How can insurers turn AI spending into meaningful returns?

Many use AI in isolated parts of the business, limiting its financial impact.

Insurers are increasing artificial intelligence (AI) spending to reduce claim costs and improve underwriting, but many are still struggling to generate meaningful returns because of outdated systems and fragmented data.

In a March report, Boston Consulting Group, Inc. said AI investment amongst property and casualty insurers is expected to more than triple to 1.9% of revenue in 2026 from last year.

The consultancy said insurers that integrate AI more deeply across operations could cut costs by about 20% and increase gross written premiums by as much as 5%.

Insurers are already using AI in underwriting and claim processing to automate routine tasks, improve pricing accuracy, and detect fraud earlier. However, many do so in isolated parts of the business, limiting its financial impact, Boston Consulting said.

A separate April survey by A.M. Best Company, Inc. found that nearly 60% of insurers expect AI to reshape their business models between 2027 and 2029.

“Legacy systems can create significant barriers when implementing AI because they simply were not built for this type of data integration,” Kaitlin Piasecki, industry research analyst at A.M. Best, said in the report.

Sridhar Manyem, a senior director of Industry Research and Analytics at A.M. Best, said AI produces unreliable results when data is fragmented or poorly governed.

Despite the challenges, the appetite for the technology remains strong. A.M. Best found that two-thirds of insurers plan to increase AI investment from 2026 to 2028.

Among insurers already using AI tools, 63% reported modest productivity improvements, whilst 11% said they achieved measurable gains in output per employee.

Questions to ponder:

1. How quickly can insurers modernise legacy systems to support wider AI deployment?
2. Will insurers see stronger returns from AI in underwriting, claims, or fraud detection first?