Capability in the Age of AI: India’s GCCs and the Future of White-Collar Work

AI and the New Geography of Enterprise Work

Artificial Intelligence (AI) has become a general-purpose technology for enterprise operations. Besides reducing computation cost, AI applications lower the costs of prediction, classification, search, software generation, and content production across functions such as software engineering, customer support, finance, logistics, research, and design. Its labour market effects are therefore best understood not as the automation of entire jobs but as the reallocation of tasks between humans and machines.^[1]

The task-based framework developed by economists Daron Acemoglu and Pascual Restrepo provides the analytical anchor for this report. AI creates a displacement effect when machines substitute for labour in tasks that can be performed algorithmically. It also creates a reinstatement effect when new tasks emerge in supervision, interpretation, governance, design, and coordination. The employment outcome depends on the balance among these effects, the nature of the adopted technology, and the institutional capacity of workers, firms, and governments to adapt.^[2] The extent of new job creation is determined by availability of technology and pace of innovation, which would most likely result in a weaker reinstatement effect and net job loss during the early stages of AI adoption. Recent labour market projections should be read in that spirit.

The World Economic Forum’s Future of Jobs Report 2025 projects both substantial job creation (around 170 million) and displacement (around 92 million) by 2030, and also finds that a large share of core skills is expected to change. Such estimates are survey-based and subjective. However, they point to the central policy issue: can economies build the capacity for reskilling, occupational mobility, and new-task creation quickly enough to keep human labour productive in AI-augmented systems?^[3]

For multinational corporations, this shift from labour optimisation to AI integration, changes the logic of global production. Earlier waves of offshoring were shaped largely by labour cost arbitrage. The next wave is being increasingly guided by capability arbitrage: firms require locations that combine technical talent, digital infrastructure, managerial depth, governance capability, and the ability to operationalise AI safely across functions and jurisdictions. Global Capability Centres (GCCs) are becoming central to that architecture.^[a] They can build models, govern data, redesign workflows, integrate automation into business processes, and support product and platform innovation across geographies.^[4],[5] Unlike third-party outsourcing arrangements, the captive structure of GCCs gives the parent firm direct control over intellectual property, data security, quality, compliance, and strategic alignment while drawing on offshore talent and cost advantages.^[6]

India is well placed in this transition. By FY2024, it hosted more than 1,700 GCCs, generating about US$64.6 billion in annual revenue and employing around 1.9 million professionals. Industry and government estimates suggest that the sector could reach roughly US$100-105 billion by 2030, with employment rising to at least 2.5 million and, under more optimistic assumptions, almost 4-4.5 million.^[7] The wide range of projections is important: it reflects genuine uncertainty about the pace and employment intensity of AI adoption.^[8]

The talent story is central to this positioning. India accounts for a large share of the global STEM (Science, Technology, Engineering, and Mathematics) and software engineering workforce, and recent AI hiring indicators show rapid growth in AI-related roles.^[9] Yet, labour abundance alone is no longer enough; the value of India as a GCC destination increasingly depends on the density of deployable digital, domain, governance, and AI capabilities within its workforce.^[10] Recent evidence suggests that this shift is already underway inside India’s GCC ecosystem.^[11] A large majority are scaling generative AI initiatives, investing in agentic AI, expanding enterprise automation, and building dedicated innovation teams. The significance is not that GCCs are using AI to do the same work faster, but that they are beginning to redefine what work can be owned and delivered from India.^[12]

The rest of the report proceeds as follows. Section 2 traces the evolution of India’s GCC model from cost-arbitrage centres to capability platforms. It situates this transition within the broader rise of AI-enabled enterprise transformation. Section 3 examines the counter-argument that AI could weaken the economic rationale for offshoring, before showcasing why enterprise AI deployment continues to require organisational capabilities that mature GCCs are well placed to provide. Section 4 presents an exploratory scenario analysis of possible AI-related employment effects in the Indian GCC sector between 2024 and 2030. Section 5 places these findings within India’s wider labour-market realities, focusing on scale, skills, distribution, and entry-level compression. Section 6 discusses the policy implications, including skilling reform, geographic dispersal, AI governance, labour-market monitoring, and stronger GCC-startup linkages.

India’s GCC Model: From Cost Arbitrage to Capability Platforms

India is the dominant global GCC location in terms of scale, maturity, and breadth of functions. Its centres have moved well beyond routine support work, as many now own mandates across product engineering, cybersecurity, analytics, cloud, finance, risk, compliance, and innovation. The sector’s projected growth, however, should not be read as a simple continuation of the past. AI changes both the amount of labour required and the kinds of capabilities that determine where mandates are located.^[13],[14]

Figure 1: Number of GCCs in India (Actual and Projected: 2000–2030)

Source: Authors’ own, created using Matplotlib. Data from SRKAY, PIB, and Economic Times.^{[15],[16],[17]}

India’s GCC advantage rests on a combination of talent, cost, infrastructure, policy support, and English-language capability. These conditions remain important, but they no longer fully explain the sector’s value proposition. The decisive question is whether the country can combine them with AI-ready skills, trusted data governance, reliable digital infrastructure, and the managerial capacity for complex global delivery.^[18],[19]

The evolution of GCCs spans three broad phases: the first phase, in the late 1990s and early 2000s, centred on back-office and basic IT functions. The second phase added process ownership, software development, product engineering, quality assurance, and financial modelling. The current phase is defined by centres of excellence in analytics, cloud, cybersecurity, AI, and product innovation. In this phase, the strongest GCCs are not merely executing tasks delegated by headquarters; they are increasingly sharing responsibility for global outcomes.

