June 27, 2016

Insurers and actuaries are used to handling large volumes of data and looking at patterns to establish risk. In fact this has been the basis of insurance going back to the origins of the industry. But it has become clear recently that with the rise of machine learning, big data and supercomputing – as well as consumer trends that are demanding instant gratification, fast quotes and fast claims settlement online – the industry is catapulting into a new age.

A recent white paper, looked at the transformative changes ahead for the insurance industry in terms of how it applies machine learning and artificial intelligence (AI). What are the specific potential applications? What will success look like for insurers and their customers in terms of accuracy, speed, sophistication, lower costs and lower premiums?

The story of insurers streamlining their claims process today is about applying big data and getting away from adjustor’s hand-written notes, fraud list data and claims management systems and being able to pull it all together.

It’s only 30 years ago that insurance companies worked with books, pricing manuals, written appraisals, fax machines and field reports from adjustors all written on paper. But for insurers now the key question is: am I getting the most from my data?

Consider also that a 1% improvement in loss ratio for a £1 billion insurer, with better use of data, is worth £10 million on the bottom line.

Analytics can flag up claims for closer inspection, priority handling or other action. With fraudulent insurance claims in the UK running at 350 cases every day, worth over £1.3 billion a year, there has never been a greater need to seek new claims solutions.

Predictive analysis and machine learning can identify fraud patterns effectively at every stage in the claims process, by a combination of modelling, rules, text-mining and database searches.  Analytics also has an important role to play in streamlining settlement.

Settling a claim at speed can traditionally be a costly process, with risk of overpaying, as any insurer settling a cluster of home damage claims for a particular area will know.

Analytics also has a role to play in managing the overall claims management activity. It makes sense to put more experienced adjustors on more complex claims, with better scores for assigning claims to the most appropriate adjustor and minimising the risk of litigation.

The authors of the Insurance Nexus white paper talked to AIG and Zurich about how they are going to use machine learning in the future.

In 2012 AIG launched its Science Team, looking at using data and modelling to identify business and education opportunities, introducing change management in its value chain. In 2015 Zurich launched the trial of its advanced, global predictive analytics initiative.

“I think there is tremendous potential for our industry to use machine learning to do things faster and smarter,” says George Argesanu, Global Head of Analytics, AIG.

“There’s not going to be a big bang followed by a new order of the universe, but slowly and surely, we are getting to a more accurate pricing of risk. Claims processes are becoming a lot more efficient, fraud will likely be caught more often and most importantly of all, more and more losses are and will continue to be prevented.”

Monika Schulze, Global Head of Marketing, Zurich, says that one of the most important touch points for this new world is in fraud mitigation.

“That’s where I see insurance applying machine learning, to improve the P&L,” she says. “Then claims management which is also very important. It is a much faster process and it is easier to reduce errors by using machine learning to process large amounts of data.”

George Argesanu and Monika Schulze are both speaking at the upcoming Insurance Nexus conference, the Insurance Analytics Europe Summit (5-6 October, London).

Follow the link to the LexisNexis Risk Solutions website to find out more about how we support insurance providers.