KPMG used an innovative machine learning approach to show how artificial intelligence could help give policyholders an answer in a matter of minutes.
For people diagnosed with a critical illness, knowing that an insurance policy will pay out can make dealing with their diagnosis a little bit easier. But across the insurance industry, the number of weeks policyholders can wait to get their claim assessed is excessive.
Our client wanted to understand the possibility of giving policyholders an answer faster. We used an innovative machine learning approach to demonstrate how artificial intelligence could help give policyholders an answer in a matter of minutes.
Our analysis found it was possible to accurately identify valid claims by understanding the actual information claim assessors use to make a determination. We introduced the machines to the input from which thousands of decision trees were produced by the machine assessors. The machines were able to assess the important features driving decision-making.
Scroll down to continue reading...
This, in turn, allowed our client to understand that decisions could be made by answering fewer questions at first notification of loss. Based on the answers, we were able to automate the decision-making process, with machines replicating the reasoning applied by humans. The machines were quickly able to learn how to make the same assessment as a human claim advisor in a significant percentage of cases – and they could do this within seconds.
Our work on this important prototype highlights the potential of artificial intelligence to transform the quality of service policyholders can expect from their insurers in the future. Capitalising on artificial intelligence to accelerate claims’ assessment also holds real promise as a differentiator for insurers striving to maintain a competitive edge in a market that is already feeling the impact of digital disruption.