AI recommendations increase mis-selling risks for insurers

Thursday, 17 September 2026 15:22 - - {{hitsCtrl.values.hits}}

AI could help insurers identify protection gaps across Asia, but automated recommendations increase the need for stronger conduct controls

Insurers are looking to artificial intelligence to reshape how customers select coverage, moving from product-led sales towards recommendations based on individual risks and protection needs.

Vanessa Lou, Asean Insurance Leader at EY, said AI could combine information on customers’ circumstances, existing protection, and relevant risks to identify coverage gaps that they may not recognise themselves.

Arthur Calipo, Asia Pacific Insurance Leader at Deloitte, sees particular potential in Asia, where 700 million to 750 million customers have entered the financial system over the past decade. As incomes rise, he expects more consumers to move from deposits and savings towards protection and wealth products.

“AI can actually model customers’ actual risks and show them gaps in terms they recognise and understand,” Calipo said.

The technology could also support more personalised engagement and conversational sales, but a greater role for algorithms in recommending products creates a conduct challenge. Customers may increasingly rely on automated recommendations when deciding what protection to buy.

“The fact that there is an algorithm that makes the recommendation doesn’t change who is accountable for it,” Lou said.

Lou identified explainability, governance, continuous testing, and clear ownership as critical safeguards. AI recommendations should remain within insurers’ existing risk and compliance frameworks, whilst models require ongoing monitoring as customer behaviour, data, and risks change.

Calipo said governance should develop alongside experimentation and deployment rather than operate as a separate workstream. He also highlighted the sensitivity of personal data held by life insurers, increasing the importance of controls as automated decision-making expands.

For insurers, the opportunity is to use AI to identify unmet protection needs and make complex products easier to navigate. The execution risk is allowing personalisation to advance faster than the controls governing how products are recommended and who remains responsible for customer outcomes.