Acrisure Cuts 2,250 Jobs as Broker AI Automation Splits the Distribution Channel

On May 20, 2026, Acrisure CEO Greg Williams sent an internal letter announcing the elimination of approximately 2,250 positions, representing roughly 11% of the firm’s global workforce. The letter, which was first reported by Insurance Journal and subsequently confirmed by Crain’s Grand Rapids Business and MLive, cited “advances in technology, AI, and digital platforms” as fundamentally changing “how businesses operate, how clients expect to be served, and how value is created.”

The reduction is the largest known AI-driven headcount cut in insurance distribution. It arrives just three months after HUB International deployed Anthropic’s Claude to 20,000+ employees reporting 85% productivity gains, and one month after Baldwin Group expanded its own Claude relationship firm-wide. Both HUB and Baldwin explicitly framed their AI investments as augmentation tools that enhance existing staff. Acrisure is making the substitution case explicit: AI and automation are replacing work that humans used to do, and fewer humans are needed as a result.

The contrast is not academic. These are three of the largest insurance brokers in the United States, deploying AI on overlapping timelines with diametrically opposed workforce strategies. For actuaries building pricing models, expense projections, and workforce assumptions, the question is no longer whether broker AI adoption will reshape distribution economics. It is which model produces better outcomes for policyholders, carriers, and the brokers themselves.

Inside the Acrisure Reduction: Scale, Scope, and Stated Rationale

Acrisure is the sixth-largest insurance broker globally by revenue, generating approximately $4.8 billion in 2024 and approaching $5 billion in 2025. The firm employs roughly 19,000 to 20,000 people across insurance brokerage, cybersecurity, real estate services, and asset management. Founded in 2005 by Greg Williams and Ricky Norris in Grand Rapids, Michigan, Acrisure grew primarily through aggressive acquisition, absorbing 155 firms in 2021 alone. Its growth trajectory, from $38 million in revenue to nearly $5 billion over approximately 11 years, was fueled by private equity capital from Bain Capital ($2.1 billion in May 2025), Blackstone (invested December 2018), and Abu Dhabi Investment Authority ($725 million Series B-2 in June 2022). The company carries a $32 billion valuation as of its most recent funding round.

Williams’s internal letter laid out the reduction in specific terms. “We will be reducing our headcount by approximately 2,250 roles,” he wrote. “This process will begin today and continue in phases into 2027.” The phased timeline is significant: this is not a one-time restructuring but a sustained program of role elimination as AI and automation capabilities expand.

The letter named the driver directly: the company must leverage “AI, data, and automation to reduce manual work and create faster, more consistent outcomes.” Williams acknowledged the human cost: “This decision affects colleagues who have contributed meaningfully to building Acrisure, and was not taken lightly. It was driven by how work must evolve as we build the company we need for the future.” Affected employees were promised comprehensive severance, extended benefits, and dedicated outplacement support.

The geographic and functional scope remains partially disclosed. North America Insurance will “organize more intentionally around our lines of business,” according to the letter, but specific departments facing the deepest reductions were not identified. The reference to reducing “manual work” points toward back-office processing, data entry, submission preparation, and routine administrative functions rather than client-facing advisory or sales roles.

The Augmentation Counter-Model: HUB and Baldwin

The contrast with other major brokers deploying AI simultaneously could not be sharper. HUB International, also ranked among the top six global brokers with roughly $5 billion in revenue and 20,000+ employees, announced its firm-wide Claude deployment on February 25, 2026. HUB disclosed three headline performance metrics from its early implementation: 85% productivity gains in targeted use cases, 2.5 hours saved per employee per week on average, and 90%+ user satisfaction across early implementations. HUB has not disclosed layoffs connected to its AI deployment.

HUB President and CEO Marc Cohen framed the technology as a “force multiplier that will help accelerate HUB’s competitive advantage, which lies in combining cutting-edge technology with our scale, deep carrier relationships, product innovation, and decades of institutional knowledge to create transformational client experiences.” The language is deliberate: technology amplifies what people already do well. It does not replace them.

Baldwin Group (NASDAQ: BWIN) followed with its own firm-wide Claude expansion on May 4, 2026, positioning itself as the first named insurance customer of Anthropic’s $1.5 billion enterprise AI services joint venture. CEO Trevor Baldwin made the augmentation framing equally explicit: “Claude doesn’t replace that judgment; it gives our colleagues more time and better information to apply it.” CTO Sandeep Bajaj emphasized that “Technology and AI are only as valuable as the business outcomes they enable.”

