More than 90% of brokers have experimented with AI but only one-quarter have embedded it into their core operations, research from the National Insurance Brokers Association finds.

A report called Technology to Trust: AI, Technology and Broking – the sixth paper in NIBA's Sector Insights series – is based on a survey of brokers and brokerages.

It follows last year’s Ready or Reacting? Shaping the Future of the Broking Profession paper, which identified technology and automation as a key issue over the next decade.

The new study shows brokers are using AI to save time on document-heavy and administrative work, and keeping human judgment at the centre of advice.

“Brokers are experimenting widely with AI,” NIBA CEO Richard Klipin said. “The next step is turning that experimentation into planned, well-governed adoption that improves outcomes for clients.”

The study finds Microsoft Copilot is the most widely used AI tool across brokerages of all sizes.

Brokerages with a detailed and resourced technology strategy are almost four times as likely to have embedded AI than those without one (35% against 9%), and large brokerages are more than twice as likely as small and medium ones to have embedded the tech (39% versus 16%).

About 86% of brokerages are using or exploring AI for policy analysis and coverage comparison. Top uses all involve reviewing, comparing and summarising large volumes of information, with brokers keeping oversight and final responsibility for decisions.

About 58% of users save at least one hour a week through AI, and the most common use of that time is client advice and relationships.

The study shows trust is the biggest barrier, with 49% of brokers citing concerns about accuracy and reliability. Data privacy (43%), regulatory uncertainty (23%) and lack of internal skills or knowledge (23%) also feature.

Only one-third of brokerages using the tech have a formal AI policy, and more than 20% have no AI-specific safeguards. More than one-third of brokerages cannot identify who is responsible for AI in their business.

Human review of AI outputs is the most common safeguard among small and medium brokerages (49%). Among large businesses, staff training on responsible AI use is the most common (62%).

About 41% of survey respondents asked for clearer guidance on responsible AI use.

“Technology will change how broking is delivered, and trust will remain the reason clients choose a broker,” Mr Klipin said.

“NIBA will use these findings to support members with practical guidance and education on responsible AI use, and to advocate for a regulatory environment that supports innovation while protecting clients.”