Cognitive computing and (re)insurance - the future?

Big data is endlessly debated in the (re)insurance world as a disruptor, though the industry has been relatively slow to get to grips with it.  

A major example of big data in action is cognitive computing, a relatively new concept in (re)insurance. In a nutshell it involves software that can process huge amounts of data very quickly, learn like a human brain and make appropriate suggestions like a human would.  

But now the concept is being explored by insurance firms such as Geico, the United Services Automobile Association (USAA) and, most recently, Swiss Re. All three are using the biggest cognitive computing engine, IBM’s Watson. 

So what are the ways that cognitive computing is having an impact on (re)insurance, and what is the future of this technology? 

The ways that (re)insurers have used cognitive computing so far fall into four main camps – client engagement, helping insurance agents, underwriting assistance and claims. 

Client engagement is the most common use of Watson, according to IBM Financial Services Europe Watson director David Robson.  

This involves Watson replacing call centre employees, he adds. 

One example of this is the United Services Automobile Association (USAA), the US military’s insurer and financial advisor. It uses Watson to answer online questions from its members about financial products. 

As part of this process, Watson was trained on more than 3,000 individual documents and with 2,000 individual questions. 

But even if an individual asks something Watson has not been directly trained to know, it has a chance of answering correctly anyway – even if the person asking the question is not confident of the subject matter, as many end consumers of financial products are. 

This is because it uses language processing software that lets it work out what someone is trying to ask. 

The second most common use of Watson is helping insurance agents, Robson goes on.  

“So Watson selling insurance, servicing insurance, Watson helping insurance agents do a better job by providing them with research, that sort of thing,” he explains.  

This is one area that Berkshire Hathaway’s Geico US insurer arm is interested in, according to chief executive Warren Buffett in a May 2015 CNBC interview. 

Buffett said that Geico, which started training Watson in 2014, is interested in using the software to help its motor insurance agents on the telephone, but that it could eventually respond to phone calls by itself. 

Watson could also help Geico with claims, he went on. 

 The third use for cognitive computing is to help underwriters directly.  

“What Watson can do in underwriting is read underwriting guidelines that a company has written, it can read a risk submission, be it life or P&C, and then having read the underwriting guidelines it can effectively help to underwrite that risk,” Robson says. 

The final use for Watson has been helping the claims process.  

“Watson can read the detail of the claim, compare it against the details in the policy and make recommendations about how somebody might settle that claim,” Robson explains.  

“The more information there is to read and the more research there is to be done, the more powerful Watson is. Because it’s a machine it can read a thousand times faster than humans.” 

North American financial services were the first to show an interest in Watson, Robson says, but he notes that their European brethren are not far behind, followed by other regions of the world.  

One recent European reinsurer to show a keen interest in the technology is Swiss Re. 

Swiss Re chief information officer Rainer Baumann says: “Swiss Re specifically decided to go into cognitive computing not because we have, today, short-term impacts that are so outrageous that we need to move there. We decided to move into cognitive computing because of the strategic long-term relevance.”  

Baumann adds that Swiss Re is using Watson on a mid-to-long term research project around life and health insurance, but that the technology has already helped Swiss Re keep its contract wording up to date. 

“In Swiss Re each contract is unique, and it’s very hard to find specific things in these contracts, for example when a legal or regulatory change happens, contracts are affected,” he says. “You can hire hundreds of people going through them, or you can have pretty advanced technology that lets you understand and interpret the contracts in a certain way.” 

In the future, Watson can also help the reinsurer with issues such as policy issuance and detecting fraudulent claims, Baumann adds. 

So what does the future hold for cognitive computing in (re)insurance? 

Baumann thinks that it could be a “game changer” for the industry.

“We well understand the potential of the technology and when you look at some of the examples that are around in other industries, you might become truly excited,” he says.  

 But the good news for smaller (re)insurance firms is that the benefits of cognitive computing will not just be reaped by large firms with deep pockets like the USAA, Geico and Swiss Re.  

“You can also imagine that as an industry leader we do not only focus on reinsurance, meaning that cognitive computing technology will not just enhance Swiss Re internally” Baumann says.  

The reinsurer has a history of sharing its research and development work with the wider industry, according to Baumann: “This is what makes Swiss Re distinctive and this is why we will not just focus on doing things for us, our focus is on helping the industry to advance.”