Superior Image Validation reduced fraudulent claims

Brief

A Motor insurance company in the US was spending too much time on manual handling of insurance claims related to vehicle accidents. In addition, there was lack of consistency with respect to approval of the claim on the basis of damages during the accidents or otherwise.

Challenge

The company’s issue was that its competitor were able to do better in terms of claim evaluation as they were using automated system for review of images. Our client also suffer few cases of fraudulent where underserving claims were processed – causing financial losses.

Our Action

• Based on gathered data from historic claims, car images (damaged and non damaged) from multiple sources, we  trained the Deep learning model
• The model was adept at assessing damaged front, side and back for the vehicle. It could find the degree of severity of the damage for the vehicle to assess if vehicle is repairable or there is total loss

Results

• Improved consistency in claim processing
• Superior imaged validation reduced fraudulent claims
• Moreover, adjudication process expedited