India’s Services Model is Splitting. Which Side are You Building For?

News

1. Anthropic Forms AI Services Firm With Wall Street Backers

2. Pixxel, Sarvam Plan India’s First Orbital Data Center Satellite

3. GalaxEye Launches India’s Largest Private Earth Observation Satellite

4. SoftBank Eyes $100 Billion Roze Listing for AI Data Centers

5. Infosys Lands Truist GCC Mandate

6. AI Dispatch | 24–30 April 2026

India’s technology services industry was built on a proposition that held for three decades: recruit graduates at scale, train them in structured processes, and deliver that execution capacity to the world at a margin no Western competitor could match. The proposition worked. It built a sector that employs close to 8 million people, generates hundreds of billions in export revenue, and establishes India as the world’s back-office for enterprise technology.

That proposition is now being disaggregated by the same technology India helped build. But the disruption is not what the displacement narrative suggests. AI is not collapsing the sector. It is splitting it into two distinct, diverging trajectories that will determine which organizations lead India’s services economy into the next decade and which are left optimizing a model the market has moved past.

Entry-level demand in India’s IT services sector has fallen 20–25%, according to a 2025 analysis by EY, as AI absorbs routine coding, testing, documentation, and support work that once employed hundreds of thousands of annual hires. NASSCOM data for FY26 shows workforce growth slowing sharply even as industry revenue expanded—a decoupling of headcount and output that would have been difficult to project five years ago.

The opposite trajectory is equally visible. India recorded close to 100% year-on-year growth in prompt-engineering talent in 2025 and a 30% hiring rate for AI engineering roles, placing it ahead of several advanced economies, per LinkedIn Economic Graph data. Roles commanding ₹58–60 lakh per annum—AI architects, ML-Ops engineers, prompt specialists, and AI risk auditors—are growing at 18% annually. Legacy process roles at ₹12 lakh are contracting.

The distance between those two lines is not a technology problem. It is a talent architecture problem. Organizations that treat it as anything less are making a strategic error whose consequences will take a decade to unwind.

THE BIFURCATION IS STRUCTURAL, NOT CYCLICAL

The conventional framing of AI’s impact on Indian technology services treats the shift as displacement: automation takes process jobs, workers suffer, and policy must respond. This framing is incomplete in a way that matters for how organizations plan.

The more precise description is value migration”. As AI systems move from generating recommendations to executing decisions within defined parameters—routing customer triage, adjusting credit limits, reallocating inventory without human sign-off at each step—the roles that survive are not those that produce outputs. They are the roles that govern, challenge, and own accountability for what autonomous systems produce.

Adnan Masood, Chief AI Architect at UST, a global digital transformation services firm, maps the transition precisely. Entry-level roles in first-line technical support, basic software testing, routine code generation, first-draft documentation, and entry-level data analytics are being absorbed rapidly by AI models. “The middle is where the pain concentrates,” he observes: “the five-to-twelve-year professional whose value is institutional knowledge of a process that AI can now learn in a weekend. That cohort has to move up into judgment work or sideways into specialization—there is no holding pattern.”

NITI Aayog’s October 2025 report, developed with NASSCOM and the Boston Consulting Group, makes the range of outcomes explicit. In a worst-case scenario—passive organizational response—headcount in India’s tech services sector falls from approximately 7.5–8 million in 2023 to 6 million by 2031, while the customer experience sector contracts in parallel. The upside scenario—active national coordination and enterprise investment—generates up to 4 million new AI-enabled roles within five years. The distance between those outcomes is determined not by technology but by whether Indian enterprises actively reshape their talent architecture or wait for the market to do it for them.

Kalyan Kumar, Chief Product Officer at HCLSoftware, identifies the management implication: “AI is not simply eliminating work; it is fundamentally reshaping how services are created and delivered. AI, in that sense, stands to elevate the service economy rather than diminish it.” Elimination is a one-time event. Rebuilding the work model is a continuous management obligation.

