Never has so much data been available to assess insurance risk from so many sources. Building for and maintaining flexibility and modularity in architecture will be key elements to success.
Historically, insurance underwriting and claims data have been manually or flat file captured via an agent, broker or directly with a carrier. Advancements have been made in pulling supplemental data from credit scoring bureaus on the consumer side and sometimes proprietary databases relating to claims and other company-specific information for commercial insurers. However, the status quo has been around for twenty years or more, leading to both internal and third party systems design that is rather rigid around a traditional underwriting and claims handling model.
With connectivity rapidly expanding and the explosion of data sources providing new and insightful information and analytics, traditional measures of underwriting risk are quickly evolving. The same is true for claims settlement rules and procedures.
For example, smartphones can provide an enormous amount of behavioral information about an activity and an insured. In the short span of a few years, data science around driving behavior has evolved dramatically both in terms of collection and data analysis. OBD-II devices installed in a car had been the standard in the industry to collect data until only a few years ago. This method proved expensive, and while costs have come down, the market has been shifting to getting similar information from a smartphone’s accelerometer and gyroscope to assess driving data. Even more recently, insurers have been exploring using other smartphone data, such as how a device is used in a car, to enhance predictive analytics regarding driving behavior.
"With connectivity rapidly expanding and the explosion of data sources providing new and insightful information and analytics, traditional measures of underwriting risk are quickly evolving"
Similarly, sensor data built into a vest worn by a worker can predict future workers’ compensation claims relating to repetitive injury. In hurricane or earthquake zones, drones can be used to assess damage when other modes of travel are difficult or impossible, improving response times, and potentially greater claims accuracy.
Moreover, with telecommunications costs dropping dramatically and the number of data sources providing updated information increasing dramatically, risks can be more real time assessed, thus creating a catalyst for a long sought after positive change in the insurer’s relationship with an insured. Carriers would have multiple touch points that provide useful insights not only at renewal or when a claim occurs. Understanding customer behaviors is strategic to predict future behaviors, which facilitate the offering of relevant products and services.
While extremely promising and exciting, the future of insurance analytics provides unique challenges to an insurance technology and information executive. Reinventing technology infrastructure and data architecture to use information in an intelligent way requires a startup mentality and an agile development approach. More and more data sources will become available, and future insights will require adaptability in ways that insurance companies can now only imagine.