Although a minority of companies today possess the advanced technology needed to leverage predictive analytics in underwriting, this would appear to be changing. A new report from Capgemini, Embracing the Power of Predictive Analytics: Are your underwriters ready for change? provides a look at the outlook for predictive models and implementation challenges, before providing recommendations.

The report states that predictive analytics has emerged as a critical tool for property and casualty (P&C) insurance carriers but only 27 per cent of those insurers studied by Capgemini in the World Property and Casualty Insurance Report 2024 possessed the advanced technology needed to leverage predictive analytics in underwriting. Among the hurdles companies must surmount, they say insurers must learn to manage clean, high quality data sources and substantially more data than they’ve had access to in the past.

Rising use of third-party data sources 

“An increased use of third-party data sources to obtain this data is unavoidable, so carriers need to assure the data is diverse, accurate and comprehensive,” they state in the predictive analytics report. “Stay current on the rapidly changing rules in this new frontier and develop an effective artificial intelligence (AI) governance framework.”

The recommendations made in the paper include involving underwriters early in the development of predictive analytics models. It recommends companies support underwriters with only the data they need to make the decisions at hand and develop and implement technologies that will help an underwriter quickly triage and prioritize work. The report also recommends conducting regular meetings with the analytics and the underwriting teams together. “Review the data and uncover how it is affecting underwriting workflows,” they state.

“Better data enables carriers to more effectively price risk, maintain favourable loss ratios and generate new revenue streams by being able to accept risks that would’ve otherwise been denied previously,” they conclude.