The Services Bifurcation

THE RESKILLING GAP IS COGNITIVE, NOT TECHNICAL—AND THAT CHANGES WHAT ENTERPRISES MUST DO

India has approximately 400,000 AI professionals. Demand exceeds supply by 53%. There is one qualified engineer available for every ten generative AI roles, according to TeamLease Digital. NITI Aayog estimates that AI talent supply meets only about half of current demand, while simultaneously, more than 60% of formal jobs are susceptible to automation by 2030.

The structural mismatch worsens over time without active intervention: roles are being automated faster than the displaced workforce can be reskilled into the roles that AI creates. But the nature of the gap matters as much as its size.

Neeti Sharma, Chief Executive of TeamLease Digital, identifies a three-part capability deficit: applied AI fluency, decision-making under uncertainty, and governance—the ability not merely to use AI tools, but to supervise them, challenge their outputs, and take accountability for what they produce. This is not a skills gap in the conventional sense. There is a gap between execution and oversight capabilities, and it is widening as AI systems assume greater operational authority.

G.S. Bhalla, Co-Founder and Chief Executive of Cosentus, an AI-focused enterprise technology services firm, frames the organizational imperative: “The real gap is not technical but cognitive. As systems get smarter, the workforce must move from execution to oversight. If done well, India will not just deliver services to the world—it will shape how intelligent systems are deployed within it.”

Srinivas Reddy, Senior Vice-President and Head of EPAM India, describes the specific hiring profile his organization is actively building: AI architects, full-stack agentic engineers, AI governance and verification leads, data and platform engineers focused on AI readiness, and product-minded engineers who shape client outcomes rather than code features. A full-stack agentic engineer—someone who can design specifications, orchestrate AI agents, integrate enterprise systems, and apply domain context across all of it—does not exist at scale in the current market.

EPAM India’s response is instructive: “This is not a talent acquisition strategy,” Reddy says. “It is a talent manufacturing strategy.” Internal academies, structured upskilling, selective hiring, and platform partnerships are the mechanisms, because the market cannot yet supply what the next phase of services delivery requires.

The governance and accountability dimension is not confined to knowledge workers. Frontline workers in retail, healthcare, logistics, manufacturing, and finance—who constitute 70–80% of the global workforce, according to Nitin Chandel, Group Vice-President and India Country Manager at UKG—are equally subject to AI-driven restructuring. Across these sectors, the value of human judgment concentrates not in doing what AI does more slowly, but in defining the boundaries of AI involvement, monitoring for performance drift, and owning accountability when outcomes fail. The governance question runs through the entire operational workforce, not only the technology team.

“Cost arbitrage is what got us here; cognitive arbitrage is what gets us to the next decade. The firms that win will stop measuring themselves in headcount and start measuring themselves in value delivered.”

— Adnan Masood, Chief AI Architect, UST

INDIA’S COMPETITIVE MOAT HAS MOVED—NOT DISAPPEARED

India’s original services moat was built on labor arbitrage: the ability to deliver structured process execution at a cost differential no advanced economy could match at scale. The moat was real. It also attracted investment in AI tooling that is now automating the processes that created it.

Masood of UST reframes what the moat actually was: “India’s moat in services was never labor costs, but the ability to deliver complex, accountable work at a scale and consistency no one else could match. AI doesn’t erode that moat; it relocates it.”

The relocation is from process execution to cognition arbitrage—what Vivek Raj, Founder and Chief Executive of Panama Hydro X, describes as the migration of value from process and SLA compliance to outcome control. “The winner is not who has jobs, but who can upgrade skill velocity faster than AI upgrades its capability.”

Over 1,700 Global Capability Centers now operate in India, and the fastest-growing roles inside them—product managers, AI and ML engineers, data scientists, cyber-risk leads—are not process-execution roles. They are the roles that direct, verify, and take accountability for AI-mediated decisions. The GCC footprint reflects exactly where the moat has relocated: toward the judgment-intensive, accountability-bearing work that clients will not trust to automation.

