Poor data quality is the biggest barrier to AI in insurance

Almost three-quarters of insurance underwriters say fragmented, siloed, and unstructured data -- not technology -- is the main barrier to AI transformation.

New research carried out by Reuters for technology transformation specialist CI&T shows that data fragmentation, unstructured formats, and siloed systems are the real roadblocks to delivering faster, more accurate underwriting and pricing.

The biggest barrier to optimizing AI and GenAI is the difficulty in extracting and using unstructured data, cited by 54 percent of respondents. A further 24 percent pointed to data being spread across disparate sources within their organization.

In response insurers are adopting Generative AI cautiously, testing in sandbox environments to mitigate risks such as hallucinations, bias, and data privacy breaches.

Just 15 percent of claims leaders believe greater personalization will significantly improve customer satisfaction, compared with 41 percent prioritizing streamlined internal processes and 39 percent favoring a blend of digital and human touchpoints.

“AI’s success in insurance won’t be determined by how advanced the algorithms are, but by the quality and accessibility of the data that feeds them,” says Mike Young, VP insurance industry growth at CI&T. “This research shows UK insurers are ready to innovate, but they need to get their data house in order first.”

The report shows that 55 percent of claims professionals surveyed believe AI-driven data entry and case assignment will have the greatest impact on claims efficiency. Additionally, 60 percent see technology as the key to offsetting rising claim costs and premiums.

You can get the full report from the Reuters site.

Image credit: Khakimullin/depositphotos.com