Less than one in three (29%) UK insurers have a fully operational AI governance strategy in place, the Sapiens Insurance AI Maturity Index study, conducted by Research in Finance, has revealed.
The challenge is particularly acute among UK specialist and mid-size insurers where the figure falls to just 9%.
The study identified key competencies required for insurers to successfully adopt AI and found that 80% of those leading AI transformation, classified in the study as autonomous leaders, have fully operational policies.
“Strong governance must be a priority not an afterthought, and our research highlights a significant opportunity particularly for UK specialist and mid-size firms who are behind on this,” commented Paul Wheeler, CEO, Sapiens.
“These firms can be more agile than their larger counterparts and they should invest in doing vital governance groundwork now, so they are better positioned to take advantage of the technology and drive long-term business value over the next couple of years.”
Encouragingly for the UK, the study found that 81% of insurers using AI are confident in the decisions that AI and agentic AI are making, which is not far behind 86% of international respondents, but the figure is understandably higher for autonomous leaders (95%).
Elsewhere, UK respondents raised concerns about the guardrails for AI not being effective and the perennial shadow of compliance red tape also lurks. One in ten UK respondents said they are not confident in checks to monitor AI and agentic AI decision-making compared to only 2% of autonomous leaders.
When asked about the top reasons for overriding AI decisions, UK insurers had by far the biggest group citing compliance and policy requirements (62%) compared with 52% of autonomous leaders. This focus on oversight may be contributing to stronger monitoring capabilities in the UK, the study said, with 53% saying they detect misaligned AI and agentic AI decisioning either in near-real time or in less than two weeks
There are notable differences across the UK insurance market, with P&C insurers (54%) slightly ahead of life counterparts (48%) when it comes to spotting misalignments within two weeks.
Specialist and mid-sized insurers (65%) are most effective at this, in contrast to 46% of major insurance organisations. As most UK insurers are still working out their governance strategies, they could be erring on the side of caution when it comes to vigilance, the study suggested. Twenty-three per cent in the UK argued there will always be human involvement unlike the Nordics (9%) and Benelux (7%), suggesting human-in-the-loop is an important part of UK insurers’ AI governance strategies.