An Update on Predictive Modeling in the Insurance IndustryBy Louis Riggio | March 10, 2017
In an interview with Claims Journal, Dan Donovan, assistant vice president of ISO Claims Solutions, Verisk – insurance solutions, spoke about changes and opportunities in predictive modeling to combat insurance fraud.
Using advanced analytics to fight fraud is reaching a stage of “maturity” in our industry, Donovan says. Years ago, predictive modeling was limited to larger insurers that could afford custom solutions. Today, the cost of the physical infrastructure to support data science has decreased. Changing technology, especially cloud computing, allows for a lower point of entry to host data, and “carriers have a better understanding of the data they have and how to take action with that data.” Donovan contends, “Carriers are seeing the benefits of analytics and getting a better feel for how to integrate them into their daily workflows and operations.” In addition, companies have developed a more thorough understanding of ways they can improve their models by supplementing their data with third-party information.
Tapping into Big Data
Still, Donovan says, “I don’t think insurance carriers have really tapped into big data and what it could be in the future.” He explains that “based on our industry interactions, we’re seeing a gap between the aspirations of senior leadership and the ability of organizations to effectively operationalize analytic solutions” in the claims departments—and elsewhere. According to Donovan, the promise of predictive modeling will be realized when the industry adopts affordable, scalable software solutions that are accessible “regardless of the size and expertise of an insurer’s internal IT resources, data science team, and SIU.”
Donovan’s vision spans all lines of business, because “any type of loss is open to predictive modeling and fraud scoring.” The most common antifraud applications are for auto and physical damage claims, however. In that space, competition to pay claims more quickly means insurers have a greater need to flag nonmeritorious claims in the pipeline.
Telematics Is a Game Changer
With so much insurance fraud being opportunistic, data from telematics systems and other elements of the Internet of Things (IoT) could forever change the way insurers evaluate claims. “IoT will help them be more objective in their evaluation of losses,” Donovan notes, for example, in cases where an accident occurs but crash data shows that the claim involves exaggerated injuries or inflated medical treatment.
Longer term, Donovan predicts that insurers’ access to real-time accident data through services like the Verisk Data Exchange™ will improve the industry bottom line in many ways: “The deterrent effect may ultimately be the biggest thing that IoT and telematics does from a fraud standpoint. Telematics can help ensure that only meritorious claims get paid—and gets them paid more quickly.”