AI Delays Cost Insurance Industry Up to $72.8 Million Daily
The insurance industry faces significant financial setbacks due to AI-driven prior authorization delays in property and casualty insurance rate filings. According to a report by InsurTech firm ZestyAI, such delays could result in potential losses of approximately $72.8 million daily. Their analysis, which reviewed over 2 million P&C rate and form filings through their regulatory intelligence platform, ZORRO Discover, suggests that these delays arise more from operational errors than from disagreements over pricing strategies.
ZestyAI's research points out that the root of these issues lies in administrative errors such as incomplete documentation, missing exhibits, inconsistent data, and unclear pricing justifications. The study, which examined 147 homeowners' insurance filings approved in prior-approval states in the latter half of 2025, noted an average rate increase of 8.49%. With direct written premiums for homeowners at $51.7 billion, each day of delay represents nearly $12 million in unrealized pricing adjustments.
Further analysis reveals that personal auto rate filings also suffer from these delays. Unchallenged submissions are typically approved in around 14 days, while those met with regulatory objections can take up to 51 days on average. Common issues leading to objections include filing packaging errors, lack of adequate actuarial support, undocumented rating factors, incomplete predictive model documentation, and unclear catastrophe assumptions.
State-specific data highlights that 70% of disputed filings in Connecticut had missing actuarial documentation, 83% of objections in Montana involved insufficient rating factor support, and wind-loss model challenges were frequent in Florida. Such issues underscore the necessity for comprehensive compliance procedures for regulatory approval.
Bryan Rehor, ZestyAI's senior director of regulatory and government affairs, emphasizes that these delays stem from mechanical execution errors, such as incorrect documentation and unsupported discount factors. He clarifies, “Regulatory delay isn’t really about regulators slowing things down – it’s a function of execution.”
Rehor further notes that every objection resets the review cycle, possibly broadening its scope and resulting in additional follow-ups. He posits that leveraging technological advancements can help maintain consistency across different jurisdictions and filing cycles, minimizing the need for frequent interventions. The findings stress insurers' growing need to refine their filing processes and enhance transparency in predictive modeling and pricing methodologies.