The pilot phase at Baldwin produced what the company described as measurable improvements in client-facing insights, productivity, and workflow efficiency, though specific metrics comparable to HUB’s disclosures have not been released. Baldwin serves over 3 million clients and recently announced a $1.03 billion merger with CAC Group that would create a combined entity with roughly $2 billion in revenue.

Side-by-Side: Three Broker AI Strategies

Metric	Acrisure	HUB International	Baldwin Group
Revenue (est. 2025)	~$5B	~$5B	~$1.6B
Total employees	~19,000	20,000+	~5,000
AI workforce strategy	Substitution (2,250 cut)	Augmentation (no cuts disclosed)	Augmentation (no cuts disclosed)
AI partner disclosed	Not named	Anthropic (Claude)	Anthropic (Claude)
Productivity metric	Not disclosed	85% gains; 2.5 hrs/week saved	Not quantified publicly
Ownership	Private (PE-backed)	Private (PE-backed)	Public (NASDAQ: BWIN)
Announcement date	May 20, 2026	February 25, 2026	May 4, 2026

A pattern emerges from this comparison that goes beyond corporate messaging preferences. Acrisure has not disclosed which AI platform or vendor it is using. HUB and Baldwin both named Anthropic as their partner and deployed a specific product (Claude) with defined use cases. The level of disclosure correlates with the workforce strategy: brokers that can point to measurable productivity gains from their AI tools tend to describe augmentation. Acrisure, which cited AI as a category rather than naming a specific tool or metric, framed the technology as a rationale for headcount reduction.

The Carrier Parallel: Chubb’s 20% Reduction Plan

Acrisure is not the only major insurance firm citing AI as a driver of sustained workforce reduction. Chubb disclosed a plan to cut approximately 20% of its global workforce, roughly 8,500 to 9,000 positions from a total of approximately 43,000, over a three-to-four-year period beginning in late 2025. CEO Evan Greenberg stated that the company aims to automate 85% of major underwriting and claims processes and extract 1.5 combined ratio points in expense savings.

Greenberg made no attempt to soften the substitution framing: “In some businesses, tech and AI will replace what humans do entirely, and in others it supports and makes them more productive.” He added that “People who embrace progress and are willing to adapt have a career at Chubb. At the same time, we plan to reduce our global employee population significantly.”

The Chubb and Acrisure cases share structural similarities. Both are PE-influenced organizations (Acrisure directly, Chubb through shareholder return pressure) with leadership that has explicitly connected AI investment to headcount reduction rather than headcount redeployment. Both cite phased, multi-year timelines. And both have identified back-office processing and manual workflows as the primary elimination targets.

The difference is that Chubb announced its plan to leverage natural attrition alongside targeted reductions, while Acrisure’s 2,250-person cut appears to involve direct involuntary separations. For a carrier like Chubb with 43,000 global employees, natural turnover can absorb a significant share of a 20% target over four years. For Acrisure, an 11% reduction announced in a single letter with a phased timeline suggests the firm intends to realize cost savings faster than attrition would naturally deliver.

BLS Data Confirms the Bifurcation

Bureau of Labor Statistics occupational projections for 2024 to 2034 reveal a structural split in insurance employment that aligns precisely with the divergent broker strategies emerging in 2026.

Back-office and processing-intensive roles face sustained contraction. The BLS projects a 3% decline in insurance underwriter employment over the decade, consistent with the automation of risk selection and pricing workflows. Claims adjusters, appraisers, examiners, and investigators face a steeper 5% decline, though the BLS notes approximately 21,600 annual openings will persist due to retirements and turnover replacement needs.

Client-facing distribution roles tell a different story. Insurance sales agents, the frontline of the broker distribution model, are projected to grow 4% over the same period, matching the average for all occupations. This growth pattern suggests that AI is not displacing advisory and relationship functions at the same rate it is replacing processing and administrative tasks.

The total U.S. insurance industry employed approximately 2.98 million workers in 2023, according to the Insurance Information Institute citing BLS NAICS 524 data. The BLS projects actuarial employment to grow 22% through 2032, far outpacing total employment growth. This juxtaposition, shrinking processing roles alongside growing analytical and advisory roles, is the labor market version of the augmentation-versus-substitution debate playing out in broker strategy.

