Deloitte, Oliver Wyman, and McKinsey Map Insurance AI Priorities for 2026: Where They Agree, Where They Diverge

From comparing the strategic frameworks that McKinsey, Deloitte, and Oliver Wyman publish for the insurance industry each cycle, 2026 marks the first year where all three firms explicitly rank AI scaling above M&A and geographic expansion as the top CEO priority. That consensus is notable. The disagreements underneath it are more useful.

Oliver Wyman released its “10 To-Dos for Insurance CEOs” in February 2026, calling AI infrastructure a “once-in-a-generation buildout.” Deloitte published its 2026 Global Insurance Outlook identifying a 65-percentage-point gap between AI intent and execution. McKinsey’s latest insurance M&A report projects accelerating deal flow with AI as a catalyst. Each firm draws from its own survey methodology, client base, and analytical lens, and the resulting divergences on combined ratio trajectories, workforce transformation speed, and private capital’s structural role create genuine strategic uncertainty for actuarial leaders trying to set assumptions for the next 12 to 24 months.

This analysis synthesizes all three frameworks, identifies the specific points of convergence and disagreement, and extracts the signals that matter most for pricing, reserving, and capital modeling work.

The Triple Convergence: AI Scaling as Priority Number One

The alignment across the three reports is striking in both emphasis and language. Oliver Wyman dedicates three of its ten CEO priorities to AI: building AI infrastructure (priority five), creating an AI-native workforce with HR and IT as co-owners (priority nine), and redesigning the operating model for what it calls “hyperspeed” (priority ten). The firm’s Quotient practice, its fastest-growing advisory team, has advised on more than $50 billion of capital investment in AI deployment across industries.

Deloitte’s “Scaling Gen AI in Insurance” companion report, based on a survey of 200 U.S. insurance executives, finds that 76% of insurers have implemented generative AI in one or more business functions. Life and annuity carriers lead at 82% adoption, while P&C carriers trail at 70%. The report cites 5x productivity gains in underwriting through AI assistant implementations and projects up to $160 billion in P&C fraud detection savings by 2032.

McKinsey frames AI as an accelerant for its core thesis: that M&A is the primary growth engine in a fragmented insurance market. The firm’s data shows AI and generative AI are “shortening deal cycles by 10% to 30%” and “cutting M&A costs by around 20%,” while 18 “arena” technology sectors (software, AI infrastructure, data platforms) now capture roughly 40% of global deal value, up from 7% two decades ago.

Supplementary data from Grant Thornton’s 2026 AI Impact Survey of 100 insurance leaders reinforces the pattern: 52% report AI-enabled revenue growth, 62% report improved decision-making, and 62% rate their AI maturity as “scaling across multiple functions.” From tracking these surveys over consecutive years, the shift from “exploring AI” to “scaling AI” as the default self-assessment is a meaningful change in how insurance executives describe their own organizations.

The Combined Ratio Divergence: Deloitte’s 99% vs. AM Best’s 96.9%

Underneath the AI consensus lies a material disagreement on P&C profitability that directly affects reserve and pricing assumptions. Deloitte projects a 99% industry combined ratio by year-end 2026, deteriorating from 98.5% in 2025 and 97.2% in 2024. AM Best forecasts 96.9% for 2026, which, while up 1.9 points from the 95.0% actual result in 2025, still implies a profitable underwriting year for the industry.

That 2.1-point gap between Deloitte’s and AM Best’s forecasts represents billions of dollars in implied underwriting income. For context, AM Best’s 2025 actual results showed net underwriting income of $39 billion, more than double the prior year, with catastrophe losses contributing just 6.9 points to the combined ratio (down from 8.4 in 2024). The core accident-year combined ratio held at 89.5%, consistent with 89.3% in 2024.

The sources of the forecasting gap appear to be assumptions about catastrophe loads, rate adequacy in softening commercial lines, and the speed at which social inflation erodes liability reserves. Deloitte’s outlook emphasizes rising claims costs from higher repair and material prices, plus third-party litigation financing straining commercial results. AM Best acknowledges these headwinds but projects more modest deterioration, with commercial lines combined ratio rising to 96.3% from 95.8%.

Three commercial lines already exceeded 100 in 2025 according to AM Best: commercial auto at 103.5, medical professional liability at 106.0, and other/products liability at 108.0. Whether those stressed lines worsen or stabilize in 2026 is a key swing factor in the combined ratio debate.

