False insurance claims – which account for as much as 10% of all European payouts – have a negative effect for genuine consumers. But recent advances in artificial intelligence are delivering promising results for customers.
In what is believed to be an industry first, data experts at Dutch insurer Aegon have developed a computer algorithm capable of quickly identifying suspicious claims. The tool, already tested at several of Aegon's divisions in the Netherlands, complements traditional methods, making them faster, more efficient, and more effective at paying out genuine claims.
"It's a win-win for both Aegon and our customers," says Olaf van Haver, Senior Data Analyst at Aegon's Center of Excellence for Digital. "Genuine claims will be identified and paid out directly, and Aegon's exposure to fraud is reduced."
So, how does it work and what are the 'red flags' Aegon is looking for? "With our smart algorithm we look for anomalies in data sources," he explains.
The goal, says Olaf, is to detect "strange behavior among the entire volume of claims we receive, such as previously rejected claims, claims made immediately after a policy was purchased, or claims made from addresses which don't exist." Every claim is given a 0 and 100 score based on an 'anomaly index'. The most irregular are shared with claims handlers for follow-up.
So, instead of profiling individual customers, the algorithm checks claim history, exchanges information with competitors, and cooperates with law enforcement agencies.
Identifying false claims is a fine art. That's why Aegon continues to rely on its investigations team to analyze cases highlighted by artificial intelligence, and provide regular feedback on the algorithm's accuracy.
Speeding up payment
"The information from our fraud-detection system is important to help our team prioritize work, and ensure that genuine claims are paid without delay," says van Haver. "We get daily feedback on which cases identified by the algorithm are worthy of further investigation." Aegon's digital experts can continually fine-tune the algorithm, so that it gets more intelligent. The team also looks for new data sources both within and outside of the company, to verify the accuracy of a claim.
Van Haver says that human intervention remains crucial to identify which claims are intentionally false, exaggerated, or the result of an honest mistake by the customer.
Spotting accidental mistakes
"Some claims can look strange to the algorithm and indeed they are strange when we look at them. But after checking with the customer and examining the evidence, it becomes clear that the claim is not deliberately fraudulent," van Haver points out.
And when there are 'gray areas', Aegon will get in touch with the customer and work out a solution. "Our aim is always to serve our customers who have a valid claim," says van Haver.
Following the trial of artificial intelligence, Aegon intends to introduce the method to other product lines and other Aegon businesses around the world.
"Our system is an innovative tool in the insurance industry: we have already achieved a lot, and there are many things we can still do to develop it further," says van Haver.
"Rolling out the use of artificial intelligence brings challenges in terms of systems integration and data availability. But, based on our experience in the Netherlands, we are sure we can apply the technology to more of our products other markets we operate in."
Aegon, and its customers, will see benefits as artificial intelligence continues to help win the battle against false claims. "If we lower the amount that we incorrectly pay for false or exaggerated claims, that helps everybody," says van Haver.