Skip to main content Skip to footer
  • "com.cts.aem.core.models.NavigationItem@5894af9b" Careers
  • "com.cts.aem.core.models.NavigationItem@4ca648af" News
  • "com.cts.aem.core.models.NavigationItem@36d62b9f" Events
  • "com.cts.aem.core.models.NavigationItem@7f4bbd89" Investors
Case study

The challenge

In the insurance business, assessing risk is not an exact science. Our client, a major global reinsurance company, wanted to examine the risk of reinsuring specific tranches of risk for its clients.

While flood insurance represents a significant opportunity for insurers, framing each risk must be balanced by accurate pricing. To get a clearer view of flood risks and the ability to model risk factors by geography down to individual ZIP codes, the client turned to Cognizant for a solution.

Our approach

Cognizant helped the reinsurer develop an intelligent algorithmic process to aid the underwriting process and boost efficiency. This involved analyzing flood hazard maps developed by the National Flood Insurance Program, publicly available census data and housing information. The solution overlays the geospatial data with data from geographic information systems (GIS) and our client’s internal data on historical claims. Our AI and Analytics team used R software with ArcGIS for geospatial data extraction. We identified potential attributes that affect market opportunity and risks for areas down to individual ZIP+4 codes. We then depicted these risks using a dashboard with visualizations built on the RShiny platform.

AI delivers intelligent underwriting

Using natural language processing to automatically examine digitized documents and combining that information with geospatial data on flooding, our client can now accurately assess the frequency and severity of flood risk by ZIP code. 

Our AI-driven machine learning solution leverages subject matter expertise with data science to make predictive underwriting faster, more efficient and more accurate. The client can now assess coverage and model the factors that could drive the market, including behavioral patterns.

83%

accuracy in modeling potential market

10-fold reduction

in underwriting throughput time

25% improvement

in case acceptance