For reserving actuaries, the choice of which forecast to anchor on has practical consequences. A 96.9% assumption suggests continued favorable development and room for conservative IBNR selections. A 99% assumption signals that the favorable development cushion is thinning and that actuaries should stress-test their picks against a near-breakeven scenario. The 19-year streak of favorable prior-year reserve development that AM Best documented through 2025, with year-end 2024 reserves showing a $9 billion deficiency that was nearly $10 billion better than originally estimated, provides some comfort. But streaks eventually end, and the consulting firm divergence on 2026 is a useful signal that the consensus is less settled than industry-level figures might suggest.

The 65-Point Gap: Why AI Intent Fails to Become AI Action

Deloitte’s most striking finding is the chasm between what insurance executives say about AI and what they do about it. While 90% of surveyed executives agree on the urgency of reinventing the employee value proposition to reflect human-machine collaboration, only 25% have taken tangible action to elevate human skills. That 65-percentage-point gap between recognition and execution defines the industry’s AI challenge more precisely than any adoption statistic.

Oliver Wyman’s CEO Forum survey of 415 CEOs across industries (representing roughly 10% of global market capitalization) provides additional texture. Among the findings: 67% of CEOs remain in planning or pilot stages on AI, not scaling. Only 12% qualify as “AI ROI leaders” demonstrating firm-wide impact above 10%, a decline from 17% the prior year. And 53% say it is too early to assess AI ROI, up from 41% in 2025. For insurance and asset management CEOs specifically, 58% are prioritizing the redesign of roles and workflows as AI reshapes business models.

Deloitte’s gen AI survey identifies the failure modes more precisely. The top barrier to scaling AI in insurance is not funding; underfunding was notably absent from the ranked barriers. Instead, the primary obstacles are: lack of business line support, poor data and AI foundations, legacy IT infrastructure, and inadequate cross-functional collaboration. Talent availability and existing skillsets rank as the areas where firms are “least prepared.”

Grant Thornton’s insurance-specific data adds a governance dimension: only 24% of insurance leaders are “very confident” they could pass an independent AI governance review within 90 days. Another 68% say AI controls exist but evidence is “fragmented across teams and tools.” When 44% report that governance and compliance challenges contributed to AI project failure, the picture becomes clear. The binding constraint on insurance AI is not capital or technology; it is organizational readiness and governance infrastructure.

Patterns we’ve seen across the AI governance research confirm this framing. The gap between Deloitte’s 76% “implemented gen AI” figure and Oliver Wyman’s 12% “AI ROI leaders” figure is not a contradiction. Most carriers have deployed point solutions (chatbots for customer service, document extraction for claims) without achieving the enterprise-wide integration that generates measurable financial returns. Pilots proliferate; scaled deployments with board-level ROI evidence remain rare.

Oliver Wyman’s Hyperspeed Operating Model vs. Actuarial Governance

Oliver Wyman’s tenth priority for insurance CEOs, redesigning the operating model for “hyperspeed,” deserves particular scrutiny from actuarial leaders. The concept calls for shorter decision paths, faster feedback loops, and an AI-powered insights engine that enables execution to match the pace of market change. As the report frames it: “Everything has moved at hyperspeed: AI is shortening innovation cycles, competitors are making deals, and geopolitics are shifting quickly.”

The tension with actuarial governance is immediate and structural. ASOP No. 56 requires actuaries to evaluate model appropriateness, assess data quality, perform validation and testing, and ensure adequate governance controls before models influence business decisions. Peer review cycles, which the American Academy of Actuaries recommends as part of quality control, add additional time to model deployment. When Oliver Wyman advises CEOs to operate at “hyperspeed,” actuaries hear a framework that compresses the very review processes their professional standards mandate.

This is not hypothetical. The 82% adoption, 7% scale gap documented in recent Sedgwick data reflects, in part, the friction between management’s deployment speed and the governance layer that actuaries and risk managers impose. Oliver Wyman’s recommendation that HR and IT become “true co-owners” of AI transformation is directionally correct, but notably absent from the framework is any mention of the actuarial function’s role in model governance. For carriers where the chief actuary or appointed actuary signs off on AI-influenced pricing and reserving models, the hyperspeed vision needs a governance chapter that the consulting report does not provide.

