EXL Q1 2026: AI Revenue Hits 60% and Reshapes the Insurance Vendor Model

From tracking quarterly filings across the top insurance AI vendors since 2024, the 60% revenue threshold marks a structural shift that BPO-era metrics no longer capture. EXL (NASDAQ: EXLS) reported $570.4 million in Q1 2026 revenue on April 29, with data and AI-led services now accounting for 60% of total company revenue, up 28% year over year. The insurance segment grew 12.6% to $193.9 million as carriers shifted from legacy digital operations engagements to higher-value AI workflow deployments. Management raised full-year 2026 guidance to $2.30 billion to $2.33 billion in revenue and $2.18 to $2.23 in adjusted diluted earnings per share.

Most coverage treats this as a standard beat-and-raise quarter. The deeper story is what happens to the vendor business model when AI becomes the majority revenue stream. EXL's numbers, set against Verisk, Guidewire, and the broader vendor ecosystem, reveal a structural transformation in how carriers procure and pay for analytical capabilities.

The Q1 2026 Financial Picture

EXL's Q1 2026 results came in well above consensus expectations. Total revenue of $570.4 million represented 13.8% year-over-year growth on a reported basis and 13.4% on a constant-currency basis. Sequential growth was 5.1%, a strong acceleration from the typical Q4-to-Q1 seasonal pattern.

Segment	Q1 2026 Revenue	Q1 2025 Revenue	YoY Growth	Gross Margin
Insurance	$193.9M	$172.0M	12.6%	37.7%
Healthcare & Life Sciences	$151.9M	$125.6M	21.0%	45.3%
Banking, Capital Markets & Diversified	$127.4M	$117.7M	8.1%	36.9%
International Growth Markets	$97.1M	$85.7M	10.9%	34.1%
Total	$570.4M	$501.0M	13.8%	38.9%

Adjusted operating margin expanded 40 basis points to 20.5%, while adjusted diluted EPS grew 20.2% to $0.58, up from $0.48 a year earlier. Adjusted EBITDA reached $127.9 million at a 22.4% margin. The company won 16 new clients in the quarter and reported headcount of roughly 67,000 across six continents, with revenue growing three percentage points faster than headcount, a signal of improving productivity metrics.

The guidance raise was meaningful. EXL now expects $2.30 billion to $2.33 billion in full-year 2026 revenue, up from prior guidance of $2.275 billion to $2.315 billion, reflecting 10% to 12% year-over-year growth. Adjusted EPS guidance rose to $2.18 to $2.23, up from $2.14 to $2.19, implying 12% to 14% growth over 2025.

Insurance Segment: From BPO Workhorse to AI Revenue Engine

Insurance remains EXL's largest segment at 34% of total revenue, and the Q1 performance reveals how the AI transition is reshaping this vertical specifically. The $193.9 million in insurance revenue grew 12.6% year over year and 4.4% sequentially, returning to consistent double-digit growth after a slower period in mid-2025. Insurance gross margins improved to 37.7% from 36.6% a year earlier, a 110-basis-point expansion that reflects the mix shift toward higher-margin AI engagements.

EXL's insurance operations span claims processing, underwriting support, subrogation (where EXL operates the Subrosource platform, one of the largest subrogation providers in the United States), regulatory reporting, and customer operations. The company was named a leader in the Everest Group Customer Experience Services in Insurance Operations Peak Matrix Assessment 2025, a recognition that reflects the breadth of its carrier client relationships.

The segment-level story mirrors the company-wide pattern: carriers are migrating existing EXL engagements from labor-intensive digital operations to AI-augmented workflows. This is visible in the financial data. Reported digital operations revenue declined 2% year over year across the company, while data and AI-led revenue grew 28%. Total operations still grew 10%, meaning the AI migration is additive rather than purely cannibalistic, at least for now. Carriers are not simply relabeling the same work; they are expanding scope as AI capabilities enable new use cases in claims triage, underwriting document processing, and regulatory compliance.

This transition has been accelerating since EXL launched its Insurance LLM in September 2024 in partnership with NVIDIA, using the NeMo framework and H100 GPU infrastructure. The LLM was trained on over a decade of domain-specific insurance data and, according to EXL's patent filings, uses an ensemble architecture with multiple expert models for claims processing, fraud detection, risk assessment, and customer support. The company holds 10 AI patents granted in a single year covering document extraction (Xtrakto.AI), entity recognition, knowledge graphs, the Insurance LLM itself, regulatory reporting automation, and property risk analytics.