The sheer breadth of the sector reinforces its policy significance. Technology and Banking, Financial Services, and Insurance (BFSI) remain dominant, but GCC activity now spans retail, healthcare, pharmaceuticals, manufacturing, engineering, telecommunications, energy, and logistics. A single centre may host functions across software engineering, data science, finance, HR, procurement, cybersecurity, compliance, and product design. This multi-functional character is what makes the AI transition consequential: automation affects not entire occupations but a portfolio of tasks across the enterprise.^[20],[21]

AI adoption is intensifying this transformation. Findings from the EY India GCC Pulse Survey 2025 on generative AI, agentic AI, enterprise automation, and innovation teams indicate a systematic shift from labour-cost arbitrage toward capability arbitrage. India is being leveraged not simply because labour is cheaper, but because mature GCCs can combine technical talent, operating knowledge, and AI deployment experience at scale.^[22] Recent industry evidence sharpens this transition. India’s GCC ecosystem is no longer expanding only as a larger offshore delivery base; it is increasingly being recast as a strategic layer of global enterprise operations. Newer estimates from the Nasscom-Zinnov 2026 report place the ecosystem at more than 2,100 centres, employing over 2.3 million professionals and generating close to US$100 billion in annual revenue, but the more important shift is qualitative rather than numerical.^[23] Indian GCCs are moving from “delivery engines” to “enterprise nerve centres”, with multinational firms assigning India-based teams greater responsibility for products, AI systems, platforms, engineering, research, and business outcomes.

Over 1,200 Indian GCCs now host AI and machine learning (ML) capabilities, while some 250 run dedicated centres of excellence, indicating that India is being used not merely for execution but for enterprise transformation.^[24] It also complicates the future of India’s wider export-services ecosystem: AI-focused GCCs may not end offshoring, but they are likely to internalise more strategic work that previously flowed to third-party IT services firms, especially in product engineering, platform ownership, enterprise AI deployment, automation, and transformation programmes.^[25]

The economic footprint of this ecosystem extends beyond direct headcount and revenue. GCCs channel foreign direct investment into office infrastructure, labs, digital systems, and training programmes; they generate tax revenue, and they support indirect and induced employment. They have played a pivotal role in elevating India’s services exports (see Figure 2), which provide a comfortable surplus to balance the merchandise trade deficit. Existing multiplier estimates are dated, however, and should be treated as indicative rather than current measures of aggregate impact.^[26]

Figure 2: India’s Services and ICT Exports 1975–2023/24 (Current US$)

Source: Authors’ calculations using World Bank World Development Indicators data.^[27] Figure created using Matplotlib.

GCC linkages with India’s startup and innovation ecosystem are also material. Corporate procurement, partnerships, and service demand create revenue opportunities for domestic technology firms, while exposure to global standards can deepen local capabilities. These linkages are important because they are one pathway through which GCC activity can create value beyond the employees directly on GCC payrolls.^[28]

The central policy question, therefore, is not whether GCCs are important. It is whether the conditions that produced India’s GCC advantage will hold when AI changes what work is done, where it is done, and how much human labour is required to do it.

Automation Vs. Orchestration: Why GCCs Still Matter

The argument that AI strengthens India’s GCC model rests on a complementarity assumption: enterprise transformation requires human orchestration, domain judgement, and organisational capability that GCCs can provide. That assumption must be tested against the strongest opposing case. The risk is straightforward: if generative and agentic AI systems can perform code generation, software testing, customer support, document processing, financial modelling, data analysis, and report writing at near-zero marginal cost, the old case for large offshore labour pools weakens. In that scenario, the relevant comparison is no longer onshore labour versus offshore labour; it is labour of any kind versus algorithmic execution.^[29]

Early evidence from advanced economies like the United States and the United Kingdom support this notion. AI tools have generated sizeable productivity gains to the tune of 20-60 percent and 15-30 percent, respectively, in controlled and field settings, while some studies show weaker demand for novice workers in AI-exposed occupations. While novice workers might gain from AI in completing simple tasks, the evidence for complex tasks is mixed. Although findings differ across studies, there is consensus that high-wage and cognitive occupations face the greatest risk. There are also some indications of AI lowering demand for junior-level roles but the magnitudes differ across countries: 13 percent in the US and less than 1 percent in Denmark. The implications are direct for an economy like India, where GCCs and technology services have historically absorbed fresh engineering graduates into formal high-skill work. AI may reduce entry-level absorption even if aggregate GCC output continues to grow.^[30],[31]

The concern is sharper when viewed through the lens of the displacement-reinstatement framework. If GCCs mainly deploy AI as a cost-saving tool for routine white-collar tasks, and if the new tasks created are too few or too senior to absorb affected workers, the result could be a progressive hollowing out of employment even as revenue per employee rises.^[32] However, the evidence at present does not support a simple substitution story. Three considerations matter.

First, automating bounded tasks is different from transforming enterprise workflows. Enterprise AI requires clean data pipelines, validation, regulatory compliance, human accountability, cross-functional redesign, and change management. These are interdependent organisational activities, not discrete tasks that can be handed to a model. Mature GCCs are well placed to perform this integration work. Second, the employment evidence remains early-stage and context-dependent. The latest Indian Economic Survey suggests that productivity gains are real but uneven, and that broad-based labour-market displacement has not yet appeared at the economy-wide scale. In labour-abundant economies such as India, low relative labour costs and uneven automation readiness may further slow the transition.^[33]

Third, Indian GCCs are not passive objects of disruption. Many are building AI capabilities, centres of excellence, governance teams, and new product mandates. This is the reinstatement channel in practice, where tasks are emerging around AI governance, model supervision, data quality, compliance automation, AI product management, and human-machine workflow design.

The optimistic conclusion remains conditional. If India’s talent pipeline, data governance frameworks, and AI infrastructure lag behind competing jurisdictions, high-value mandates can migrate elsewhere. The future of GCCs is therefore not determined solely by AI but is co-produced by intentional and firm strategy and public policy.