The practical resolution is not to reject speed but to distinguish between AI applications that require full actuarial validation (pricing models, reserve projections, capital model components) and those that can move through lighter governance (operational efficiency tools, document processing, customer service automation). Blurring that distinction creates both regulatory and professional risk.

Private Capital as a Permanent Structural Force

Oliver Wyman’s fourth CEO priority, “decide your role in the asset-management-led insurer model,” acknowledges private capital as a “permanent competitor, partner, and benchmark” for traditional insurers. This framing aligns with Deloitte’s data showing reserves ceded to sidecars nearly tripled between 2021 and 2023, reaching $55 billion, while private placements account for 21.1% of total insurance assets under management, up from 20% in 2023.

The private capital angle matters for AI strategy because PE-backed carriers and platforms approach technology investment with fundamentally different time horizons and return expectations. McKinsey reports that PE deal value climbed 54% to $1.2 trillion globally in 2025, with sponsors active across life back-book platforms, specialty carriers, MGAs, TPAs, and insurance infrastructure technology. MGAs, favored for their capital-light and high-margin business models, account for roughly 5% of deal flow.

Deloitte’s data shows 61% of CFOs and CIOs globally expect private credit to provide the highest returns over the next year, with even higher expectations in the Americas (64%) and Asia-Pacific (69%). For actuaries performing asset-liability matching and reserve credit analysis, the growing penetration of private credit and alternative assets into insurance portfolios creates valuation and liquidity risk considerations that traditional fixed-income-heavy portfolios did not present.

The AI investment implications diverge by ownership structure. PE-backed platforms tend to deploy AI with shorter payback requirements (two to three years) and higher tolerance for disrupting existing workflows. Mutual carriers and legacy stock companies, with their longer time horizons and stronger actuarial governance cultures, tend to approach AI as incremental improvement rather than transformational rebuild. Oliver Wyman’s framing of private capital as a “permanent benchmark” suggests that traditional carriers will increasingly face pressure to match the AI-enabled expense ratios that PE-backed competitors achieve, even if their governance standards and workforce obligations require different implementation paths.

M&A as the AI-Enabled Growth Engine

McKinsey’s primary thesis centers on M&A as the dominant growth mechanism in a structurally fragmented insurance market. The data is compelling: global deal value rose 43% to $4.7 trillion in 2025, with transactions exceeding $10 billion jumping to 60 deals, the highest count since the post-COVID 2021 peak. Insurance-specific M&A reached approximately $104 billion in 2025, up from $88 billion the prior year, with average deal size growing to $1.1 billion.

The fragmentation thesis resonates particularly in European markets, where McKinsey notes the top five players in the UK, Italy, and Germany hold less than 55% of overall market share. In the U.S., roughly 39,000 independent agencies illustrate the brokerage market’s fragmentation. The insurance brokerage market, valued at $359 billion in 2026, is projected to reach $521 billion by 2031, a 9.8% compound annual growth rate that attracts both strategic and financial acquirers.

Divestitures surged 30% in 2025 to $1.6 trillion globally, the highest since 2021, with particular focus on life and annuity back-book disposals, exits from smaller geographies, and carve-outs of non-core books. For actuaries, the M&A acceleration creates demand for due diligence expertise, particularly around loss reserve adequacy, embedded value calculations for life blocks, and the integration of disparate data platforms and pricing models that follow any acquisition.

McKinsey’s observation that technology “arena” sectors now command 27x EV/EBITDA multiples compared to 16.5x for established industries helps explain why insurtech and insurance infrastructure technology attract disproportionate investment. When AI capabilities become a valuation driver, carriers without credible AI strategies face both competitive disadvantage in operations and lower acquisition multiples in capital markets.

Workforce Compression and the Junior Role Question

The workforce implications across the three reports demand attention from actuarial leaders and exam candidates alike. Oliver Wyman’s CEO Forum survey found that 43% of CEOs across industries are now reducing junior roles, a sharp increase from 17% in 2025. Another 45% are holding headcount flat (up from 31%). Mega firms report 5x more AI-driven cost savings than midsize companies, suggesting that scale advantages in AI deployment translate directly into workforce restructuring.

