WED 12:45 - 13:30

This is a 2017 presentation

In this ON BOARD-session, we will look further into the realization of a data-driven system that takes decisions using the prescriptive analytics within the primary process of Schadegarant.

If you get damage to your car and the insurer of the vehicle is affiliated with Stichting Schadegarant, you can recover from one of the repairers within the Schadegarant network for damage repair. When the repairer offers a calculation of the repair, the insurer has the option to ask a damage expert to review and evaluate the damage and corresponding calculation.

The insurer has to make a decision if they want to consult a damage expert, for that they rely on Schadegarant.

The use of damage experts is costly, and Schadegarant and its members want to use as few “unnecessary” experts as possible. To optimize this, Xomnia has developed data-driven models using machine learning.

Based on many historical files, an expectation pattern can be drawn upon new cases. It is possible to estimate how much the costs are. Also, unexpected parts and repair methods can be recognized. In addition, Xomnia has also developed the ability to manually add business logic to the advisory system.

All this has been brought together in a system that is addressed through a web service, which, thanks to the deployment of the Google Cloud Platform (GCP), combined with Kubernetes, ensures high uptime and performance.

Sign up for this session here.

This session is made possible by Xomnia.