I had a blast at the SAS Financial Services Executive Summit. Representatives from banks and insurance companies came together at SAS headquarters in Cary, North Carolina to talk about the future of the financial services industry. The agenda included many interactive sessions and lots of brainstorming with fellow attendees while mapping out ideas on whiteboards. During one of these exercises attendees were asked to pretend that we could fix any issues in our data with the push of a button. One of the attendees suggested that we could program a button to “automatically find fraud.” Wouldn’t that be nice?
Insurance fraud detection is challenging. While fraud is definitely increasing (the latest ForeCAST NICB quarterly Questionable Claim Referral report suggests a 23 percent increase over two years) it still remains a relatively rare event when compared to the population of legitimate insurance claims. Rare events like insurance fraud are difficult to predict. So, naturally it would be nice to push a single button and have all those problems fixed immediately. We’re not quite there yet.
The next part of our hypothetical exercise involved figuring out ways we could fix at least some of the challenges and move towards this goal of push-button fraud detection. What we discovered is that most insurance companies suffer data challenges related to fragmented sources, information silos, poor data quality and a lack of discrete elements to draw from since the most useful information is often captured in unstructured text fields like claim notes. Many fraud detection solutions today focus on trying to detect rare events but they fail to address these corresponding data issues making it difficult to be successful.
What we all quickly realized is that one of the most important aspects of insurance fraud detection is addressing data quality and integration issues before trying to detect anything. When asked how we might be able to do that today, one attendee responded simply, “SAS.”
I couldn’t agree more. Solutions like the SAS Fraud Framework for Insurance include data quality, integration and text analytics technology to address these very issues up front. The famous inventor and science fiction author Arthur C. Clarke said that, “Any sufficiently advanced technology is indistinguishable from magic.” In this case, the magic comes when companies are able to fix their data problems first and then apply advanced analytics in a hybrid approach to detect those rare events.
James Ruotolo is the Principal for Insurance Fraud Solutions at SAS. Connect with him on Twitter or LinkedIn.