Deloitte adds a cautionary finding: employees perceive employers as 2.3 times less empathetic and human when AI tools are offered. That perception gap matters for an industry already struggling with talent attraction. Grant Thornton’s insurance-specific data shows only 7% of the insurance workforce is “fully ready” to adopt AI, with 47% rated as “mostly ready” and the remainder requiring significant upskilling. Meanwhile, 29% of insurance leaders cite talent and upskilling gaps as a top barrier to AI scaling.

For the actuarial profession specifically, the junior role compression creates a pipeline question. Entry-level actuarial positions have traditionally served as the proving ground where exam candidates gain practical experience. If AI tools handle an increasing share of data preparation, preliminary analysis, and routine calculations that historically occupied junior staff, the volume of entry-level positions may contract. The Morgan Stanley projection of $9.3 billion in P&C expense ratio savings through AI assumes significant labor productivity gains, and junior actuarial roles are not exempt from that calculation.

At the same time, the governance and validation gaps documented across all three consulting reports suggest growing demand for experienced actuaries who can evaluate AI model outputs, certify compliance with ASOPs, and serve as the bridge between management’s AI ambitions and regulatory requirements. The profession’s challenge is managing both trends simultaneously: fewer traditional junior positions, but increasing demand for mid-career and senior practitioners with AI governance expertise.

What This Means for Actuarial Leaders

Synthesizing the three frameworks produces six actionable signals for actuaries and insurance executives.

Anchor your reserve assumptions deliberately. The 2.1-point combined ratio gap between Deloitte (99%) and AM Best (96.9%) is material. Neither forecast is inherently more credible; they reflect different assumptions about catastrophe loads, social inflation trajectories, and rate adequacy in a softening market. Reserving actuaries should document which external benchmarks inform their selected picks and stress-test against both endpoints. The three commercial lines already above 100 (commercial auto, medical professional, products liability) warrant line-specific attention regardless of which industry-level forecast you favor.

Bridge the 65-point governance gap in your own organization. If 90% of executives at your carrier agree that AI transformation is urgent but your organization has not taken tangible steps to retrain staff or build governance infrastructure, you are in the majority. The consulting firms agree that the binding constraint is organizational readiness, not capital. Actuaries are well positioned to lead governance framework development because ASOP No. 56 already provides the foundational requirements. The gap is in translating those principles into specific, documented procedures for AI model validation, monitoring, and disclosure.

Define which AI applications require actuarial governance and which do not. Oliver Wyman’s “hyperspeed” framework is not inherently incompatible with actuarial standards if the scope of actuarial oversight is clearly delineated. AI tools that process documents, automate customer interactions, or streamline operations can move through lighter governance. AI systems that influence pricing indications, reserve selections, or capital model calibrations must move through the full ASOP No. 56 validation process. Drawing that line explicitly prevents both regulatory risk and organizational friction.

Factor private capital dynamics into your ALM and reserve credit analysis. With reserves ceded to sidecars at $55 billion and private placements at 21.1% of insurance AUM, the asset side of the balance sheet is shifting faster than many actuarial models reflect. Carriers measuring AI ROI against expense ratio targets need to account for the different investment return assumptions that private credit portfolios introduce.

Prepare for M&A-driven actuarial demand. At $104 billion in insurance deal volume and growing, M&A creates recurring demand for loss reserve adequacy opinions, embedded value calculations, integration planning for data and model platforms, and ASOP-compliant assumption bridging between acquirer and target methodologies. Actuaries with deal experience will find their skills increasingly valued.

Manage the talent pipeline proactively. The 43% of CEOs reducing junior roles represents a structural shift, not a cyclical one. Actuarial departments that allow junior headcount to decline without creating alternative entry pathways (rotational programs, AI-augmented analyst roles, embedded data science positions) risk hollowing out the mid-career talent pool that governance and validation work depends on. The profession needs fewer people doing manual data manipulation and more people doing model evaluation and regulatory translation, but the transition requires deliberate design rather than passive attrition.

The consulting firm consensus on AI as the top insurance priority is real and well-documented. The disagreements on timing, magnitude, and implementation approach are where the actionable intelligence lives. For actuaries navigating 2026 strategic planning and budget cycles, the combined ratio divergence, the governance gap data, and the workforce compression signals deserve explicit inclusion in reserve assumptions, staffing plans, and governance roadmaps.