Most insurers are willing to let AI handle repeatable operational tasks, but they draw a firm line at consequential decision-making, according to new research from ISG commissioned by mea Platform. The global survey of senior insurance leaders found that while 83% support AI for repetitive work, 86% insist that decisions distinguishing a company must stay with people.
The study surveyed leaders across underwriting, operations, claims, technology, and transformation in North America, Europe, and Asia. It examined 20 operational activities, including submission intake, triage, quote generation, bordereaux processing, claims adjudication, and compliance screening. The findings reveal an industry cautiously adopting AI Agents & Automation while building guardrails around governance.
Governance defines trust in AI systems
The research defined governed AI as a system drawing from controlled policy wording, endorsements, underwriting appetite, and claims guidance. Permissions determine what the system may read and what it may decide, and a designated person can override or stop the system at any point. The AI operates according to an organization's appetite, thresholds, and referral triggers.
For high-consequence underwriting and claims decisions with limited oversight, 75% of respondents favored an insurance-specific model or a governed hybrid approach. Only 6% said they would trust a general-purpose model on its own for these decisions. The preference is clear: insurers want domain-specific systems with organizational controls, not off-the-shelf AI making judgment calls.
Operational capacity creates a bottleneck
Carriers estimated that one in nine broker submissions is declined or left unquoted because operations cannot keep up, even when the carrier has an appetite for the risk. This capacity gap helps explain why 83% of respondents support automating repeatable work. Insurers see AI as a way to process more submissions without adding headcount.
Among organizations already using AI in operations, 61% reported productivity improvements and 51% reported faster cycle times. Respondents expect operating costs to fall 16% over two years. When asked what would most improve brokers' perceptions of carriers, 64% cited pricing, 52% cited ease of doing business, and 51% cited AI-driven speed and completeness of submission responses.
AI-native operations remain rare
The study found that 96% of insurers have AI-led redesign on their agenda, but current adoption tells a different story. Only 13% have reached an advanced operating posture where AI plays a central role in workflow design. Another 52% intend to reach that stage within two years. Less than 1% operate on a fully AI-native basis today, where AI executes defined processes end-to-end while people establish policies and manage exceptions. About 12% expect to reach that stage within two years, and 27% of the market remains in evaluation mode.
Under the approach most insurers prefer, underwriters and brokers retain responsibility for risk decisions while AI handles repeated tasks and provides information that supports human judgment. The technology is positioned as an enabler of faster, more consistent operations, not a replacement for underwriting expertise.
Why this matters for insurance professionals
The research confirms that AI adoption in insurance is accelerating, but the focus is narrow: automating high-volume, repeatable work to close the submission gap. For underwriters and claims professionals, the immediate impact will be less manual triage and faster access to structured information for decision-making. The study also signals that carriers investing in AI for Insurance are linking adoption directly to broker satisfaction metrics. Speed and completeness of responses now rank alongside pricing as factors that shape broker perceptions, which means operations teams will face growing pressure to deliver AI-assisted turnaround times.
Your membership also unlocks:
