AI Spotlight: How Predictive Analytics are Reshaping Post-Hurricane Insurance Claim Response

Share |

In the past two weeks, America has seen catastrophic devastation from both Hurricane Harvey and Hurricane Irma.  As a result, thousands of Americans are struggling with all the damage incurred to their property. As homes and cars are still submerged in floodwaters, insurance companies have to find ways to quickly and effectively resolve insurance claims to provide fast relief.  To do this, some companies have been deploying predictive analytics and machine learning tools to respond to insurance claims – and there have been plenty.  Esurance is one such company. 

Prior to Hurricane Harvey’s landfall in Texas, Esurance deployed a fleet of planes with cameras to take pictures in Texas.  After Hurricane Harvey hit, it did the same thing to provide complete “before and after” evidence.  These images are then fed into a computer and rapidly analyzed to detect the precise levels of damage and flooding to vehicles.  The program assesses vehicle year, make, model, and water depth.  Post-Harvey, claimants whose cars have been completely submerged automatically receive a full refund.  To provide fast-action relief, Esurance has been distributing electronic payments to customers whose cars have been destroyed.  Based on what other claimants file, the program is able to learn and improve the accuracy of its estimates of damage.  The benefit of responding this way removes the need for full on-site inspections to quickly settle claims that people desperately need to have resolved.

These disaster relief efforts by Esurance are not without a human component.  Aside from the obvious facts that humans are necessary to fly the planes, create the algorithms, and verify claims, Esurance also employed a disaster response team to Texas.  This team is able to settle incoming claims in person in conjunction with the software.  The machine learning tool acts to supplement human work rather than completely displacing it.  Having on site teams that are privy to data from the software also helps to quickly resolve claims in areas that are completely inaccessible.

SIIA’s issue brief on artificial intelligence (AI) and the future of work shows that in many cases, machine learning is a useful tool to supplement human work to increase efficiency and effectiveness.  Rather than displacing human work 100%, AI tools simply enhance human work.  Esurance’s machine learning software is a prime example of this.  Humans and machines working together are helping provide the better relief for victims of Hurricane Harvey.

The devastation inflicted by Hurricane Harvey will be felt for many more months.  The outpouring of aid and support to the region is immense and continuing.  Unfortunately, hurricane season is not yet over.  Machine learning and predictive analytics show promise for the future of disaster relief.  The use of these tools will hopefully become more prevalent to enhance aid for victims of hurricanes, tornadoes, fires, and other natural disasters.  In addition, the applications of machine learning in predicting the occurrence of natural disasters as well as responding to them will also improve so that companies and responders can provide the best support possible.

Diane Diane Pinto is the Public Policy Coordinator at SIIA. Follow the Policy team on Twitter @SIIAPolicy.