Morgan Stanley Research published an 86-page report in January 2026 projecting that AI automation will take 200 basis points off P&C expense ratios by 2030, worth $9.3 billion of additional operating income across 16 carriers and five brokers in its coverage universe.

The headline is a 2030 steady state. The part that changes a 2026 filing is the shape of the path to it, which the report models as a J-curve that goes negative first.

Key Takeaways

  • The $9.3 billion is the 2030 gap between operating income with AI ($92.1 billion) and without ($82.7 billion), an 11% difference, from a cohort expense ratio of 28.5 against 30.5.
  • For 2026 the model shows $6.0 billion of gross cost savings but only 10% of it, $600 million, reaching operating earnings, against $3.0 billion of implementation cost: a $2.4 billion operating income reduction.
  • That puts the 2026 post-AI operating margin at 14.7% against a 15.2% pre-AI baseline, 50 basis points worse, with breakeven in 2027.
  • BCG finds only 38% of P&C insurers currently generating value at scale from AI in core workflows, against a model that assumes full savings realization by 2030.
  • Progressive carries the cohort's lowest automation rate at 20.7% and planned to hire more than 12,000 people in 2025, which is savings absorbed as growth rather than expense ratio.
27.7 → 25.3
U.S. P&C expense ratio, 2014 vs. 2024 (AM Best)

The Forecast Asks AI to Repeat a Decade in Five Years

The starting point is AM Best's series on US P&C underwriting expenses, which fell from 27.7 in 2014 to 25.3 in 2024.

That 2.4 point decline was not evenly sourced. Other acquisition expenses, covering rent, technology and administrative overhead beyond commissions, supplied 1.9 points of it. General expenses supplied 0.5. Commission and brokerage ratios stayed roughly flat, so distribution economics did not change; occupancy and operating overhead did, with AM Best flagging the shift to hybrid and remote work as a contributor.

Morgan Stanley's projection is a 2.0 point reduction in the cohort expense ratio between 2026 and 2030, from 30.5 to 28.5. That is the same order of improvement the industry took ten years to produce, compressed into five, from a different cost category.

The $9.3 billion headline is a difference, not a saving. It is the gap between projected 2030 operating income of $92.1 billion with AI and $82.7 billion without, an 11% uplift, applied across the premium base of those 16 carriers. Gross savings are larger; implementation costs consume much of them in the early years.

The J-Curve Is the Part That Reaches a 2026 Expense Provision

The report models implementation cost exceeding realized savings in the first year, so early adopters carry a margin drag before the efficiency compounds.

Year Pre-AI Margin Post-AI Margin Net Effect
2026 15.2% 14.7% -50 bps
2027 ~15.3% 15.4% +10 bps
2028 ~15.4% 15.6% +20 bps
2029 ~15.5% 16.2% +70 bps
2030 15.6% 17.4% +180 bps

The 2026 line does the work. Morgan Stanley identifies $6.0 billion of gross cost savings across the cohort that year, of which only 10%, or $600 million, flows through to operating earnings. Against that sit $3.0 billion of AI implementation costs, netting to a $2.4 billion operating income reduction. Post-AI margin lands at 14.7% against the 15.2% that would have prevailed without the investment.

For anyone setting an expense provision, the sign matters more than the size. A rate indication that reflects the 2030 steady state understates the expense load in the filing year, and it does so most for the carriers automating hardest. The relationship inverts: through 2026 and into 2027 the aggressive adopter carries the higher expense ratio, and the ranking only flips as the curve crosses in 2027 and widens through 2029 and 2030.

Chubb is the cleanest test of that, having committed to 1.5 combined ratio points of run-rate savings, 85% automation of key underwriting and claims functions, and a 20% headcount reduction covering roughly 8,600 of 43,000 employees, all of which lands in quarterly statements. Morgan Stanley nonetheless puts Chubb's 2030 earnings uplift at 9%, below AIG's 13% and well below Assurant's 27%, because uplift tracks workforce composition rather than announced ambition.

The Model Measures Capability, Not Behavior

The automation rates come from the Anthropic Economic Index mapped through Department of Labor O*NET occupational codes and LinkUp job posting data, which is a more rigorous build than a top-down analyst estimate. It also measures a different thing than the forecast needs.

The index captures what tasks AI can automate today. Business and finance occupations show 94.3% theoretical AI coverage while actual adoption sits far lower, and BCG finds only 38% of P&C insurers currently generating value at scale from AI in core workflows. Projecting full realization across the cohort by 2030 requires either a sharp acceleration or the assumption that these 16 large, well-capitalized carriers are disproportionately inside that 38%. The second is plausible and unstated.

The larger gap is behavioral. Savings only reach the expense ratio if automation potential converts into headcount reduction or redeployment, and the cohort is split on that. Chubb is reducing. AIG reports higher submission volumes without additional human capital. Progressive carries the lowest automation rate in the cohort at 20.7%, still shows an 8% earnings uplift on scale alone, and planned to hire more than 12,000 people in 2025. A carrier that takes its AI productivity as volume rather than cost produces a smaller and slower expense ratio improvement than the model books.

The broker side shows the same distinction more starkly. The five brokers carry a higher average automation rate of 25.1% and a projected 350 basis point margin improvement, nearly double the carriers' 180, yet Morgan Stanley estimates they need five years to capture just 50% of it. Capability is ahead of realization there by the report's own arithmetic. PwC's work adds the cost that sits outside the model: businesses cutting entry-level roles aggressively report burnout among senior staff absorbing displaced work and thinner talent pipelines, which is an expense that arrives later and in a different line.

Further Reading

Sources

AI in insurance P&C insurance actuarial pricing
Feedback

We are seeking feedback on how to improve the site and deliver high-quality content relevant to actuaries. Help us make it better.

Submit feedback

Stay ahead with daily actuarial intelligence - news, analysis, and career insights delivered free.

Subscribe to Actuary Brew Browse All Insights