AI and the Employment Intensity of GCC Growth

The preceding sections argued that AI is unlikely to make the Indian GCC model obsolete in the medium term. The more plausible risk is subtler: AI may raise the value and productivity of GCC operations while reducing the employment intensity of their growth, particularly in routine cognitive and entry-level functions. A scenario exercise is therefore useful because it translates this qualitative claim into a bounded quantitative framework. It does not attempt to forecast the future size of GCC employment with precision. Instead, it estimates how far employment could fall below an optimistic 2030 baseline under different assumptions about AI adoption, task exposure, and the creation of new AI-augmented roles.

This distinction is important. The exercise should not be read as a prediction that AI will eliminate a fixed number of GCC jobs. It is a structured uncertainty analysis. Its purpose is to identify the scale and timing of potential deviations from a high-growth baseline and to clarify the policy response required if AI adoption is gradual rather than front-loaded—i.e., assigning the majority of capital or budget for initial stages of a project or policy.

What the Model Estimates

The model estimates displacement relative to a counterfactual 2030 employment baseline of 4.5 million GCC workers. This baseline is drawn from the high end of industry projections and should be interpreted as an optimistic upper-bound growth path rather than a neutral forecast. The choice of a 4.5 million baseline is deliberate: it allows the model to estimate how far employment could fall short even under a favourable GCC expansion scenario, rather than measuring absolute job losses from the current workforce. The outcome variable is therefore not absolute job loss from the current workforce, but the shortfall from a possible 2030 employment level if AI adoption reduces the demand for labour in exposed task categories.

The model is built at the level of functional clusters within GCCs. Each function is assigned a workforce weight, an observed AI exposure score, a theoretical exposure ceiling, and a task heterogeneity adjustment. The exposure gap captures the distance between current AI use and the potential frontier of AI-enabled task performance. The task heterogeneity term reflects the fact that functions differ in their susceptibility to automation: routine cognitive tasks are more exposed than judgment-intensive, managerial, or domain-specific activities.

Model Specification

For each function f, scenario s, and year t, gross displacement is estimated as the product of the employment baseline, the function’s workforce share, the change in AI exposure, task heterogeneity, the displacement coefficient, and an India adjustment factor. The main relationship can be written as:

Here, D_s,t represents gross displacement in scenario s at time t; B is the 4.5 million employment baseline; w_f is the workforce share of function f; ΔE_f,s,t is the scenario-specific change in AI exposure; TH_f is the task heterogeneity multiplier; β is the displacement coefficient; and λ_s is the India adjustment factor. Net displacement after reinstatement is then calculated as:

where ρ denotes the reinstatement assumption. Two reinstatement sensitivities are used: 20 percent, representing limited new-role creation in lower-maturity centres and 40 percent, representing stronger new-task creation in mature integrator-style GCCs.

Table 1: Estimation Parameters and Interpretation

Term	Interpretation	Role in the Exercise
B	2030 GCC employment baseline of 4.5 million	Defines the counterfactual high-growth path against which displacement is measured
w_f	Workforce weight of each GCC function	Allocates baseline employment across ER&D/software, BPM, finance/compliance, management, and corporate support
ΔE_f,s,t	Change in AI exposure under scenario s and year t	Captures movement from observed AI use towards the theoretical exposure ceiling
TH_f	Task heterogeneity multiplier	Adjusts for differences in functional vulnerability to AI-enabled automation
	Displacement coefficient	Translates a 10-point rise in exposure into an estimated employment effect
λ_s	India adjustment factor	Scales the displacement coefficient to reflect India-specific adoption and labour-market conditions
ρ	Reinstatement effect	Offsets gross displacement by allowing for new AI-augmented roles

Source: Authors’ formulation

Scenario Design

The scenarios are organised as a 2×3 design. The first dimension is the speed of convergence toward the theoretical AI exposure ceiling: slow, moderate, and rapid. These correspond to 50 percent, 67.5 percent, and 87.5 percent convergence by 2030. The second dimension is the path of convergence. Linear scenarios assume that AI adoption proceeds at a constant annual pace. Exponential scenarios assume front-loaded adoption, where displacement pressure is sharper in the early years before moderating as the easy gains from automation are exhausted.

This design intentionally separates endpoint effects from timing effects. Scenarios A and D, B and E, and C and F reach the same 2030 displacement endpoint. They differ in the trajectory through which they get there. This matters because a gradual path gives policymakers and firms time to sequence curriculum reform, apprenticeships, and reskilling, while a front-loaded path requires transition systems to be in place before displacement becomes visible in aggregate employment data.

Table 2: Adoption Scenarios and Interpretation

Source: Authors’ formulation

Calibration and the Limits of the Estimate

The scenario model should be read as a calibrated stress test and not a point forecast. Its purpose is to test whether AI could reduce the employment intensity of India’s GCC growth relative to an optimistic 4.5 million-worker baseline by 2030. This is consistent with broader evidence on AI and work: the World Economic Forum’s ‘Future of Jobs Report 2025’ projects both job creation and job displacement by 2030, while estimating that 39 percent of key labour-market skills will change over the same period.^[34] The model, therefore, includes both a displacement channel and a reinstatement channel, rather than assuming that automation produces only job losses.

The calibration follows three principles. First, AI exposure is treated at the task and function level, not at the level of whole occupations. This reflects the report’s core argument that AI is more likely to automate software testing, documentation, reporting, support, and routine analysis than domain judgment, compliance accountability, cross-functional coordination, or enterprise redesign. Second, an India adjustment factor is used because the impact of AI adoption in India will be mediated by labour costs, skills heterogeneity, infrastructure readiness, and institutional capacity. This is also the approach taken by NITI Aayog’s ‘Roadmap for Job Creation in the AI Economy’, which uses a Work-Worker-Workforce framework and shows that adverse employment outcomes are possible in technology services and customer experience unless firms and policymakers actively manage the transition.^[35] Third, the reinstatement term captures new AI-augmented roles in model supervision, AI governance, data quality, compliance automation, product ownership, and human-machine workflow design.

