Insurers Widen Gap Between A.I. Leaders and Laggards, SAS Research Finds
Insurance companies that build strong governance into their artificial intelligence systems are converting that investment into returns at a rate their peers simply cannot match, according to research by SAS in partnership with IDC.
The study, based on a survey of 2,699 decision-makers across 28 countries, found that 59% of insurance “leaders,” organizations scoring 80 or higher on the report’s Trustworthiness Index, expect strong or high A.I. return on investment. Among “laggard” insurers with low trustworthiness scores, that figure was 0%, making insurance the only focus industry in which the laggard cohort produced no strong or high ROI signal at all.
The report measures trustworthiness across five dimensions:
- Data Quality & Governance.
- Model Governance & Oversight.
- Explainability & Fairness.
- Responsible A.I. Policy.
- And Audit & Accountability.
Insurance’s overall Trustworthiness Index score rose to 63.2 in 2026 from 45.6, a gain the report called one of the largest among the industries it tracked. Seventy-eight percent of insurance leaders have cascaded a responsible A.I. policy to every employee, compared with just 3% of laggards, and automated, continuous validation and independent governance functions exist across a majority of the leader group while remaining “effectively absent” among laggards, the report said.
Leaders Chase Growth, Laggards Chase Savings
The report found a clear divide in how insurance leaders and laggards prioritize A.I. use cases. Asked what they most want A.I. to achieve, 48.5% of laggards named cost savings above all else, compared with 23.3% of leaders. Leaders instead concentrated on outcomes tied to value creation: operational efficiency (37.2% vs. 27.3% for laggards), new revenue growth (30.2% vs. 15.2%) and increasing profits (30.2% vs. 21.2%). The report described this as consistent with a broader finding across the research: deploying A.I. chiefly to cut costs is among the least effective ways to use it.
For leaders, A.I. is embedded across the decision intelligence stack, including risk assessment, underwriting, regulatory compliance, pricing, capital management, claims and fraud, the report said. One head of data and governance at a global insurer in the U.K., quoted in the report, said: “100% explainability is key. The underwriting ops team now gets a monthly report. Instead of the ops team having to review every single document, now they only focus on about 10% of the effort to complete the task manually rather than the whole thing.”
Autonomy Brings New Worries for Leaders
As insurance leaders push further into automation, their concerns are shifting from foundational data issues toward the challenges of managing autonomous systems at scale, according to the report. Thirty-four percent of high-trust insurers cited managing autonomy and human oversight as a top concern, compared with 21% of low-trust peers, which the report attributed to leaders actually running these deployments today. By contrast, 52% of laggards flagged data privacy and security as a top concern, versus 42% of leaders, suggesting laggards are still working to resolve underlying data infrastructure issues.
Within the five trustworthiness dimensions, insurance’s gains moved largely in tandem, with year-over-year increases ranging from 11.3 to 16.2 points. Model Governance & Oversight was the biggest mover, up 16.2 points, as insurers built out approval workflows and oversight structures to support automated underwriting and claims decisions, the report said. Responsible A.I. Policy moved the least, up 11.3 points, but remained the sector’s highest-scoring dimension at 69.6 out of 100, reflecting a dimension with less room left to improve.
The report said insurance has completed much of the foundational governance and policy work, with the next challenge being to sustain that pace as deployment scales so that capability and user confidence continue to advance together.
Obtain the full report here. &