What the 60% Threshold Actually Measures

EXL's claim that data and AI-led revenue now represents 60% of total company revenue requires careful parsing. The company does not separately break out AI-led revenue as a GAAP line item. Instead, it uses an internal classification system that categorizes engagements based on the delivery model: work delivered through AI-augmented workflows counts as "data and AI-led," while traditional process execution counts as "digital operations."

At 60% of $570.4 million, AI-led revenue was approximately $342 million in Q1 2026, up from roughly $268 million a year earlier. That 28% growth rate significantly outpaced overall company growth of 13.8%, confirming that the AI category is pulling revenue away from the digital operations category while also generating net new demand.

CEO Rohit Kapoor provided context on the earnings call: "The commercial model is changing much more towards a fixed-fee and milestone-based payment and an outcome-based model." This is a significant shift from the traditional BPO pricing structure of per-FTE or per-transaction fees. Under fixed-fee and outcome-based arrangements, the vendor's margin improvement from AI-driven productivity does not automatically flow back to the client as a rate reduction. Instead, the vendor captures the efficiency gains as margin expansion while the client pays for outcomes rather than inputs.

The 60% figure also reflects a definitional choice. EXL classifies entire engagements as AI-led when AI is the primary delivery mechanism, even if human operators remain involved. This is consistent with how the company describes its operating model: a "human-in-the-loop AI-led" approach rather than a fully automated one. The distinction matters because it means the 60% figure captures the commercial and contractual shift toward AI-centered engagements, not purely the revenue attributable to AI software in isolation.

Vendor Ecosystem Comparison: Who Is Winning the AI Revenue Race?

EXL's 60% AI revenue share becomes more striking when placed alongside the other major technology vendors serving insurance carriers. Each occupies a different position in the value chain, and their Q1 2026 results reveal divergent trajectories.

Verisk (VRSK) reported Q1 2026 revenue of $783 million, up 4% year over year on a GAAP basis and 4.7% on an organic constant-currency basis. Subscription revenues grew 7% and now represent 84% of total revenue. Verisk released seven new client-facing AI modules in Q1 and targets 25 for the full year through its "Core Lines Reimagine" initiative. The company does not disclose AI-specific revenue, but its pure subscription model means essentially all revenue benefits from embedded analytics and machine learning. Verisk's challenge is different from EXL's: it needs to layer generative AI and agentic capabilities on top of an already-analytics-heavy revenue base, making the incremental AI contribution harder to isolate. Management described the quarter as a "trough" period, citing low catastrophe activity, tough prior-year comparisons, and a federal government contract work stoppage.

Guidewire (GWRE) reported Q1 2026 revenue of $356.6 million, up 22% year over year, driven by its cloud platform migration. Annual recurring revenue guidance was raised to $1.22 billion to $1.23 billion. Guidewire signed eight cloud deals in Q1, including major wins at The Hartford and Sompo. The company's AI strategy centers on its ProNavigator acquisition and PricingCenter product, which embed AI into the core policy administration and claims management workflows that carriers already run on Guidewire's platform. Guidewire's AI revenue is not separately disclosed, but the cloud ARR acceleration suggests that AI-enhanced modules are a meaningful contributor to deal velocity.

Sapiens (SPNS) generated approximately $595 million in 2025 revenue with annual recurring revenue exceeding 60% of total sales and recurring revenue at 79%. The company is in the middle of a SaaS transition that creates a 2% to 3% revenue headwind as perpetual license revenue converts to subscription. Sapiens launched AI-powered CoreSuite updates in 2025 and has deepened alliances with Microsoft Azure and AWS for co-selling, but it remains primarily a core systems vendor with AI as an enhancement layer rather than the revenue driver.

Vendor	Q1 2026 Revenue	YoY Growth	AI Revenue Share	Primary Model
EXL	$570M	13.8%	60% (disclosed)	AI-augmented services
Verisk	$783M	4.0%	~84% subscription (analytics-embedded)	Data/analytics platform
Guidewire	$357M	22.0%	Not disclosed	Core systems + cloud
Sapiens	~$149M (est.)	~8%	Not disclosed	Core systems + SaaS migration

The comparison reveals that EXL is the only major insurance vendor explicitly quantifying and growing AI-specific revenue as a disclosed metric. Verisk's subscription model is inherently analytics-driven, so the concept of "AI revenue share" is less meaningful. Guidewire and Sapiens treat AI as a feature within core platform subscriptions rather than a separately measured revenue stream. EXL's decision to measure and report the AI mix shift gives investors and carrier procurement teams a clearer signal of how fast the business model is actually changing.