The model’s assumptions should therefore be interpreted transparently. The 4.5 million baseline defines the high-growth counterfactual; workforce shares allocate that baseline across GCC functions; exposure gaps measure movement from current AI use toward a theoretical automation frontier; task heterogeneity adjusts for differences in automatability; the displacement coefficient converts additional exposure into reduced labour demand; the India factor scales this effect to domestic conditions; and reinstatement offsets part of gross displacement through new tasks.

The model has three limitations. It does not estimate wage effects, productivity gains, or second-order employment generated through GCC-linked consumption, startup partnerships, or local services. It also does not model firm-level differences between low-maturity cost centres and high-maturity innovation centres beyond the reinstatement sensitivity. Finally, the estimates are sensitive to the 4.5 million employment baseline. The results should, therefore, be read as an employment-intensity stress test, not as a full general-equilibrium estimate of AI’s impact on India’s technology labour market.

Size of the Employment Shortfall

The results indicate that the estimated employment shortfall from the 4.5 million baseline is modest in aggregate terms but meaningful for the segment of the labour market most directly affected by GCC hiring. Under the most conservative scenario, gross displacement is approximately 40,183 jobs by 2030. Under the most aggressive scenario, it rises to approximately 150,686 jobs. These estimates are not large relative to India’s total labour force, but they matter for the formal, high-skill entry routes through which engineering, Business Process Management (BPM), and early-career technology workers enter global production systems.

Table 3: Net GCC Worker Displacement Under Different Scenarios

Source: Authors’ estimates

The identical 2030 values for A and D, B and E, and C and F are a feature of the scenario design, not an error. The paired scenarios share the same endpoint but differ in timing. In the exponential cases, displacement is concentrated earlier in the period, especially between 2025 and 2027. The policy implication is that the timing of adjustment support matters as much as the final magnitude of displacement. A government or firm response calibrated only to 2030 totals would miss the transition pressure created by front-loaded adoption.

The reinstatement sensitivities also matter. A 20 percent reinstatement assumption reduces the highest displacement estimate from 150,686 to 120,549, while a 40 percent reinstatement assumption reduces it further to 90,412. However, reinstatement does not eliminate net displacement in any scenario. This supports the report’s broader argument: new AI-augmented roles in governance, model supervision, data quality, compliance automation, AI operations, and product ownership can offset some displacement, but they should not be assumed to absorb all the workers affected by automation in routine functions.

Transition Risk, Not Sectoral Collapse

In relative terms, the scenario results are not large enough to imply a collapse in GCC employment, but they are significant for policy. A gross displacement of 40,000-150,000 jobs amounts to roughly 0.9-3.3 percent of the optimistic 2030 GCC employment baseline of 4.5 million jobs. More importantly, relative to the projected increase from about 1.9 million GCC workers in FY2024 to 4.5 million by 2030, the estimated displacement represents around 1.5-5.8 percent of incremental job creation. Annualised over 2025-2030, this implies an average shortfall of roughly 6,700-25,100 jobs per year before reinstatement effects. The policy concern, therefore, is not aggregate sectoral decline, but the possible compression of new formal high-skill opportunities, especially for early-career workers.

The scenario exercise strengthens the report in three ways. First, it clarifies that the disappearance of the GCC model is not the central labour-market risk, but rather a reduction in employment relative to optimistic growth expectations. Second, it shows that the main policy uncertainty concerns timing and scale: front-loaded AI adoption (stronger displacement effect) requires earlier policy intervention than a linear diffusion path (gradual AI adoption). Third, it links employment outcomes to GCC maturity. Centres that remain focused on cost arbitrage and routine execution are more exposed to net displacement, while transformer and integrator centres are better positioned to recreate work through higher-value mandates.^[b]

The defensive protection of existing tasks is, therefore, not the appropriate policy conclusion; it is active transition management. India should encourage GCCs to move up the maturity ladder, expand earn-and-learn pathways for early-career workers, monitor task-level shifts in hiring, and build skilling systems before displacement is visible in lagged employment statistics. Read in this way, the model provides a disciplined bridge between the report’s conceptual claim and its policy recommendations: AI can deepen India’s GCC advantage, but without a deliberate institutional response, that advantage may become more productive and more unequal at the same time.

India’s Labour Market Reality: Scale, Skills, and Uneven Gains

Any serious treatment of AI and labour markets in the Indian context must move beyond the GCC sector itself and confront the structural features of the labour market in which GCCs are embedded. The policy significance of GCCs cannot be assessed in isolation from the economies in which they operate. Three dimensions are especially important: the question of scale, the depth of the skills challenge, and the distributional implications of a growth model centred on high-skill enclaves.

India’s GCC sector is impressive by any measure of the global services industry; however, its weight within the domestic labour market requires careful perspective. As of FY2024, GCCs employed approximately 1.9 million professionals, while newer FY26 estimates place employment at over 2.3 million. India’s total labour force, as of 2023–24, exceeds 600 million. Even with the higher FY26 estimate, GCC employment accounts for less than 0.4 percent of the workforce.^[36] Even the most optimistic projections of 4 to 4.5 million by 2030 would bring that share to just under 1 percent. These are not trivial numbers in absolute terms, as 1.9 million high-skill jobs generate significant purchasing power, tax revenue, and urban economic activity. However, they are not, by themselves, a labour market strategy for a country where absorbing millions of new entrants annually into productive work is the central employment challenge.

The broader labour market context is sobering. The IHD–ILO (Institute for Human Development-International Labour Organization) India Employment Report 2024, drawing on two decades of PLFS (Periodic Labour Force Survey) and National Sample Survey (NSS) data, documents that nearly 90 percent of India’s workforce remains in informal employment.^[37] The slow structural transformation from agriculture to non-farm sectors, which had been underway since the early 2000s, reversed after 2019 with a measurable increase in agricultural employment—widely interpreted as distress-driven absorption rather than productive reallocation.