Submission Quality and Downstream Actuarial Effects

The divergent workforce strategies at major brokers carry direct implications for the data actuaries use to price risk, develop reserves, and validate underwriting models. Submission quality is not a peripheral concern: it is the upstream input to every carrier pricing decision.

The Substitution Pathway and Data Risk

When a broker eliminates 2,250 processing and back-office roles, the question for carrier actuaries is who, or what, now performs the work those roles handled. If AI systems take over submission preparation, data entry, and policy administration, the quality of that output depends entirely on the AI’s training data, validation procedures, and error-correction mechanisms. Human processors introduced errors, certainly, but they also caught anomalies, flagged unusual risk characteristics, and applied contextual judgment that automated systems may not replicate.

Actuaries building loss development factor assumptions for lines serviced heavily by substitution-model brokers should consider whether submission quality will shift during the transition. A phased multi-year reduction means the workforce operating in 2027 will be structurally different from the one producing the experience data in actuaries’ current loss triangles. The credibility standard in ASOP No. 25 does not directly address upstream data pipeline changes of this nature, but the principle of evaluating whether historical experience remains representative applies.

The Augmentation Pathway and Data Enrichment

Brokers on the augmentation track present a different actuarial dynamic. If HUB’s 85% productivity gains translate into more complete submissions, faster turnaround, and richer supplemental data, carriers receiving those submissions may see improved risk selection without changing their own models. The shift of agentic AI into the producer channel is already reshaping the submission pipeline, with tools designed to extract, validate, and enrich exposure data before it reaches the carrier.

For actuaries, augmentation-model broker submissions may warrant different credibility assumptions than substitution-model broker submissions. The former should, in theory, produce higher-quality data over time. The latter may produce equivalent or better data in steady-state (fully automated systems can be highly consistent), but the transition period introduces uncertainty about data quality that current actuarial standards do not explicitly address.

Expense Ratio Projections

Carrier expense ratio assumptions in rate filings typically model the carrier’s own operating costs. But submission handling costs are shared between broker and carrier: every manual touch a carrier underwriter adds to clean up an incomplete submission is an expense that would disappear if the submission arrived complete. If substitution-model brokers reduce headcount but maintain or improve submission quality through automation, carriers receiving those submissions should see measurable reductions in underwriting processing expenses. If quality drops during the transition, carrier costs could temporarily increase.

Actuaries preparing rate indications for commercial lines with significant broker distribution channels should consider modeling a range of submission quality scenarios rather than assuming the historical average will persist. The variance across broker AI strategies makes a single-point assumption less defensible than it was even 12 months ago.

The Private Equity Dimension

Acrisure’s ownership structure is relevant to understanding why the firm chose the substitution path. The company has raised billions from Bain Capital, Blackstone, and Abu Dhabi Investment Authority. Its $32 billion valuation implies investor expectations of sustained revenue growth and margin expansion. An 11% headcount reduction directly improves the expense ratio, a metric that PE-backed firms optimize aggressively as they position for eventual exits or public market listings.

HUB International is also PE-backed, having been acquired by Hellman & Friedman in 2018. But HUB’s AI strategy emphasizes revenue enhancement over cost cutting: the six-pillar deployment includes digital direct-to-customer experiences and agentic workflows that could drive new revenue streams. Baldwin, as a public company, faces quarterly earnings scrutiny that incentivizes both growth and cost discipline, yet chose the augmentation framing over workforce reduction.

The divergence may reflect different stages of the PE lifecycle rather than fundamentally different views on AI capability. Acrisure, with a recent $32 billion valuation and significant debt from its acquisition-driven growth model, faces pressure to demonstrate that its revenue per employee can rise substantially. Eliminating 2,250 roles while maintaining revenue immediately improves that ratio. HUB, at a similar revenue scale but with a different leverage profile, may have more flexibility to invest in AI as a growth tool rather than deploying it as a cost-reduction mechanism.

Regulatory Considerations

State insurance regulators have been focused primarily on carrier AI use, particularly in underwriting and claims decisions that directly affect consumers. The NAIC Model Bulletin on AI, adopted in late 2023 and now implemented in nearly half of U.S. states, requires insurers to establish written governance programs for AI systems, document decision-making processes, and ensure AI-influenced decisions comply with existing anti-discrimination and consumer protection statutes.