The Commercial Model Transformation

The shift from per-FTE pricing to fixed-fee and outcome-based commercial models is the most structurally important development in EXL's Q1 results, and it has direct implications for how carriers budget for and evaluate vendor AI services.

Under the traditional BPO model, carriers paid EXL based on headcount deployed or transactions processed. This created a straightforward cost-per-unit structure where the carrier captured productivity improvements as the vendor became more efficient. If EXL could process 20% more claims per employee through automation, the carrier would expect to pay less per claim in the next contract cycle.

Under the new model that Kapoor described, the pricing mechanism shifts. Fixed-fee engagements lock in total cost regardless of how many people or how much AI the vendor uses to deliver the work. Milestone-based pricing ties payments to specific deliverables rather than ongoing effort. Outcome-based pricing links fees to measurable results such as claims cycle time reduction, underwriting accuracy improvement, or regulatory compliance metrics.

For carriers, this changes the procurement calculus in several ways. First, it makes vendor AI capability a competitive differentiator rather than a cost-pass-through. A vendor that can achieve better outcomes through superior AI earns better margins without necessarily charging the carrier more. Second, it shifts the productivity risk from the carrier to the vendor: if EXL's AI works well, EXL captures the margin; if it underperforms, EXL absorbs the cost of supplementing with human labor. Third, it creates natural switching costs, because migrating from an outcome-based vendor engagement to an in-house capability requires the carrier to rebuild not just the technology but the entire performance measurement framework.

The margin data supports this interpretation. EXL's adjusted operating margin expanded to 20.5% despite revenue growing faster than headcount by only three percentage points. Under a pure FTE model, revenue-per-employee improvements would be competed away. Under a fixed-fee model, those productivity gains accrue to the vendor.

Carrier Build-vs.-Buy Implications

EXL's results arrive at a moment when carrier AI strategy is fragmenting into three distinct approaches, and the economics of each are becoming clearer.

Carriers building proprietary AI systems, such as AIG with its three-layer stack of Anthropic Claude, Palantir Foundry, and internal tools, or Travelers with its $1.5 billion annual technology budget, are making large upfront investments with uncertain but potentially transformative payoffs. AIG has disclosed a 55% reduction in time-to-quote for E&S submissions. Travelers has deployed 10,000 Anthropic Claude assistants across its engineering organization and consolidated four claims call centers to two. These are real productivity gains, but they require sustained capital allocation and specialized AI talent.

Carriers licensing platform vendors like Guidewire and Verisk get AI as an embedded feature of their core systems or analytics subscriptions. The Hartford and Sompo signed Guidewire cloud deals in Q1 2026. Verisk's MCP connectors now embed ISO Indications and XactRestore analytics directly into Anthropic's Claude. This model requires less internal AI expertise but gives the carrier less control over the analytical capabilities and their evolution.

Carriers outsourcing to services companies like EXL get operational execution and AI capability bundled together. The carrier does not need to hire AI engineers, manage model training pipelines, or build inference infrastructure. EXL handles all of that within the engagement, deploying its patented Insurance LLM, Xtrakto.AI document extraction, and knowledge graph systems on the carrier's data. The trade-off is dependency: the analytical methods themselves are proprietary to EXL, protected by 10 patents, and the carrier has limited visibility into or control over how the AI evolves.

EXL's 60% AI revenue mix changes the calculus for the outsourcing option. When a services vendor's revenue was primarily labor-driven, the carrier could credibly argue that it could replicate the capability internally by hiring equivalent staff. When 60% of the vendor's revenue comes from AI-augmented delivery with patented methods, outcome-based pricing, and a domain-specific LLM trained on a decade of insurance data, the replication argument weakens significantly. The carrier is no longer buying labor arbitrage; it is buying an AI capability stack that would cost tens of millions and years to build internally.

Celent's third annual GenAI in Insurance survey found that 48% of insurers globally are now in production with generative AI, with adoption expected to reach the "late majority" phase in 2026. Meanwhile, 22% of surveyed insurers plan to have an agentic AI solution in place by year-end 2026. For carriers that fall into the majority adopting AI through vendor relationships rather than building internally, EXL's Q1 results suggest the vendor option is becoming more capable, more expensive to replicate, and more deeply embedded in carrier workflows with each passing quarter.

