Take AI seriously if you want to really transform underwriting and claims: DGTAL CEO

Artificial intelligence in reinsurance has reached the stage where we should be past the talk and the tests. The technology is already reshaping underwriting, claims and capital allocation, but only for those taking it seriously.

Key points:
Production, not pilots, matters
Combined ratios improve with AI
Investors expect faster insights

That is the view of Arndt Gossmann, chief executive officer and co-founder of DGTAL, who believes reinsurers must move past proofs of concept and pilots to real-world implementation. “Take it seriously – that is my message,” he emphasised. 

“The use of AI in production will help CEOs, CUOs and engineering officers to quickly and significantly improve the KPIs of their business.”

For Gossmann, the immediate opportunities are clear: underwriting and claims. “These are the two key areas of interest. In underwriting, AI enables reinsurers to request more data, analyse it better and build much deeper analytics. 

“That will enable improved pricing decisions, especially in a market that is no longer so hard, and where competition is increasing.”

On the claims side, he sees equal potential: “By analysing the claims that come in, and by going broader and deeper, AI gives much better predictions for the future. The interesting part is the feedback loop between claims and underwriting. 

“If companies had analysed claims development more quickly during the Covid pandemic, they would have been able to anticipate this in pricing within the next year.” 

“As long as you don’t have systems in production, you cannot judge whether it’s working.”

The impact is tangible. Wherever AI is systematically put into production, Gossman explained that “those who do it see significant improvement in their combined ratios. That means moving towards 10%, which should be enough motivation to look into it”.

Yet the industry is still mired in pilots and setbacks. Gossmann cited a recent MIT study: “95% of companies are so far disappointed with the performance of AI. Only 80% have looked at it. Out of these, just 50% have gone into a pilot, and only 40% of those into production.

“As long as you don’t have systems in production, you cannot judge whether it’s working.”

His tip is to choose technology providers already running in production. “There are lots of offerings. However, there are a few but still a couple of players – not only us – where you can see things are working. 

“Concentrate on providers with clients in production, not just proof-of-concepts.”

Unlocking claims data is another hurdle. Most remains unstructured. “The existence of a data warehouse doesn’t mean data is available in an analysable form,” Gossmann warned. “If you have a good data warehouse and good AI, however, you can cut through. AI makes huge amounts of data available and DGTAL deliver the front end for this.” 

The ability to query an entire portfolio in real time, rather than waiting weeks for an analytics team, would mark a fundamental shift. The business case is straightforward: efficiency and margins. 

“If you want to improve your combined ratio, 10% is realistic to achieve within a year. We’ve seen clients reach that level. If you cut the claims ratio by 10%, that’s a reason to get up in the morning.”

But there are barriers. Reluctance to adopt new tools, poor past experiences, lack of leadership support and weak change management all slow progress. “A head of AI is not leadership support,” he stated.

“The C-suite has to understand they need to look into this. I’m not saying they should put AI everywhere, but they should use it where it makes sense.”

Looking ahead, Gossmann predicts AI will move from back-office efficiency to the fundamentals of pricing and capital allocation. “It will change how risk is priced and how capital is allocated because investors will expect different ways of pricing.” 

He pointed to claims status updates after catastrophes: “You can give a status within one or two weeks. Investors sometimes have to wait six to eight weeks for the first idea. Once they know it’s possible, they will ask for it daily.”

His final advice to CEOs at Monte Carlo is uncompromising. “If you take it seriously, you can see obvious improvement in the performance of your business in an unprecedented way. 

“Yes, it’s tough work to find the right providers, the right technology and to organise and implement it. But it will be rewarding for the C-suite and shareholders. My message actually is – have some fun with it.”

Arndt Gossmann is the chief executive officer and co-founder of DGTAL. He can be reached at: a.gossmann@dgtal.io.

For more news from Monte Carlo Today, click here.

Did you get value from this story? Sign up to our free daily newsletters and get stories like this sent straight to your inbox.