Insurance technology company mea Platform has commissioned new research from ISG, the global technology research and advisory firm, examining how insurers are approaching AI across underwriting, claims and operational workflows.
The study finds that insurers are broadly open to using AI for repeatable work, but most want the technology to operate through insurance-specific models or within controls set by their own organisations.
The research found that 83% of the global insurance market would support AI executing repeatable work, while 75% said they would only allow it where an insurance-specific or governed model is used. At present, fewer than 1% of the market operates on a fully AI-native basis, according to ISG’s research.
The study draws on responses from senior insurance leaders across underwriting, operations, claims, technology and transformation. It examines 20 operational activities, ranging from submission intake and triage to quote generation, bordereaux processing, claims adjudication and compliance screening, with respondents across North America, Europe and Asia.
The findings point to a distinction insurers are making between routine operational processes and decisions requiring human judgement. Some 86% of respondents said consequential decisions, which the research describes as those that create genuine company distinction, should remain with people.
The research commissioned by mea Platform suggests insurers instead see AI playing a larger role in the work surrounding those decisions. By handling repeated processes and organising information, AI could give underwriters and brokers more complete evidence when assessing a risk, while reducing the time required to reach a decision.
Operational capacity is also emerging as a factor in how much business insurers can process. According to the research, carriers estimate that one in nine broker submissions is declined or left unquoted because their operations teams cannot keep pace, despite the insurer having an appetite to write the risk.
Among insurers already using AI in operations, 61% report productivity improvements and 51% report faster cycle times. Respondents also expect operating costs to fall by 16% over the next two years.
The research indicates that insurers are considering AI in terms of capacity as well as efficiency. When asked what would most improve brokers’ perception of their businesses, 64% selected pricing, 52% ease of doing business and 51% AI-driven speed and completeness of submission responses.
The findings also show a clear preference for controlled AI systems when decisions carry greater consequences. For high-consequence underwriting and claims decisions made with limited oversight, 75% of respondents said they would trust an insurance-specific model or a governed hybrid. Only 6% said they would accept a general-purpose model operating independently.
ISG’s research, commissioned by mea Platform, uses “governed” to describe systems that operate against controlled policy wording, endorsements, underwriting appetite and claims guidance. Such systems can have permissions that distinguish between what they are allowed to read and what they are permitted to decide. A named person can override or stop the system, while the technology operates according to the organisation’s own appetite, thresholds and referral triggers.
Martin Henley, Chief Executive of mea Platform, commented: “Insurers are ready to hand the repeated work to AI, and they will do it on one condition, which is a model that understands insurance and operates inside their own rules. That is a demanding standard and the right one, because the prize on the other side is growth. Carriers are turning away business they had the appetite to write because operations cannot keep up, and winning that back is worth more than any saving.”
The research also identifies a gap between insurers’ plans for AI and their current operating models. Some 96% of insurers have AI-led redesign on their agenda, while 13% currently have an advanced operating posture in which AI is central to workflow design. Respondents expect that figure to rise to 52% within two years.
Fully AI-native operations remain rare. ISG defines this model as one in which AI executes defined processes from end to end, while people set policy and manage exceptions. Less than 1% of insurers operate in this way today, with 12% of respondents expecting to reach that position within two years.
Henley added: “Insurers know exactly where they want the line between AI and human judgment. Getting there is the part they have not solved. The gap between 96% with redesign on the agenda and the 13% that have reached an AI-centred operating model defines the next two years.”