Figure 1: Structural Transformation Since 2000

Source: ILO^[38]

The real wages of regular workers have either stagnated or declined since 2012, and self-employed real earnings have fallen since 2019. Over two-thirds of unskilled casual agricultural workers did not receive the prescribed minimum daily wage in 2022.^[39] This is the labour market against which GCC employment projections must be evaluated. The question is thus whether GCC growth will generate meaningful spillovers for the wider economy or create high-productivity enclaves that coexist with, but do not substantially transform, a vast informal sector characterised by low wages, insecure work, and limited access to social protection.

On the one hand, India’s 28 percent share of the global STEM workforce and its leading position in AI hiring growth paint a bright picture.^[40] The other side is less reassuring: according to the India Employment Report 2024, 75 percent of Indian youth could not send an email with an attachment, 60 percent could not perform basic file operations such as copy and paste, and 90 percent could not use a mathematical formula in a spreadsheet.^[41] These are not measures of advanced AI fluency but are basic digital literacy benchmarks that a majority of India’s young workforce fails to meet. The proportion of educated youth among the total unemployed nearly doubled, from 35.2 percent in 2000 to 65.7 percent in 2022,^[42] suggesting that the expansion of higher education has outpaced the economy’s capacity to generate jobs commensurate with the qualifications being produced.

This creates a paradox for the GCC sector. At the top of the talent distribution, India possesses a globally competitive layer of engineers, data scientists, and AI specialists, as reflected in the 2.5 AI skill penetration rate recorded by the Stanford AI Index 2025.^[43] However, this elite layer is thin relative to the total graduate output, and GCC employers consistently report difficulty in finding candidates who combine technical skills with the domain expertise, communication skills, and problem-solving orientation required for higher-value roles. A 2025 roundtable with GCC industry leaders highlighted talent acquisition and retention as among the most pressing operational challenges, particularly as GCC mandates shift from routine execution to innovation and product ownership.^[44]

The gap between the top and the median of the talent distribution is where the skills mismatch aggravates. India produces approximately 1.5 million engineering graduates annually, but only a limited proportion are considered employable by the IT and GCC sectors without significant additional training.^[45] An IIM Ahmedabad survey found that while 55 percent of white-collar workers have adopted AI tools and 48 percent have received some training, 68 percent fear that their roles could be partially or fully automated within five years, and 40 percent believe their current skills will become redundant.^[46] This is not a workforce that feels equipped for the transition described by the optimistic GCC narrative. If the projections of 2.5 to 4 million GCC employees by 2030 are to be realised, and if these employees are to perform the high-value AI-augmented roles that justify the “innovation arbitrage” thesis,^[c] India’s skilling infrastructure must be reformed at a pace and scale that current institutional arrangements do not yet appear to support.

The distributional question is a further dimension that the GCC discourse rarely engages with. GCC workers are, by Indian standards, well compensated. AI-skilled roles command a wage premium estimated at 28 percent against 12 percent for comparable digital-skilled positions.^[47] However, three distributional concerns deserve attention. First, it is unclear how the productivity gains from AI adoption within GCCs are being shared. If AI raises output per employee while headcount growth slows, then the additional value accrues primarily to multinational parent firms in the form of higher margins, rather than to Indian workers in the form of higher wages or to the Indian economy in the form of expanded employment. The available data does not yet allow a definitive answer; however, the question should be asked explicitly, because the answer determines whether AI-driven GCC growth is a shared prosperity story or a corporate efficiency story with limited domestic distributional benefits.

Second, the GCC wage premium, while real, may be widening inequality within the Indian labour market rather than narrowing it. A sector that employs fewer than 2 million people at salaries several multiples above the median and operates alongside a vast informal economy where real wages have stagnated creates a dual labour market dynamic. This has implications for housing costs, urban planning, and social cohesion—particularly in the cities where GCCs are concentrated. This is not an argument against GCC growth, but it is an argument for complementary policies that ensure that the benefits diffuse beyond the directly employed.

Third, and most consequentially for the near term, is the question of entry-level hiring. The evidence reviewed in the preceding section points to a structural risk that is specific to the AI era. GCCs have historically served as a major entry point for fresh engineering graduates, offering them access to global production systems, exposure to frontier technologies, and a pathway into the formal high-skill economy. If AI tools enable smaller, more senior teams to perform work that previously required large junior cohorts, this entry-level absorption function weakens. The NITI Aayog’s Roadmap for Job Creation in the AI Economy puts numbers to this risk. In a worst-case scenario, the headcount in the tech services sector could decline from 7.5–8 million in 2023 to 6 million by 2031, and the customer experience sector could see a decline from 2–2.5 million to 1.8 million. These are not forecasts; however, they clearly describe the downside risk: without proactive transition management, India’s most celebrated knowledge economy sectors could become less reliable generators of new formal employment than in earlier phases of digital offshoring.^[48]

From GCC Growth to GCC Strategy: India’s Policy Agenda

The analysis leads to a conditional conclusion: AI-driven restructuring can deepen India’s GCC advantage, but only if public policy and firm strategy build the capabilities that make complementarity possible. The priority is not to protect every existing task; it is to accelerate movement toward higher-value mandates while ensuring that workers have pathways into the new task categories being created.

Building a National GCC Strategy

India does not currently have a unified national policy framework for GCCs. Several state governments, including Karnataka, Telangana, Tamil Nadu, Kerala, and Gujarat, have developed their own incentive structures, ranging from land and infrastructure subsidies to sector-specific investment promotion.^[49] These efforts have contributed to the ecosystem’s growth; however, they are fragmented, sometimes duplicative, and not designed to address the systemic challenges introduced by the AI era. What is needed is a centrally coordinated strategy that does not displace state-level initiatives but rather provides a common architecture for talent pipeline development, data governance standards, regulatory alignment, and infrastructure investment.

