Brokers can use artificial intelligence to slash the time spent processing submissions, triaging risks and preparing placements, but what happens when the rest of the insurance chain does not move as quickly?
That question became a theme during a Dive In Festival panel on AI and digital transformation, where insurance and technology leaders examined how automation is changing distribution, broker-carrier relationships and the skills the industry will need as more routine work disappears.
A broker may become dramatically more efficient internally, but those gains mean little to the client if “quotes still take days to return,” according to David King, co-founder and co-CEO of Artificial.
“If a broker can process things in maybe 20% of the time it takes now, but they still have to walk into the Lloyd’s building or send an email and wait for 20 quotes to come back, and 18 of them take a week, the whole process still takes a week,” King said. “The customer experience is still a week.”
The issue seems to cut into one of the biggest challenges facing brokers as AI moves deeper into insurance distribution. King said the market already has technology addressing individual tasks including ingestion, placement and appetite codification; the bigger hurdle is connecting those tools across brokers, insurers and the wider insurance ecosystem.
Ed Short, head of digital at Arch Insurance International, said technology increasingly requires technical specialists at carriers to work directly with their counterparts inside brokerages rather than routing every interaction through traditional placement teams.
According to Short, every broker Arch works with now has at least some digital interaction with the insurer, although the extent varies considerably.
In some markets, transactions begin digitally and become manual when risks grow more complicated. London and wholesale placements often start from the opposite position: highly complex, relationship-driven transactions where firms are identifying pieces of the process that can be digitised. This makes collaboration between brokers and carriers increasingly important if investments on either side are to translate into faster placements.
But technology is unlikely to diminish the value of the relationship itself. Martin Reith, chairman of Artificial Labs and chief executive of the Corinthian Fund, said AI could take over more of the transactional work while leaving brokers with greater capacity for negotiation and client conversations.
“The low-value transactional piece of an insurance risk will be efficiently dealt with, allowing for high-value conversation and negotiation to take place,” he said, adding AI could ultimately strengthen brokers’ trading relationships by stripping out processing work and allowing more time to be spent on complex risks.
Faster placement also raises a more fundamental question: which decisions should remain with brokers and underwriters?
For Reith, AI should provide information while people retain accountability for the ultimate decision, particularly where the broker’s experience and understanding of the client influence an underwriter’s view of a risk. “If I trust that broker and think they have good experience and good knowledge, that all helps to form a good outcome for my client,” he said.
King said explainability should be a hard boundary for automated insurance decisions.
“If you haven’t got the accountability, the audit trail and a reference to a deterministic process where the decisions have been made, then it shouldn’t be allowed,” he said.
Bias becomes particularly significant as those systems scale. Elizabeth Woolaston, chief of markets at Artificial, believes firms need to be able to audit AI-supported decisions and identify where unwanted bias may have influenced an outcome.
“I think it would be untrue of us to say that we can get rid of any bias within our AI,” she said. “But what we are able to do, or what organisations are able to do, is review the decisions that have been made and make them better and less biased for the future.”