The Revenue-per-Employee Productivity Signal

One metric from EXL's Q1 results deserves separate attention: the three-percentage-point spread between revenue growth (13.8%) and headcount growth (approximately 11%). This gap is small by software company standards but significant for a services business with 67,000 employees.

Annualized Q1 revenue per employee works out to roughly $34,000, up from approximately $30,000 two years earlier. For a services company that historically scaled revenue by adding headcount, consistently growing revenue faster than people signals that AI is enabling genuine productivity improvement, not just revenue reclassification.

The productivity gains are not evenly distributed across segments. Healthcare and Life Sciences, with its 45.3% gross margin, generates significantly more revenue per unit of effort than Insurance at 37.7% or International Growth Markets at 34.1%. As AI adoption deepens across all segments, convergence toward higher-margin delivery models could drive meaningful operating leverage over the next several years.

For carrier procurement teams evaluating EXL's pricing, this metric matters because it reveals the vendor's margin structure. If EXL is growing revenue 14% while adding 11% more people, the incremental revenue per new hire is higher than the base, which means either the new work is inherently higher-value or AI is making existing workers more productive. Either way, it gives EXL room to absorb fixed-fee pricing without sacrificing margins.

The Insurance BPO Market Context

EXL's transformation sits within a broader insurance BPO market valued at $64.3 billion in 2025 and estimated at $68.4 billion in 2026, according to Mordor Intelligence. The overall market is growing at roughly 6% annually, driven by carrier demand for cost efficiency and digital workflow modernization.

EXL's insurance segment, at $194 million quarterly and roughly $776 million annualized, represents approximately 1.1% of this market. That modest share belies EXL's influence: because its engagements are concentrated among large and mid-market carriers where AI adoption decisions are made, EXL's commercial model innovations (outcome-based pricing, AI-augmented delivery) set precedents that ripple through the broader vendor ecosystem. Competitors including Cognizant, WNS, and Genpact are all pursuing similar AI-led transformations, but none have achieved or disclosed an AI revenue mix as high as EXL's 60%.

The market dynamic creates a feedback loop. As more carriers experience AI-augmented service delivery through vendors like EXL, their expectations for vendor capabilities rise, making it harder for smaller or less AI-advanced vendors to compete for renewal business. This consolidation pressure may accelerate M&A activity in the insurance services space, a trend already visible in EXL's own acquisition strategy and in the broader pattern of technology-driven vendor consolidation.

Why This Matters for Actuaries

EXL's Q1 results carry several direct implications for actuarial practice.

Vendor risk assessment is becoming an actuarial concern. When carriers outsource claims processing, underwriting support, and regulatory reporting to a single vendor with proprietary AI methods, the concentration risk is not just operational; it touches reserving adequacy (if the vendor's claims triage AI introduces systematic bias), pricing accuracy (if underwriting support models drift), and regulatory compliance (if the vendor's automation produces results that do not align with ASOP requirements). Actuaries involved in enterprise risk management and appointed actuary reporting should be evaluating third-party AI vendor dependencies as a component of operational risk.

Outcome-based vendor pricing changes expense ratio dynamics. Under traditional BPO arrangements, the carrier could forecast vendor costs with high precision because pricing was input-based. Under outcome-based arrangements, vendor costs may vary with performance, creating a new source of expense ratio variability that loss ratio and expense analyses need to incorporate. For carriers reporting under LDTI or IFRS 17, the treatment of variable vendor costs within the contractual service margin or liability for remaining coverage calculations may require additional actuarial judgment.

The build-vs.-buy analysis now requires AI capability benchmarking. When actuarial leadership participates in technology strategy discussions, the relevant comparison is no longer "Can we hire enough people to do this internally?" The question is whether the carrier can build, train, and maintain an AI capability stack that matches what EXL deploys through its Insurance LLM, 10 patented methods, and decade of training data. For most mid-market carriers, the answer is increasingly no, which makes vendor selection a strategic decision with multi-year actuarial implications.

Model validation scope is expanding. If a carrier uses EXL's AI-augmented claims processing, the outputs of that system feed directly into reserving data. Actuaries relying on ASOP No. 23 (Data Quality) and ASOP No. 56 (Modeling) need to determine how much validation they require of vendor AI models that influence their own actuarial work product. The NAIC's 12-state AI Evaluation Tool pilot, now underway, will likely establish expectations for how carriers document and validate third-party AI vendor capabilities.