Fraud and Abuse Detection

It is widely believed that most insurance fraud goes undetected. Traditional rule-based fraud detection processes can, by definition, only uncover fraudulent practices that have been used in the past. It would be blind to new and unknown types of fraud which could make up a significant amount of the fraudulent activity.

Claim Analytics has developed a unique solution to this problem, our solution can be used in conjunction with existing fraud programs and will identify new fraudulent activity. Our solution does not require any prior knowledge as to which claims in the past were fraudulent. Instead, our models are trained to identify atypical practices to flag for further investigation.

Application Validation

Applications to buy insurance should be reviewed to protect against fraud, under-disclosure and simple errors. Many companies focus on application fraud using activity based requests, while overlooking other potential dangers that can arise from simple errors. Our approach is to flag anomalous activities by identifying patterns specific to your data and your business. This can build on existing safeguards, and it has the advantages that it does not rely on previously identified suspicious activity and can be automated in the application process.

Provider Fraud

Our provider fraud models have effectively identified fraudulent provider activities that were not known in the past as they are trained to identify service providers with atypical treatment practices. This approach will result in more false positives – think of a dentist that specializes in patients with dental phobia. However, by identifying those providers whose practices differ from their peer group this modeling approach can be very successful at uncovering fraudulent and abusive practitioners and uncovering unknown and novel schemes.

Claim Fraud

With effective application and provider fraud programs in place, claim fraud largely relies on individual investigation. We can help flag suspicious claims by using a predictive model trained to identify atypical claim activity.


What our clients say..

“I was joking with a colleague the other day that the reports are like those drug-sniffing dogs at the airport. They sniff out the files that need further investigation.”