AI is set to improve (re)insurance inefficiencies, but humans will still need to be in the loop as the complexity of the industry requires human expertise, panellists said on the third day of the Singapore International Reinsurance Conference on Wednesday (Nov 6).
“If you look at the reinsurance industry, it’s a little different by line of business and geographically, but collectively less than 70 cents on the dollar in our industry goes towards claims,” said Anne Lohbeck, Chief Underwriting Officer, Specialty at Swiss Re. “We’re lucky as an industry that that number isn’t publicly front of mine for people, because it speaks to the level of inefficiency of the industry and how we originate risk, how we aggregate risk, how we transfer risk, and in how we administer claims.”
This level of inefficiency makes the reinsurance industry ripe for disruption, said Lohbeck. And while the price of risk has to be covered, it’s no longer a question of whether disruption will happen, but when, she added. “On all of those pieces that don’t go towards claims payment, we can become more efficient, and that’s something that’s really meaningful.”
Still, AI won’t replace humans any time soon, as human judgement still incorporates elements that AI can’t fully replace, she added. “I think now we’re at a point where we’re looking at what the human and AI can do in a very complementary way, and are focusing AI efforts on those elements that are classically considered tedious tasks by our teams, so that’s been a key focus.”
Lohbeck made the comments at a wide-ranging panel on AI that discussed how AI has been a game changer for many in the industry.
While humans should still be at the center of any strategy, AI is a strategic differentiator, said Sofia Kyriakopoulou, Group Chief Data and Analytics Officer at reinsurer SCOR.
“We know data is super central in reinsurance, and for SCOR, AI is probably a differentiator, a strategic differentiator,” she said. “We have in-house experts, and we have actuaries, very technical people, people who spend their days crunching data and developing analytics, and our job is really to augment them, to take some of their previous tasks away and to provide to them more data to enhance their decision-making.”
“With AI, for us, it’s a bet you have to pursue with high conviction,” Kyriakopoulou said.
Insurers are already moving to use AI to bolster their core capabilities, said Chen WenWen, Director and Lead Analyst at S&P Global Ratings. “(AI has) helped them enhance their competitive advantage, to compete with their close peers on product offerings,” Chen said.
AI is expected to transform specialty lines, said Lohbeck, as startups create models that can predict risk and handle claims. All this means that AI can get (re)insurers to a more efficient, effective level.
“Applying something like this on a larger scale, and allowing us to have better assumptions on a real-world projecting level, before it even comes to underwriting, has dramatic, dramatic potential,” she said.
Machine learning can also help with notoriously difficult to adjust claims, like business interruption. “Very similarly, you can use the same approaches that we’re using on the underwriting side to then becoming better in claims handling,” Lohbeck added.
Big Tech unlikely to take a piece of the complex pie
Despite the inefficiencies in (re)insurance, Big Tech is unlikely to try and get into the insurance world quickly. “There’s a lot of complexity in insurance,” said John Maroney, CEO at Global Asia Insurance Partnership.
Insurance is one of the few services and goods that people can’t get their full pricing on, Maroney added, as locking in insurance pricing for the next three to five years can be pretty difficult. Thus, Big Tech is more likely to pick up an embedded insurance approach or picking up a different distribution channel. “I’m not saying I don’t see it happening, but I think it’s probably more going to happen in isolated places,” Maroney said.
Ultimately, it’s a game of partnerships and frameworks, said Kyriakopoulou. “To profit from AI, you can only do it if you’re a new book player (that) can quickly integrate all these super powerful components,” she added. “So it is absolutely a game of integration and finding the right capability, being able to plug it in, monitor it, govern it, and solve the business problem in an incredible time to market fashion.”
Every single piece of data needs to be tagged, and data governance will need to be an essential part of the data foundation, Kyriakopoulou said. “We absolutely cannot do it without these partnerships,” she added.
AI is expected to create individual aspects of efficiency, said Lloyd’s chief Commercial Officer Dawn Miller on Tuesday (Nov 5), a move that is “fantastic for the individual business unit”.
Lloyd’s has “two dozen different AI proofs of concept going on in the business”, Miller added. “We’re seeing great advancements coming in through the lab,” she said.
“It’s a bit of a journey, and if we have maturity to realise there may be some negative dips in that journey with some of the unintended consequences of this coming up,” Miller added. “But we will come out the other side.”
