KPMG has warned insurers to “pick up the pace” in adopting artificial intelligence.

Despite wide acceptance that AI could transform operating models, the industry approach is “cautious optimism”, it says, and projects are taking too long to reach production.

“The slow pace of implementation is creating significant delays in progress compared to other industries.

“Despite the use of AI in a handful of areas and pilots ... insurance companies are not gaining an advantage on competitors by using this technology more widely. Doing so could enable businesses to work faster, more flexibly and develop more sophisticated models.”

A survey by the consulting group found 58% of insurance CEOs said it would take three to five years for generative AI to provide a return on investment, and there is “still hesitancy around wider deployment, exacerbated by challenges around the speed at which AI is evolving, data quality, bias and regulatory compliance”.

Many insurers are using AI in actuarial models and fraud prevention, and in voice recognition, chatbots and “next best action” prompts for agents.

But life insurer Prudential’s CTO Anette Bronder says many AI activities are triggered by IT departments.

“We need to flip this around and create an appetite within business functions ... Being able to respond quickly to business needs will be critical; if this doesn’t change, insurance organisations will likely be too slow,” she said in KPMG’s report.

PassportCard CEO Alon Ketzef says insurers are typically slow adopters “for the right reasons sometimes”.  

He recommends established insurers create a separate cloud and AI business.

“When it gains momentum, I would move pieces of the old business into the new business. I am afraid that any other approach – with incremental improvements here and there – is not going to move the needle at the speed needed.”

AI can transform insurance claims, but KPMG says “doing so appears to be slow, with testing and implementation processes often taking several months to complete”.  

“Wider [AI] adoption among insurers has been relatively slow and siloed to date, and there are growing concerns around the accurate output of the technology ... Senior leadership teams will need to quickly identify and develop their AI strategy, carefully navigating the balance between embracing innovation, understanding their barriers to adoption and mitigating the emerging risks.”

It recommends an internal “maturity assessment” for clarity on current capabilities and to identify areas to prioritise.