The IFSCA (Global In-House Centres) Regulations, 2025, notified in December 2025, which established a modernised regulatory framework for financial-sector GCCs operating within India’s International Financial Services Centre, is a useful precedent.^[50] These regulations introduced flexible operating models such as captive centres, shared services, build-operate-transfer, joint ventures, and hybrid structures. They eased employee transfer norms, permitted the limited servicing of Indian group entities, and streamlined governance and compliance requirements. The framework demonstrates that purposeful regulatory design can make India more attractive for high-value GCC operations. Although the IFSCA framework applies specifically to global in-house operations tailored for the financial sector within the IFSC ecosystem, it offers a useful regulatory template for how India can design targeted, sector-sensitive rules for higher-value GCC activity. The principle should be extended beyond financial services through a national GCC policy council. This can operate in coordination with the proposed AI Economic Council recommended by the Economic Survey 2025–26. It could align incentives across sectors, track the evolution of GCC mandates, and ensure that India’s regulatory environment keeps pace with the changing nature of work in these centres.

Rewiring the Skilling Pipeline

The skills mismatch documented in the preceding section cannot be addressed through incremental measures. India needs a structural overhaul of the pipeline that connects higher education to GCC-ready employment. Four specific interventions deserve priority. First, AI and data literacy must be integrated into undergraduate curricula across disciplines, not merely confined to computer science and engineering programmes. The National Education Policy 2020 provides the framework for this through its emphasis on multidisciplinary education and flexible credit pathways, but implementation remains uneven. Universities need both the autonomy and the funding to design industry-relevant courses. The Viksit Bharat Shiksha Adhishthan Bill, 2025, currently under parliamentary committee review, points toward a possible restructuring of higher-education regulation; however, its eventual implications for AI-ready curricula will depend on implementation.^[51] The Economic Survey 2025–26 rightly emphasised that countries leading in AI innovation combine classroom learning with real industry exposure, which is currently lacking in India’s higher education system.

Second, industry-academia partnerships for AI skilling should be expanded. For instance, FutureSkills PRIME has made a credible start in offering certification-based learning in emerging technologies, but its reach remains limited relative to the scale of the challenge. GCC employers should be incentivised through tax credits, co-investment frameworks, or procurement preferences to open their training infrastructure to students and mid-career professionals, effectively converting GCCs into distributed skilling platforms.

Third, earn-and-learn apprenticeship models should be formalised and scaled. India’s formal apprenticeship system remains small and manufacturing-oriented. For the services and knowledge economy, a new model in which students spend structured periods within GCC operations during their degrees, acquiring technical, organisational, communicative, and domain-specific competencies, is needed. This way, the critical skill gaps consistently identified in fresh graduates are addressed. The introduction of certification-based alternative eligibility for key managerial personnel in fund management entities, with reduced experience requirements, within the IFSCA 2025 regulations, offers a template that could be adapted for the broader GCC workforce.^[52]

Fourth, reskilling programmes for mid-career professionals must be designed with the same seriousness as entry-level pipelines. The IIM Ahmedabad survey finding that 40 percent of white-collar workers believe their current skills will become redundant is a leading indicator of a workforce transition that cannot be managed solely through voluntary upskilling. Government, industry, and educational institutions need to co-design transition pathways, including income support, modular credentialing, and placement linkages, helping workers to efficiently handle the new task categories that AI is creating through the automation of roles.

Taking GCCs Beyond the Metros

The concentration of GCCs in five or six metropolitan centres is a source of increasing vulnerability. These cities face infrastructure constraints, urban congestion, rising real estate costs, and talent-market saturation that are beginning to erode their competitive advantage. Tier-1 cities are gradually losing ground, and firms are already exploring Tier-2 locations such as Ahmedabad, Coimbatore, Kochi, Jaipur, and Chandigarh, which offer lower operating costs and reduced attrition. Policy should accelerate this shift, and not merely accommodate it.

Three levers are available. First, infrastructure investment in Tier-2 and Tier-3 cities in digital connectivity, reliable power, quality office stock, and housing should be treated as GCC-enabling expenditure with measurable returns. Second, state governments in these cities should be supported in developing GCC-specific investment promotion strategies, including dedicated facilitation desks, streamlined approvals, and targeted skill development partnerships with local universities. Third, the central government could introduce time-limited fiscal incentives, such as enhanced deductions for GCC payroll in designated emerging locations, to lower the risk for firms willing to establish operations outside the saturated metros. The objective is to distribute the gains of GCC growth more broadly across the country, which matters in its own right. It also enhances efficiency: talent costs are lower, attrition rates are lower, and a geographically diversified GCC ecosystem is more resilient to localised disruptions.

Governing AI-Intensive GCC Operations

GCCs in India now handle activities that raise important governance questions vis-à-vis the cross-border transfer of sensitive personal and financial data; the development and deployment of AI models that make consequential decisions for global enterprises; algorithmic processing subject to regulatory requirements across multiple jurisdictions; and the management of intellectual property generated through AI-assisted innovation. India’s current regulatory architecture does not provide adequate clarity on several of these dimensions. India’s Digital Personal Data Protection Act, 2023, together with the notified Digital Personal Data Protection Rules, 2025, now provides the basic framework for personal data governance. However, sector-specific guidance for AI deployment, especially in financial services, healthcare, cross-border analytics, and automated decision-support systems, remains underdeveloped.^[53]

As the European Union (EU) advances its AI Act and other jurisdictions develop their own governance frameworks, multinational firms will increasingly route AI-intensive mandates to locations that offer both capability and regulatory certainty. If India is perceived as a jurisdiction with ambiguous or weakly enforced AI governance, GCC mandates in the most sensitive and highest-value functions may migrate to competing locations. The proposed AI Economic Council should be complemented by sector-specific guidance for GCC operations, developed in consultation with industry and aligned with the governance frameworks of the jurisdictions in which GCC parent firms operate. India’s objective should be to transform into the most intelligently regulated destination for AI work, as opposed to the least regulated one.

