succeed unless it is fully automated. 

We deliver the industry’s most powerful and integrated predictive maintenance solution, data-driven operations control, and the only intelligent ecodriving system available for multi-brand fleets. 

We will lead the transportation industry towards a zero downtime future, enabling transport companies to serve more customers, better, at a lower cost.