Finding the ‘next Max Verstappen’ with data and smart algorithms.
Van Amersfoort Racing
Uptima
Predictive maintenance for spare parts forecasting.
Predict which parts you’ll need before shortages or unnecessary stock build up.
Trusted by leading organisations
Why we built this,
Maintenance organisations need to have parts available before work begins. But it is not always easy to predict which parts will be needed and when. Demand depends on maintenance patterns, parts usage and how assets are used in operations.
Too little stock can delay maintenance work and leave assets out of service. Excess stock, on the other hand, ties up capital unnecessarily and can lead to parts that are rarely used.
Uptima analyses historical maintenance data and parts usage to predict future demand for spare parts. This helps support stock and purchasing decisions.
What you can expect,
Insight into future parts requirements
The tool uses available maintenance and consumption data to predict which parts may be needed in the period ahead.
The quality and accuracy of these predictions depend on the available data, maintenance patterns and how predictable demand is.
Support for inventory and purchasing decisions
Predictions show which parts may be in short supply or surplus. Supply chain and maintenance teams can use this information when planning orders.
For example, the tool can provide insight into:
- Expected demand for each part
- Differences between predicted demand and available stock
- Parts at increased risk of a shortage
- Changes in usage patterns
- The reasoning behind ordering decisions
The standard forecasts, planning horizon and available reports are yet to be determined.
Predictions as a guide, not a definitive answer
Unexpected failures or operational changes can still occur. That is why we treat a prediction not as a certainty, but as additional information for planners.
Staff can take account of lead times, critical parts and practical knowledge that is not fully captured in the available data.
Our approach,
-
Explore the challenge and available data
We identify which parts are used, how inventory is planned and what historical maintenance data is available. We also determine which parts would have the greatest impact if there were shortages or surpluses.
This lets us focus the first application on a specific inventory problem. -
Develop and test a predictive model
We develop a model using the available data and test its predictions against historical situations. We then compare the results with the existing planning method.
We look not only at technical accuracy, but also at whether the predictions help staff make better inventory decisions. -
Integrate it into the planning process
A prediction only delivers value when staff can use it at the right time. That is why we integrate the results into the existing inventory or purchasing process.
After the first application, the solution can be extended to additional parts, assets or locations.
Not a prediction for its own sake, but a practical tool for inventory and maintenance decisions.
This helps planners respond sooner to anticipated demand, without relying solely on fixed inventory rules or experience.
What we’ve built before,
Three projects from our work with data and AI.
Further reading
Blogs and in-depth articles from our team.
24 Aug 2026
Three lessons from our AI projects
What we learned from dozens of AI projects: start with the decision, treat data quality as an organisational issue and arrange management in advance.
18 Aug 2026
From pilot to production: why most AI projects stall
Most AI projects stall not because of the technology, but because of data, teams and processes. Read what it takes to move from pilot to production.
25 Jun 2026
Responsible AI: from principles to practice
AI is developing rapidly. New models emerge almost every week, and more and more organisations are experimenting with AI.
Get in touch
Stef Stoelwinder
Business DevelopmentE-mail info@datasciencelab.nl
Phone +31 (0)20 3636 724
Frequently asked questions,
Historical data on maintenance, parts usage and stock provides an important foundation. Depending on the question, lead times, asset information and operational planning may also be relevant. We start by examining what data is available and usable.
The tool is primarily intended to show expected demand and potential shortages. Whether order recommendations or automatic ordering become part of the process depends on how you want to set it up and which systems are available. This functionality has yet to be confirmed.
That depends on the amount of data and the demand pattern. Reliable predictions are harder for parts with little historical activity or highly unpredictable demand. In those cases, fixed stock rules and practical knowledge may remain important.
Maintenance patterns, assets and business processes change. Over time, a model may therefore become less accurate. Comparing predictions with actual demand shows when the model needs adjusting or retraining.
Technical accuracy alone is not enough. The forecast must contribute to a relevant business goal, such as fewer urgent orders, less downtime or stock levels that better match demand. That is why we agree in advance which outcomes matter and compare them with the existing way of working.