Joseph Anantharaju, Co-Chairman and Chief Executive of Happiest Minds Technologies, identifies the specific new archetypes that enterprises need to plan around: Human-in-the-Loop Architects who design the points at which human judgment must override automated decisions; Agent Orchestrators; AI Operations Engineers; Specification Engineers; and AI Governance Specialists who can design, code, test, and secure AI systems. The gaps in building these roles are not only technical. “These gaps are contextual,” Anantharaju notes, “observed in data interpretation, model oversight, and responsible AI governance.”

For regulated sectors—banking, financial services, healthcare, manufacturing—the governance dimension is acute and timeline-specific. Regulated environments are increasing the premium on human judgment around risk, ethics, and accountability. Functional leaders in these sectors must build explicit human-in-the-loop accountability structures before autonomous AI, not after the first compliance failure. The sequence matters.

Raj identifies the constraint that could foreclose the opportunity: India lacks scaled reskilling pathways, especially outside Tier 1 cities, and produces too few interdisciplinary operators who can bridge domain expertise, AI fluency, and governance accountability. The talent density that makes cognitive arbitrage possible in Bengaluru, Hyderabad, and Pune is not yet replicated at the scale the opportunity requires.

MANAGEMENT IMPLICATIONS

NITI Aayog’s roadmap makes the stakes clear: timely action could generate up to 4 million new jobs; inaction risks more than 1.5 million job losses in the technology sector by 2031. Those two outcomes are not separated by global market forces or the pace of AI development. They are separated by whether India’s enterprises make the organizational choices in measurement, reskilling, and governance that determine which side of the divide they occupy. The following implications apply within a two-year action horizon.

 

Role	Action Required — By FY2027
CEO	Answer two questions before the close of FY2027: which revenue streams depend on process execution that AI can replicate within three years, and what is the credible path to judgment-intensive work that clients will not trust to automation? Organizations that still measure productivity in headcount, utilization rates, and billable hours are optimizing for a model that AI is actively rendering obsolete. Redefine the firm’s value metric from hours delivered to decisions governed.
Board	Treat workforce transformation as a governance obligation, not an HR function. Three questions belong on the agenda in the next cycle: What is the organization’s current revenue exposure to roles AI will automate within three years? What is the reskilling investment relative to that exposure? Does the organization have a credible capability pipeline for judgment-at-scale work at the margin that the market will pay for? Commission a formal workforce transition audit.
CHRO	The reskilling gap will not close through market mechanisms alone. Build internal capability academies—as EPAM India has done—for the roles that do not yet exist in the external talent market: full-stack agentic engineers, AI governance leads, and Human-in-the-Loop architects. Formalize upskilling pathways for the five- to twelve-year professional cohort before that window closes. Track skill migration rate, not headcount, as the primary workforce health metric.
CIO / CDO	Before expanding autonomous AI deployment, build the accountability infrastructure that makes that deployment governable: defined human-in-the-loop decision points, model performance monitoring protocols, and named ownership for system failure. For regulated sectors, this infrastructure must precede deployment—not follow the first compliance incident.
Functional Leaders (CFO, COO, Sector Heads)	Identify which operational decisions are currently made by AI systems within your function that carry human accountability and which do not. Where accountability is absent or diffuse, that is a governance gap, not a technology gap. Assign it explicitly to a named role before the next audit cycle.

The bifurcation of India’s services economy is not a future risk. It is a current organizational sorting process. Enterprises that define clear boundaries for AI authority, invest in the human capability to govern autonomous systems, and build accountability structures that hold up under operational and regulatory scrutiny will determine what Indian services look like in the decade ahead.

Those who do not will find the market has been determined for them.

India built the world’s services economy on the proposition that accountable, complex work could be delivered at scale. That proposition is still available. The work it describes has changed. The organizations that recognize this—and act on it before FY2027—are not managing disruption. They are defining the next version of the moat.

Editor’s Note: MIT Sloan Management Review’s AI Research Forum will make its India debut in Bengaluru on 23 July, bringing together enterprise leaders, researchers, and practitioners to examine how autonomous AI is moving from experimentation to governed deployment at scale. To speak, partner, or attend, register here