Insurance Agencies Embrace AI Despite Governance Gaps, Survey Finds

Independent insurance agencies are racing to deploy artificial intelligence tools even as the vast majority operate without formal governance frameworks, according to a new industry survey that underscores a growing tension between technological ambition and operational readiness.

The survey, conducted by the Agents Council for Technology, found that 68% of respondents are either very likely or somewhat likely to increase AI use over the next year. Yet only 13% of agencies have established formal AI policies, while more than half reported having no policy at all. The findings reveal what the council described as "a growing gap" between the promise of AI and the infrastructure required to deploy it responsibly.

Operational efficiency and staff productivity topped the list of motivations for adoption, cited by 60% and 52% of respondents respectively. But concerns about data privacy, compliance risks, and inaccurate outputs remain significant barriers. About 45% of agencies reported already using public large language models like ChatGPT, while adoption of specialized tools—policy comparison platforms, marketing automation, chatbots—lagged far behind, each registering below 20%.

"AI in its current form should be treated like a junior colleague," the council cautioned in its report. "Although it's fast and capable of consuming large volumes of information, it still requires supervision for complex or high-impact decisions."

Kasey Connors, executive director of the council, called the findings "a pivotal moment" for independent agencies, warning that "long-term success hinges on clearer governance, stronger training, and more integrated technology strategies."

The insurance sector's governance deficit mirrors broader patterns across industries. A separate cybersecurity analysis found that 63% of organizations lack AI governance policies altogether, while IBM research showed that shadow AI—employees using unsanctioned generative tools—added an average of $670,000 to data breach costs where present. Among breached organizations that experienced AI-related security incidents, 97% lacked proper access controls.

(The insurance survey comes as a 12-state pilot program begins evaluating how insurers use AI systems, marking what legal observers describe as the most wide-ranging effort yet to create uniform standards for the sector. The initiative arrives ahead of the EU AI Act's substantive obligations, which take effect in August and will require high-risk AI systems to demonstrate compliance with requirements around risk management, data governance, transparency, and human oversight.)

The insurance industry's AI adoption trajectory reflects competing pressures across knowledge work sectors. In California, legislators are advancing bills that would require 90-day notice for AI-driven layoffs and mandate human oversight of disciplinary actions. Economist Gad Levanon of the Burning Glass Institute has urged caution against what he termed "AI Derangement Syndrome," noting that actual job displacement has been minor and most AI use has involved augmentation rather than replacement. Yet labor challenges are mounting: Hollywood writers struck in 2023 over AI concerns, and Kaiser Permanente workers are now contesting AI introduction as a threat to mental health and administrative roles.

The insurance council identified multiple constraints hampering responsible AI implementation, including lack of documented processes, vendor confusion, resource limitations, portal fatigue, security gaps, and what it termed "change fatigue and tool sprawl." Nearly a third of survey respondents said they are not currently using AI, while 33% described themselves as "just experimenting" and 22% reported use only in limited areas. More than 44% of agencies rely on peer-to-peer training rather than structured programs for new technology systems.