How can insurers use AI without letting it decide claims?
Only 8% of insurers have fully integrated AI projects.
Insurers are expanding their use of artificial intelligence (AI), but analysts warn the technology should be limited to processing claims and detecting issues rather than deciding whether claims should be paid.
“Within claim management in particular, there is a big difference between using AI to gather information, spot anomalies, or process straightforward queries, as opposed to asking it to decide whether a claim should or shouldn’t be paid,” Ross Sinclair, CEO at EIP Ltd., said in a Sept. 23 statement.
Sinclair said insurers should use insurer-configured rule engines to automate claim processing whilst retaining control over claim decisions.
A survey by S&P Global Ratings found 81% of insurers and reinsurers are prioritising automation, whilst 71% are using AI to improve internal support functions. Half are targeting customer service, and 39% are using AI for risk management, including fraud detection and pricing.
“AI is increasingly used to improve customer experience, underwriting and risk management, and claim processing, with operational benefits appearing earlier and more visibly than financial gains,” Andreas Lindberg, an analyst at S&P, said in a Sept. 21 report.
The survey of 121 insurers and reinsurers found 83% remain in the early or intermediate stages of AI adoption, whilst only 8% have fully integrated AI projects across their organisations.
Insurers expect AI to deliver efficiency gains of as much as 7% and revenue improvements of up to 5% by 2028.
The proportion expecting AI-related cost savings of more than 3% is projected to rise to 80% by 2028 from 16% in 2025.
S&P said AI has yet to materially affect insurers' profitability or credit ratings despite gains in handling claims, underwriting, and customer service.
More than three-quarters of respondents identified data privacy, security, and regulatory compliance as major barriers to wider AI adoption. About 62% maintain inventories of AI models to track their use and associated risks.
Questions to ponder:
- How much of the claim process can insurers automate without letting AI decide claims?
- Will projected cost savings materialise as insurers scale AI use?
- Can insurers expand AI use whilst meeting privacy, security, and regulatory requirements?
EXPERT OPINION
As insurers expand their use of AI, they should consider not only how AI can improve their own operations, but also how their systems can interact effectively with those used by other stakeholders—including insurance brokers, loss adjusters, and clients’ risk managers. Where appropriate, interoperable systems and agreed data standards could support the secure and efficient exchange of relevant information, reduce duplication, and enable more coordinated decision-making across the insurance lifecycle.
Alongside information security and privacy, data ownership and usage rights must be clearly defined. This includes establishing who owns the data, who may access and use it, for what purposes, and how long it may be retained.