EXL Reports 19.7% Growth in Q3, Data and AI Surge 36%

EXL, an American provider of data, analytics, and operations services, reported on Tuesday (27) revenue of $634.0 million for the third quarter of 2026, a 19.7% increase over $529.6 million from a year earlier. Adjusted earnings per share were $0.62, up from $0.48, a 28.5% rise. Under GAAP accounting standards, diluted earnings per share increased from $0.36 to $0.45.

The most crucial data for buyers and sellers of IT services lies in the revenue composition. According to the statement, EXL's segment referred to as "Data and AI-led," which accounts for about two-thirds of total revenue, grew by 36% on a reported basis and 31% organically. The company also raised its revenue projection for 2026 to the range of $2.430 billion to $2.440 billion, up from the previous estimate of $2.390 billion to $2.415 billion.

What Fueled the Quarter

"We entered 2026 with strong momentum that accelerated throughout the year," said Rohit Kapoor, chairman and CEO, in the announcement. He noted that all segments of the company grew over 15% on an organic basis. EXL primarily serves insurers, healthcare companies, banks, and retailers.

Part of the reported growth came from acquisitions. EXL completed its acquisition of iMerit, specializing in data preparation for model training, and established a Model Training Solutions practice with it. In an interview with Business Standard in June, Kapoor described iMerit as an entry point for working with foundational model labs. The difference between the reported 36% and the 31% organic growth in the data and AI business largely reflects this contribution.

The company reported 14 new clients in the quarter and introduced a new brand, "Go Beyond."

The Margin That Didn't Keep Pace with Revenue

The growth did not come with a gain in profitability. The adjusted operating margin was 19.3%, down from 19.4% in the third quarter of 2025. Under GAAP, the operating margin fell from 14.4% to 14.0%. An analysis published after the earnings report attributed part of the pressure to scaling costs and increased debt related to the iMerit acquisition.

This is the point that weakens the more optimistic narrative about the sector. If AI were rapidly replacing human labor within EXL itself, one would reasonably expect margin expansion, not stability. What the quarter shows is something different: the company is selling more data and AI projects, but is still incurring costs to build, integrate, and finance them.

The contrast comes from the sector itself. Infosys, while disclosing its numbers in recent days, stated that clients have begun to expect lower prices when a supplier uses AI, and that these savings are already embedded in multi-year contracts, according to a report from The Register. The skeptical reading is that revenue growth in AI coexists with deflation in hourly labor costs. The optimistic view is that EXL, having migrated early to data and analytics, captures new budgets before experiencing deflation from the old. Tuesday's report does not resolve the dispute, but shifts the question from revenue to margin.

Where the Results Resonates

In India, where EXL concentrates a significant portion of its delivery operations, the quarter underscores a divergence within the services industry: those selling data and AI to regulated sectors, such as insurance and healthcare, are growing nearly 20%, while Infosys reports that AI is slowing part of its traditional revenue. For the Noida and Pune centers, this means hiring driven by data profiles, not operator volume.

In the Philippines, an important base for the global BPO industry, the effect is the opposite. Operations for American insurers' customer service and back office are exactly the kind of work that "Data and AI-led" projects aim to automate. When a vendor's data and AI segment exceeds two-thirds of revenue, the volume of traditional operations ceases to be the growth engine.

In the United States, home to the main clients, insurers and healthcare operators are appearing as buyers willing to pay for AI applied to claims, underwriting, and billing. The conference call with analysts is scheduled for Wednesday (28), at 10 AM New York time, and is expected to bring the question that the announcement did not answer: how much of the growth depends on new contracts and how much on selling AI to the installed base at lower prices per unit of work.