Tracking the AI Labour Market Transition

India currently lacks a systematic mechanism to track how AI adoption is affecting task composition, occupational structure, and hiring patterns within the GCC and IT-BPM sectors. The PLFS captures broad employment categories but does not disaggregate at the level needed to detect, for example, whether entry-level software engineering roles are declining while AI governance and orchestration roles are expanding. NASSCOM publishes aggregate headcount and revenue data, but not the task-level or occupational-level detail that would reveal early signs of displacement or reinstatement.

This information gap should be closed. A dedicated GCC employment observatory, which could be jointly operated by government, industry, and independent researchers, could track quarterly changes in the occupational composition of GCC hiring, monitor the ratio of entry-level to senior roles, survey firms on AI-driven task reallocation, and publish findings that inform both policy and educational planning. According to NITI Aayog’s worst-case employment projections, if the technology services sector is to lose 1.5–2 million jobs by 2031 in an adverse scenario, policymakers need early warning systems, not retrospective analyses published after the displacement has occurred.

Turning GCC-Startup Linkages into Innovation Spillovers

GCCs contribute an estimated US$15 billion annually to Indian startups through procurement, partnerships, and demand for services. This is an important and underappreciated channel through which GCC activity generates domestic value beyond direct employment. However, the linkage is largely informal and market-driven. Policy can deepen and formalise it in three ways.

First, GCCs could be encouraged to source a larger share of technology and analytics needs from Indian startups, particularly those working on AI applications relevant to the GCC’s sector. Second, joint intellectual-property frameworks should give Indian startups meaningful ownership stakes in innovations developed through GCC collaborations, rather than routing all value to multinational parents. Third, regulatory sandbox environments could be established, building on models already being utilised by the RBI (Reserve Bank of India) and IFSCA, in which Indian startups test AI-powered solutions within live GCC operations. These measures would turn buyer-supplier relationships into deeper innovation partnerships, strengthening India’s domestic technology ecosystem while giving GCCs access to local experimentation and sector-specific problem-solving.

Conclusion

India’s GCC opportunity in the AI era is tangible; however, it should be understood as a conditional development opportunity rather than an automatic employment dividend. The technology itself is unlikely to eliminate the GCC model in the medium term because enterprise deployment requires governance, domain expertise, integration, and accountability.

At the same time, it can make GCC growth more employment-light, more senior-skills intensive, and less able to absorb large cohorts of entry-level workers than earlier phases of offshoring. The policy priority is therefore to move beyond headline counts of centres, revenue, and projected employment.

India needs a strategy that raises GCC maturity, builds AI-ready talent pipelines, disperses growth geographically, provides regulatory certainty, strengthens startup linkages, and monitors task-level labour-market transitions in real time. If those institutions are built, GCCs can become platforms for capability formation and innovation. Otherwise, India may still attract high-value mandates, but the gains will remain narrow, uneven, and weakly connected to the wider labour market.

Arya Roy Bardhan is Junior Fellow, Centre for New Economic Diplomacy (CNED), ORF.

Soumya Bhowmick is Fellow, CNED, ORF.

All views expressed in this publication are solely those of the authors, and do not represent the Observer Research Foundation, either in its entirety or its officials and personnel.

Acknowledgement

The authors acknowledge the use of Claude Opus 4.8 for background research and language editing. All factual content, analysis, and final text remain the responsibility of the authors. They also thank Manish Vaidya and Siddharth Joshi at ORF for their assistance with data collection and analysis.

Endnotes

^[a] GCCs are wholly-owned offshore subsidiaries of multinational corporations, established to deliver specialised services such as IT (Information Technology), R&D (Research and Development), finance, analytics, customer support, and business operations to the parent organisation and its global network.

^[b] A Transformer GCC drives enterprise-wide change, innovation, and business model transformation. It owns the end-to-end design and execution of disruptive strategies rather than just executing tasks assigned by headquarters. An Integrator GCC acts as a centralised, cross-functional hub that unifies fractured global operations, data silos, and technology architectures across different business units and geographies.

^[c] Innovation arbitrage is the practice of identifying valuable intellectual property (IP), technologies, or business models in one market which are then monetised by another market. Instead of just sending routine development and support tasks to lower-cost regions, multinational corporations establish advanced AI labs and R&D hubs. They tap into highly skilled, localized talent pools to develop novel products that can be deployed globally, extending the company’s competitive runway.

^[1] Flavio Calvino, Daniel Haerle, and Silvia Liu, “Is Generative AI a General-Purpose Technology? Implications for Productivity and Policy,” OECD Artificial Intelligence Papers, no. 40 (Paris: OECD Publishing, 2025), https://doi.org/10.1787/704e2d12-en.

^[2] Daron Acemoglu and Pascual Restrepo, “Automation and New Tasks: How Technology Displaces and Reinstates Labor,” Journal of Economic Perspectives 33, no. 2 (2019): 3–30, https://doi.org/10.1257/jep.33.2.3.

^[3] World Economic Forum, The Future of Jobs Report 2025, January 7, 2025, https://www.weforum.org/publications/the-future-of-jobs-report-2025/

^[4] India Brand Equity Foundation (IBEF), “Global Capability Centres Are Transforming India’s Corporate Landscape,” IBEF Blog, October 18, 2024, https://www.ibef.org/blogs/global-capability-centres-are-transforming-india-s-corporate-landscape

^[5] Calvino, Haerle, and Liu, “Is Generative AI a General-Purpose Technology? Implications for Productivity and Policy.”

^[6] Srivardhini K. Jha and Anilesh Seth, “Global Capability Centres: Emerging Opportunities and Challenges,” IIMB Management Review 37, no. 3 (2025): 00595, https://doi.org/10.1016/j.iimb.2025.100595.