Broker AI automation raises different regulatory questions. Insurance brokers and agents operate under state licensing requirements and suitability obligations that assume human judgment in the recommendation process. If AI systems take over functions that were previously performed by licensed individuals, such as coverage comparison, risk assessment for recommendation purposes, or policyholder communication, state regulators may ask whether the remaining human oversight satisfies the suitability standard.

Colorado has moved furthest on AI governance. SB24-205, the Colorado Artificial Intelligence Act, took effect on February 1, 2026, with compliance deadlines extending through July 2026. The law addresses “consequential decisions,” defined as determinations affecting “a consumer’s access to or eligibility for” insurance or “differentiated price, cost sharing, compensation, or other material terms.” A follow-up bill, SB26-189, signed in May 2026, revised certain provisions based on early implementation experience.

Whether broker-side AI automation falls within the scope of consequential decision regulations remains untested. If an AI system at a substitution-model broker effectively selects which carriers to approach, which coverage forms to recommend, or how to present risk information to underwriters, those functions touch the suitability obligation even if the final placement decision involves a human signature. The gap between adoption ambition (82%) and scaled deployment (7%) that characterized the carrier AI landscape 12 months ago is now appearing on the distribution side, and regulators have not yet caught up.

What the Industry Data Shows About AI and Insurance Employment

The broker AI split mirrors broader industry trends that multiple research firms have documented. According to Deloitte’s December 2024 research, improved efficiency and productivity is the most significant realized benefit from generative AI implementations in insurance. BCG research has found that insurance leads other sectors in AI adoption, though an “implementation gap” persists between adoption announcements and realized value.

The insurance AI ROI challenge is particularly acute for substitution-model deployments. Augmentation investments can be evaluated by measuring productivity gains (HUB’s 85% figure, for instance) without requiring workforce reductions to justify the spending. Substitution investments must demonstrate that the eliminated roles were truly redundant, meaning the work either stops being done or is performed at equivalent or better quality by the automated system. If quality degrades, the short-term cost savings may be offset by downstream losses from poor submissions, compliance failures, or customer attrition.

ScienceSoft noted in Q1 2026 that agentic AI is the most actively pursued AI type in insurance. The shift toward agentic systems, which can execute multi-step workflows autonomously, makes the substitution pathway technically feasible in ways that earlier AI tools did not support. Simple chatbots could not replace a submission processor; an agentic system that extracts data from applications, validates it against carrier appetites, and assembles a complete submission package potentially can.

Why This Matters for Actuaries

The Acrisure reduction is not an isolated headcount story. It is the first large-scale test of whether the insurance distribution channel can sustain significant workforce reduction without degrading the quality of risk data that carriers and actuaries depend on.

Three specific actuarial implications deserve attention:

Loss development assumptions may need segmentation by broker AI strategy. If submission quality diverges between augmentation-model and substitution-model brokers, aggregate loss development patterns may mask offsetting trends. Actuaries working on commercial lines with mixed broker channels should consider whether stratifying experience data by submission source improves the accuracy of indicated rates.

The insurance workforce crisis is evolving, not resolving. The industry narrative of a talent shortage coexists with Acrisure cutting 2,250 jobs and Chubb planning to eliminate 8,500 to 9,000 over four years. The resolution is that the shortage and the surplus are occurring in different roles: analytical, advisory, and technical positions remain scarce while processing, administrative, and back-office roles face contraction. Actuaries, whose BLS growth projection of 22% places them firmly on the growth side, should recognize that the workforce surrounding them is shrinking even as demand for their work increases.

Expense assumptions in rate filings require updating. The distribution channel’s cost structure is changing in real time. If major brokers achieve the efficiency gains they claim (or that their headcount reductions imply), the commission and brokerage expense that flows through rate filings will reflect a different cost basis than the one embedded in historical factors. Actuaries using trended expense ratios should consider whether the trend is capturing a structural shift rather than a cyclical fluctuation.

The insurance distribution channel has spent decades building relationships, placing risk, and servicing policies through a labor-intensive model. Acrisure’s 2,250-job reduction and HUB’s augmentation-first deployment represent the two poles of how that model changes under AI pressure. The next 12 to 18 months will reveal which approach produces better policyholder outcomes, carrier relationships, and financial performance. For actuaries, the task is to build pricing and reserving models flexible enough to accommodate either outcome, or more likely, a distribution landscape where both strategies coexist and the data implications of each must be modeled separately.