— Insurance companies are expanding their use of artificial intelligence while some are reducing investment in the employees responsible for working with the technology, according to new research examining operational pressures across the industry.

Covenir's 2026 Insurance Operations Leaders Trends Report found that 70% of surveyed insurance operations leaders have AI running in live business functions, compared with 58% in 2025. However, 20% identified training and development as an area where they plan to reduce investment, while only 7% said they were actively protecting training budgets.

The findings come from a survey of 152 U.S. insurance operations decision-makers at carriers, managing general agencies and insurtech companies conducted in February and March 2026.

AI adoption is producing measurable benefits, particularly in customer service and claims administration. Among respondents identifying areas of improvement, 31% cited increased customer satisfaction, 29% reported reduced handling times and 23% identified faster claims resolution.

Those gains have not eliminated workforce pressures. Nearly two-thirds of respondents said their teams were more stretched than ever, including 91% of executives. Mid-sized organizations with 500 to 999 employees reported particularly high strain, despite 92% of respondents in that company-size category reporting increased budgets.

David Squibb, president and CEO of Covenir, argues that employee training needs to be included in AI implementation budgets rather than treated as a separate expense that can be reduced when financial pressures increase.

The concern is particularly relevant to claims handling, where automated systems may assist with document processing, customer communications and claims decisions but still require employees to evaluate exceptions and manage complex policyholder interactions.

The survey identified first notice of loss as the most frequently reported point of breakdown between an insurer's promised service and the policyholder's actual experience. Forty-two percent of respondents selected FNOL as the primary problem area, followed by billing and payments at 21%.

Workforce investment is also changing among organizations with more advanced AI deployments. Among respondents whose companies have AI operating across multiple business functions, 54% identified headcount as the area where they plan the largest investment reductions in 2026. That compares with 11% among organizations using AI in only one production function.

The report cautions that the relationship does not establish whether AI-enabled efficiency, technology spending or other factors are driving those workforce decisions.

Data management presents another challenge. Although 81% of respondents considered operational insights critical to business success, 29% said their organizations were not using those insights, and another 18% reported difficulty turning collected information into operational decisions.

Covenir's findings also suggest insurers increasingly expect outside service providers to contribute to their technology strategies. Twenty-nine percent of respondents identified embedded AI capabilities as a new expectation for outsourcing partners.

For claims departments evaluating additional automation in 2027, the findings point to practical considerations involving employee preparation, accountability for automated recommendations and the ability to act on operational data. Squibb recommends assigning responsibility for handoffs between AI systems and employees, measuring their performance and making training part of the original deployment budget.