Evaluating the AI Threat to India’s IT Model

Image Source: Getty Images

Global Capability Centres (GCCs), expanding across the Indian subcontinent, are offshore units established by multinational companies (MNCs) to handle tasks requiring specialised capabilities. These primarily include functions such as Engineering, Research and Development (ER&D) and Business Process Management (BPM). Initially set up as cost-efficient subsidiaries, they have now evolved into strategic captive centres for their parent companies. Employing roughly 2.36 million professionals and generating around US$98.4 billion in annual revenue, GCCs represent a cornerstone of India’s services-led growth model. India is now home to around 2117 GCCs as of 2026, a number that is rapidly increasing every year.

GCCs occupy a strategically disproportionate position in the Indian economy relative to their headcount. They anchor high-skill urban employment, generate stable foreign-exchange inflows that buffer the current account, and serve as a critical rung on the skill ladder, absorbing engineering and management graduates whose alternatives in domestic industry are limited. Through agglomeration effects, they sustain demand for ancillary services in real estate, hospitality, and consumer goods across tier-1 and, increasingly, tier-2 cities, while each new GCC announcement reinforces India’s positioning as a preferred destination for delivering high-value services to global capital. Artificial intelligence (AI) has emerged as an imminent threat to the demand for GCC services, which could directly affect the employment intensity of this sector.

Table 1: GCC Penetration over the period 2021-26

Source: GCC Landscape in India 2026

An employment slowdown here is not analytically equivalent to a slowdown in any other sector. The Indian growth model is services-led and skill-intensive precisely because manufacturing has not absorbed the demographic dividend at the pace once anticipated. If the GCC sector enters a phase of revenue growth sans employment growth, the country loses one of the few channels currently capable of converting tertiary education into formal, high-productivity jobs at scale. A muted hiring response also dampens entry-level wage growth, weakens the talent pipeline that sustains the sector itself, and transmits negative demand shocks back into the urban consumption economy.

If the GCC sector enters a phase of revenue growth sans employment growth, the country loses one of the few channels currently capable of converting tertiary education into formal, high-productivity jobs at scale.

The broader academic consensus is that augmentation rather than displacement is the likely outcome of AI adoption in the workplace (Figure 1). Routine-based tasks would be more prone to automation, while tasks that involve complex thought and decision-making would require cooperation between human judgment and generative AI (GenAI) agents. Even with a cooperative effort to amalgamate the two, residual displacement is inevitable. However, it is equally important to focus not just on the displacement of existing employees, but also on the eventual slowdown in hiring as augmentation enhances existing employees, creating labour market imbalances.

Figure 1: Indian AI Talent Growth

Source: GCC Landscape in India 2026

This displacement pressure, however, coexists with a parallel demand-side shift that complicates a simple substitution narrative. India is not merely deploying AI within its GCC ecosystem but is actively building it. India is identified as the second-largest employer of enterprise AI talent globally, with over 250,000 AI/ML professionals across GCCs accounting for approximately 28 percent of global GCC AI talent. India also leads all GCC markets in AI hiring intensity. This bifurcation, with routine task compression on one side and high-skill AI/ML demand on the other, is precisely what makes the intensive margin the critical analytical lens: aggregate employment figures will increasingly obscure a hollowing out at the middle even as headline talent numbers remain robust.

Tasks at Risk

This hiring slowdown will be felt most acutely in ER&D and BPM, where routine, replicable task composition makes these functions most amenable to AI augmentation. Across regions, ER&D-led  GCCs account for 60 percent of Americas-origin GCCs, 43 percent of EMEA-origin GCCs, and 48 percent of APAC-origin GCCs by India headcount, while BPM-led GCCs represent 19 percent, 31 percent, and 28 percent respectively, making these two functions the most exposed to displacement pressure. Considering a more consolidated, value-driven projection baseline of 2.5 to 3.4 million employees in 2030, the scale of AI-driven workforce compression warrants significant attention from the MNCs owning these GCCs, as well as the government. The very concentration of risk in these two functions, combined with the scale of the projected workforce, makes targeted intervention both urgent and analytically complex.

The exposure of these two GCC functions to AI augmentation is well documented as routine, biased technological change, which is identified as the dominant driver of job polarisation across advanced economies. This reduces demand for middling occupations relative to higher- and lower-skilled ones, leading to cost reductions for functions like BPM from automation. This lowers their market prices and can increase demand for these services, expanding BPM output while simultaneously reducing BPM employment. Revenue and output growth improve, but employment begins to wane.

