Saving hundreds of thousands GBP
Significant cost savings by earlier fraud detection, every year
Detecting more fraud cases than previous rule-based solution and prioritizing them
Less false positive
Saving time and bettering customer experience by lowering number of false positive alerts
"DataSentics have provided invaluable support to Esure Group for 18 months. Hard working, diligent and always starting with deep engagement with stakeholders, the team were able to grasp very diverse business challenges and leverage state-of-the-art data science to solve them and drive an impact across the enterprise."
About the client
Esure Group plc is a leading personal lines insurance company providing a wide range of insurance products, including home, car, multi-car, and travel insurance. Serving over 2.35 million customers. Their trusted brands include Esure, Esure Broker, Sheilas' Wheels, and Sheilas' Wheels Broker.
Esure Group faced the persistent challenge of the insurance industry, claim fraud. While collaborating with DataSentics, Esure went on its digital transformational path. By harnessing cutting-edge machine learning on Databricks, they integrated digital behavior data, unveiling a new fraud detection model. The goal was to decrease false positive rates as well as catching fraud earlier in the process, which ultimately makes the process more effective and less costly.
The cost of inefficient fraud detection
The insurance industry has long grappled with the challenge of claim fraud. It affects customer satisfaction through delayed payouts and extended investigations, adding stress. Furthermore, investigation costs and pressure from regulators compound the issue. Improper payouts not only dent profitability but also foster similar behavior among policyholders. The existing fraud model suffered from a high false positive rate. It operated with traditional product-centric data, leading to delayed fraud detection and increased operational costs.
Enhanced fraud detection with digital insights and advanced AI techniques
We examined how people purchase insurance online using digital data. To achieve this, we employed cutting-edge machine learning and artificial intelligence techniques, specifically LightGBM. These advanced methods were utilized on top of the Databricks lakehouse platform and AWS cloud to enhance our approach.
We improved the traditional rule-based fraud model by enriching it with new digital data. This innovation was then integrated into the financial crime team's operations. It refined their strategy for handling fraud by making it easier to prioritize and address new cases.
Serving 2 millions of clients, and saving hundreds thousands pounds yearly
The solution is now operational for over 2 million Esure clients, resulting in significant cost savings. By detecting potential fraud cases earlier, the model contributes to annual savings amounting to hundreds of thousands of GBP. This not only emphasizes the model's concrete impact but also highlights its role in enhancing the organization's financial health.