ExlService Holdings Inc.

Revenue comes from a small number of major client-facing industry groups, built mainly around insurance and healthcare-related work alongside a broader banking and international client base, priced through several different arrangements within the same business: some tied to hours or staff assigned, some to transaction volume or business outcomes, and some to recurring subscription or software fees. No single customer supplies a dominant share of revenue, though its largest clients together account for a meaningful minority of it.

This system scales chiefly by adding delivery capacity, staff and workstations concentrated in a handful of low-cost offshore locations, rather than by building one large fixed asset or a network that becomes more valuable as more participants join it. Several of its return-on-capital and free-cash-flow ratios sit in the upper part of the range for the peer group CompanyGraph maps it against, it has returned a significant amount of capital to shareholders through buybacks over time, and it has recorded positive net income in every one of the past several fiscal years, a combination that reads more like a mature, cash-generative delivery model than an early, fast-scaling one.

The system depends on a workforce concentrated mainly in India and the Philippines, on third-party vendors supplying software, cloud infrastructure, data communications and licensed datasets, and on keeping pace with fast-moving AI technology, each named as a specific risk in its own filings. Its largest cost by far is the pay and benefits of its own people, so what it depends on to operate is chiefly labor rather than physical materials.

Large and mid-sized enterprises across a handful of named industries, including insurance, healthcare, banking and capital markets, retail, media and energy, rely on it to run outsourced operations and analytics on their behalf; no single customer supplies a dominant share of revenue, but a small group of its largest clients together account for a meaningful minority of it. In its healthcare business specifically, it says it sits between the party paying for care, the party providing it and the person receiving it, coordinating activity among all three.

CompanyGraph places this company within a fairly large group of similarly structured, expertise-driven service businesses, so operating this kind of labor-leveraged, data-and-analytics model is not by itself unusual. Within that shared shape, the company's own account points to long-standing relationships with a large share of major insurers and banks, process knowledge built up inside a small number of regulated industries, and a combination of human expertise with embedded AI as what it says sets it apart, though these are the company's own claims about itself and CompanyGraph has no independent way to judge whether a rival could reproduce them.

Its own filings describe two different contract shapes side by side: shorter, project-based engagements, and longer digital-operations contracts that run for several years and require extended notice, or an early-termination fee, to exit. Even so, the filings say most client contracts can be ended without cause and do not obligate the client to any minimum volume of work, so the extended length of the digital-operations contracts is the main disclosed source of friction in leaving, rather than a volume commitment or a large committed order backlog, neither of which the company discloses.

Its own risk disclosures list attracting, training and retaining enough skilled staff and leaders among the pressures named first, and its most significant cost by a wide margin is the pay and benefits of its own people rather than materials or plant, both consistent with a business that scales by leveraging expert talent rather than by running a large fixed asset. It also discloses a specific, finite amount of physical seating capacity across its delivery locations that can only be expanded by building or leasing more space, so growth beyond that footprint requires deliberate capital spending ahead of demand.

Its own filings describe a client base where a relatively small group of large, mostly domestic clients supplies a meaningful minority of revenue, most contracts can be ended without cause and do not obligate any client to a set volume of work, and operations and the people who deliver the work are concentrated mainly in India and the Philippines, so a disruption specific to either country would affect a large share of delivery capacity at once. It also names dependence on outside vendors for software, cloud services and data as a source of possible delay, alongside keeping pace with AI and the legal and reputational risk AI itself introduces, among the pressures it lists first.

Its own filings name a wide range of privacy, anti-bribery and data-protection regimes across the countries where it operates and serves clients, disclose ongoing Indian tax disputes under appeal, and flag exposure to trade restrictions and tariffs that could affect clients' willingness to keep using an offshore delivery workforce. They also name the Indian rupee, Philippine peso, British pound and South African rand as significant currency exposures against the US dollar, hedging most though not all of that exposure, and list keeping pace with fast-moving AI technology, the legal and reputational risk AI itself introduces, and competition that includes its own clients building the same capabilities inside their organizations, among the pressures named first.