AI for leasing companies

For leasing companies, AI means training a model on data you already have: contracts, sale proceeds, work orders, tyre changes and parts in the workshop. You can use it to estimate residual values more accurately, spot unusual wear earlier and identify which parts the workshop needs. These are among the AI applications we build.

You can find more of our work in the sector under mobility and logistics.

With a model trained on cars you have already sold: make, model, fuel type, mileage, contract term, damage and the proceeds from remarketing. The RDW publishes open data for each vehicle registration, including the list price, date of first registration, an assessment of the recorded mileage and outstanding recalls.

The challenge is a new model or powertrain with no sales history. In that case, the model draws on comparable cars, and you factor that into the quote. Read more about forecasting.

By bringing together tyre changes and mileage for each vehicle registration. For Athlon, we used Python, Snowflake and Qlik to build dashboards that give the asset management team detailed insight into the fleet’s tyre use and wear patterns, so it can identify and address anomalies.

A predictive model is the next step: identifying which cars will need new tyres sooner than expected. Read more about business intelligence.

Yes, if work orders are recorded by vehicle registration and type of work. Under a closed-cost arrangement, the leasing company bears any higher costs; under an open-cost arrangement, the difference between estimated and actual costs is settled at the end of the contract. In both cases, a better estimate helps when preparing a quote.

The challenge lies in the work orders themselves: each dealer uses different codes, and free text must first be categorised. If you want to anticipate faults, read about predictive maintenance.

All participants in the Keurmerk Private Lease scheme register lease contracts with the BKR. An AI system that assesses the creditworthiness of natural persons is classified as high-risk under Annex III of the AI Act; according to the European Commission, those rules apply from 2 December 2027. Systems used to detect financial fraud are excluded.

Your legal counsel or compliance department determines whether your approval model falls into that category; we can help you assess it. See the AI Act and Responsible AI.

By monitoring and retraining it. A new generation of electric cars, a price change by a manufacturer or a shift in demand in the used-car market changes what a car will sell for. The prediction then becomes less accurate. This is called drift.

Decide in advance who receives an alert and who approves a new version. You can read how this works under MLOps; your team can manage the model, or we can do it through management and continuous improvement.

When the main question is what happened last quarter, rather than what will happen next. In that case, a Power BI dashboard using your contract and work order data is enough. A fixed allowance in your cost calculations may also be sufficient if a particular model performs the same way every year.

If contracts, work orders and sale proceeds sit in separate systems without a shared vehicle registration identifier, start with the data foundation. If a simpler solution is a better fit, we’ll tell you.