Insurance claim fraud detection

Identify which claims that are submitted to a fraud investigation team are actually fraudulent.

3

Consecutive record months of operational performance resulting in high levels of staff morale

Results

200%

Of the original fraud collected with just
95% of the cost of investigations

Benefit

This solution reduces wasted effort and cost investigating claims that are not in fact fraudulent. This also has the additional impact of enhancing staff morale and satisfaction from the product houses.

Context

Emerge’s client has a team that investigates insurance claims to determine if they are fraudulent. They complete this function for multiple product houses which each have their own nuances and complexities. There is limited capacity in the team and they were looking for a system to prioritise their investigations and minimise wasted effort.

Methodology

Using all the data available at the time of claim including details about the policy, the policyholder, and the claim, Emerge built an AI solution to prioritise which claims are most likely to be fraud and which are almost certainly not fraud. The team prioritised the claims that are likely to be fraud, and invested more effort to find the fraud. The claims that the AI solution identified as not being fraud were automatically approved without protracted investigation

  • Emerge ML Twitter
  • Emerge ML LinkedIn

© 2020 by Emerge