The benefits of AI in insurance are coming gradually and the technology has not yet “transformed the industry”, new research finds.

Generative AI is boosting productivity but adoption is “uneven across the industry”, a report from insurance analytics group SAS and researcher Economist Impact says. Companies relying on legacy systems are moving more slowly than insurtechs and cyber specialists. 

“Many insurers are still tailoring AI to specific use cases rather than deploying it at scale,” the report says. 

SAS insurance adviser Thorsten Hein says it is “not yet the case” that AI agents can perform all kinds of complex tasks. Their greatest value is in repetitive and data-intensive jobs. 

“Some elaborate tasks, such as calculating complex risk models, still require integration with specialised industry solutions and human expertise,” he said. 

Tech-led change may accelerate as recent advances in generative AI speed coding work, streamline customer service operations and enable real-time analysis of cyber and climate risks. AI has long played a role in underwriting and fraud detection. 

Insurance leaders envision an agentic AI hybrid workforce in which AI agents and humans work together on tasks such as underwriting, product development and claims processing, the report says. 

Experts from Zurich, HDI Global, Tokio Marine and Manulife were consulted for the paper and revealed a sector cautiously optimistic about AI’s potential. 

“Insurers’ workforces will become hybrids of human employees and agents collaborating closely, with some agents working largely independently under human oversight,” Manulife global chief analytics officer Jodie Wallis said. 

The research involved a survey of more than 500 executives from 17 countries on the future of insurance. 

It found wide variance in regulatory requirements across nations is a challenge for global companies.