Displacement at the Margin

As AI capabilities advance from traditional machine learning to agentic AI, displacement occurs along both extensive and intensive margins. The extensive margin involves unemployment generated by layoffs and is therefore visible and accountable. On the other hand, intensive margins are difficult to track. Reductions in work hours and steady productivity gains among existing employees, while new hiring slows, cannot be tracked as closely as extensive margins. In these cases, intervening appears unnecessary precisely when it is most needed, as the effects remain invisible against the more tangible signal of rising unemployment.

As intensive-margin displacement increases, conventional policy action fails to account for it due to its inconspicuous nature. The conventional economic metrics, such as increasing unemployment, are easier to respond to swiftly, while indirect effects through productivity gains and reduced work hours are not tracked through conventional labour market metrics. Data from the Periodic Labour Force Survey and GDP contribution metrics, which track employment levels and output, are not designed to capture work hour reductions or hiring slowdowns at the function level. The dilemma with this problem is that planning to implement a policy at a stage of economic success may face heavy resistance from stakeholders, investors, and management. Meanwhile, waiting for productivity losses to show and then acting may exacerbate the problem till it is too late to intervene.

Implications for the GCC Sector

GCCs are currently responding to AI adoption primarily through workforce reskilling and upskilling initiatives, emphasising AI/ML, data engineering and advanced analytics. However, neither reductions in work hours nor hiring slowdowns are tracked as risk indicators. Attrition has, in fact, declined from 13 percent in 2023 to 9 percent in 2025, masking the intensive margin effects accumulating beneath the surface. The primary workforce challenges at this juncture are attracting niche digital talent, rising people costs, and retention challenges; effectively sidelining the intensive-margin indicators (work-hour reductions, productivity absorption, and hiring slowdowns) that could have signalled displacement before it became pronounced.

Existing policy instruments, calibrated to unemployment rather than to the rate of augmentation, will not detect the problem until the damage is already done.

Assuming that current conditions persist as India enters 2030, the GCC sector may feature a leaner workforce with highly productive AI augmentation. The displaced population will seek employment in ER&D and BPM functions with no demand, slowly inflating unemployment as demand for these roles goes unmet in a sector that has quietly stopped growing its headcount. The market will not correct itself due to a lack of distress signals. Existing policy instruments, calibrated to unemployment rather than to the rate of augmentation, will not detect the problem until the damage is already done.

Toward a Calibrated Policy Response

Closing this blind spot will require parallel action on measurement and intervention at both the government and corporate levels. On measurement, India’s labour market statistics need to evolve beyond unemployment and headline payroll figures. The PLFS and EPFO data ecosystems should be augmented with function-level reporting from GCCs that capture hours worked, hires-to-headcount ratios, and revenue per employee at the unit level. A voluntary GCC observatory, based on a public-private partnership (PPP) model, could publish anonymised quarterly indicators on AI penetration, task automation, and hiring intensity, providing the early-warning signal that conventional metrics miss.

The objective is not to slow AI adoption, but to ensure that India’s GCC sector remains a high-employment engine through the transition rather than a high-revenue one quietly shedding its labour-market role.  

For instruments, fiscal and regulatory levers should be calibrated for augmentation rather than only for displacement. Special Economic Zone (SEZ) and Software Technology Parks of India (STPI) incentives, currently linked to revenue and exports, could be partially indexed to net workforce growth or apprenticeship intake, rewarding GCCs that scale headcount alongside output. Public investment in reskilling should pivot from generic AI literacy toward verified placement in adjacent high-skill functions, with funding tied to outcomes rather than enrolment.

Organisational policies inside GCCs themselves are equally critical, since intensive-margin shifts are first visible to management. Parent MNCs should be encouraged, through disclosure norms or stewardship codes, to publish function-level AI adoption metrics alongside their workforce data. Internal job markets, redeployment guarantees, and minimum-hire commitments tied to productivity gains can institutionalise the principle that augmentation should expand the pie rather than shrink the workforce. The objective is not to slow AI adoption, but to ensure that India’s GCC sector remains a high-employment engine through the transition rather than a high-revenue one quietly shedding its labour-market role.  

---

Arya Roy Bardhan is a Junior Fellow with the Centre for New Economic Diplomacy at the Observer Research Foundation.

Siddharth Joshi is a Research Intern with the Centre for New Economic Diplomacy at the Observer Research Foundation.

Disclaimer: Claude 4.6 has been used for data visualisation in this piece. 

The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.