Insurers leverage AI to unlock legacy claims data

Register now

The insurance industry has seen a great amount of digital transformation in recent years, but two areas that have not reaped the benefits yet are policy development and claims handling. However, with advancements in artificial intelligence, this is set to change.

That's according to Nathan Root, VP of insights and data for insurance at consultancy Capgemini Financial Services. Root, formerly an SVP in charge of underwriting transformation at CNA, says that AI's impact on claims can drastically reduce cycle time.

Insurers are using AI technology to analyze their old paper claims, submission applications, and policies. These digitized documents can then be subject to analytics that can inform carriers' policy and claim strategies going forward.

“Within the claims process, insurance companies are looking to deliver a better outcome and an overall better customer service experience,” Root says.

For example, he says, AI is allowing companies to dramatically cut claim routing times. Companies he is working with are using AI and data analytics to create a system where a new claim can automatically be routed to an adjuster depending on the severity of the claim and the experience of the adjuster. This cuts the process of assigning a claim from days to just minutes.

In order to deliver this improved experience, Capgemini works with clients to combine the growing abilities of AI software to understand natural language with existing technology, such as character recognition, to digitize the data in their new claims.

Root described the maturity of the big data and analytics capabilities when it comes to mining old claims as “mainstream" -- but it hasn't yet peaked, he says. For example, there's still a lot of potential for insurance companies to use AI to automate data collection and analysis.

“It has not yet hit its threshold of maturity,” he said. “In five years, it will be more mainstream and more mature. It will get more innovative with machine learning.”

For reprint and licensing requests for this article, click here.