Insurance Telematics: Risk Assessment of Connected Vehicles

Abstract

Insurance Industry, especially the Motor Insurance is at an inflection point. Insurers are struggling to reduce their 'Loss Ratios'. With the connected and autonomous vehicles, the underwriting is becoming more complex as there are more 'rating variables'. Even today, vehicles with ADAS and Telematics get discounted premiums, but there is no model to calculate the `likelihood of an accident of that vehicle', now that ADAS has made the vehicle safer. In this paper we have used Visual Telemetry to identify those rating variables and based on our current ongoing pilots we have created a scoring system for those rating variables. Our research objective is to build an underwriting model which will help actuaries to calculate the risk of every vehicle. These scores help us to calculate the 'likelihood of accidents'. Using the likelihood and legacy data from the insurers, we then calculate the 'claim frequency' and 'claim severity'. The insurance premium is then calculated as the product of claim frequency and claim severity. We are trying to gather more data from our clients like ADOBE & BAJAJ Allianz who are using our ADAS and consecutively improving the scoring techniques and scores. This scoring methodology is very specific to Indian Driving and Traffic data. Insurance Regulatory and Development Authority of India (IRDA) has been supplying mandates to various Indian Insurers to validate role of telematics in improving underwriting profits and decreasing loss ratios. In our current tested model, we have shown that accident frequency could be reduced upto 15% which in turn means lower loss ratios. With our Risk Assessment methodology, the motor insurance for the connected and autonomous vehicles could be revolutionized, however the role of regulators still stands crucial for such models to be implemented with the insurers.