^[7] Press Information Bureau, Government of India, “From Policy to Prosperity: GCCs Leading India’s Growth Journey,” December 11, 2025, https://www.pib.gov.in/PressReleasePage.aspx?PRID=2202046&reg=3&lang=2

^[8] India Brand Equity Foundation (IBEF), “Indian Global Capability Centres (GCC) Industry to Hit US$ 100 Billion by 2030, Generate Over 2.5 Million Jobs,” IBEF News, November 19, 2024, https://www.ibef.org/news/indian-global-capability-centres-gcc-industry-to-hit-us-100-billion-by-2030-generate-over-2-5-million-jobs

^[9] Press Information Bureau, Government of India, “AI@Work: Driving Productivity, Jobs, and Innovation,” February 12, 2026, https://www.pib.gov.in/PressReleasePage.aspx?PRID=2226912&reg=48&lang=2

^[10] Press Information Bureau, “From Policy to Prosperity.”

^[11] EY, “India’s GCCs Are Leading the Shift to Intelligent, AI-Native Enterprises,” https://www.ey.com/en_in/insights/consulting/global-capability-centers/india-s-gccs-are-leading-the-shift-to-intelligent-ai-native-enterprises

^[12]India Brand Equity Foundation, “Indian Global Capability Centres Industry to Hit US$100 Billion.”

^[13] NASSCOM, India GCC Landscape Report: The 5-Year Journey, https://nasscom.in/knowledge-center/publications/india-gcc-landscape-report-5-year-journey#

^[14] EY, Future of GCCs in India: A Vision 2030, 2023, https://www.ey.com/content/dam/ey-unified-site/ey-com/en-in/newsroom/2022/4/documents/ey-future-of-gccs-in-india-a-vision-2030.pdf

^[15] Karunjit Kumar Dhir, “The Evolution of Global Capability Centres in India: A Journey of Two-Decades,” SRKay Consulting Group, December 2, 2024, https://www.srkay.com/newsroom/the-evolution-of-global-capability-centres-in-india-a-journey-of-two-decades.

^[16] Press Information Bureau, “From Policy to Prosperity.”

^[17] IANS, “93% of Workforce at India’s GCCs Is Gen Z and Millennial: Report,” ET HRWorld, March 12, 2026 https://hr.economictimes.indiatimes.com/news/industry/majority-of-indias-gcc-workforce-comprises-gen-z-and-millennials-report-finds/129504288

^[18] Shrikanth G., “The GCC Boom: India’s Journey from Cost Arbitrage to Innovation,” Dataquest, https://www.dqindia.com/features/the-gcc-boom-indias-journey-from-cost-arbitrage-to-innovation-10654025

^[19] Nitika Goel, Richa Kejriwal, and Sachit Bhat, “Global Capability Centers (GCCs): All You Need to Know About GCCs,” Zinnov, September 26, 2024, https://zinnov.com/centers-of-excellence/global-capability-centers-101-all-you-need-to-know-about-gccs-blog/.

^[20] ETBFSI, “US Financial Firms Make a Beeline for Setting Up GCCs in India,” February 19, 2026, https://bfsi.economictimes.indiatimes.com/articles/us-financial-firms-make-a-beeline-for-setting-up-gccs-in-india/128524636.

^[21]  EY India, “58% of India GCCs Investing in Agentic AI; Two-Thirds Creating Dedicated Innovation Teams to Globalise Ideas,” November 2025, https://www.ey.com/en_in/newsroom/2025/11/58-percent-gccs-in-india-investing-in-agentic-ai-two-third-creating-dedicated-innovation-teams-to-globalize-ideas-ey-gcc-pulse-survey-2025

^[22] EY India, “58% of India GCCs Investing in Agentic AI.”

^[23] Zinnov and NASSCOM, India GCC Landscape Report 2026, 2026, https://zinnov.com/centers-of-excellence/zinnov-nasscom-india-gcc-landscape-2026-report/

^[24] Zinnov and NASSCOM, India GCC Landscape Report 2026.

^[25] The Indian Express, “Global Firms’ AI Centres Are Reshaping Indian IT Industry, but Who Has the Top Jobs?” May 14, 2026, https://indianexpress.com/article/explained/explained-sci-tech/global-capability-centres-india-ai-engineering-growth-10688637/lite/

^[26] NASSCOM, GCC Value Proposition in India, 2021, https://nasscom.in/system/files/publication/NASSCOM_GCC_Value_Proposition_in_India_24062021.pdf

^[27] World Bank, “ICT Service Exports (BoP, Current US$) – India,” World Development Indicators, March 31, 2026, https://data.worldbank.org/indicator/BX.GSR.CCIS.CD?locations=IN

^[28] Deloitte, “Engineering Innovation,” http://deloitte.com/in/en/Industries/global-gcc/perspectives/engineering-innovation-gcc.html

^[29] Deloitte, Global Outsourcing Survey 2024, https://www.deloitte.com/global/en/issues/work/global-outsourcing-survey.html

^[30]  R. Maria del Rio-Chanona et al., “AI and Jobs: A Review of Theory, Estimates, and Evidence,” arXiv, September 18, 2025, https://arxiv.org/abs/2509.15265v1

^[31] Stanford Digital Economy Lab, “AI and Labor Markets: What We Know and Don’t Know,” October 10, 2025, https://digitaleconomy.stanford.edu/news/ai-and-labor-markets-what-we-know-and-dont-know/

^[32] Acemoglu and Restrepo, “Automation and New Tasks.”

^[33] Government of India, Economic Survey 2025–26, Chapter 14: Evolution of the AI Ecosystem in India, Ministry of Finance, 2026, https://www.indiabudget.gov.in/economicsurvey/doc/eschapter/echap14.pdf

^[34] World Economic Forum, The Future of Jobs Report 2025.

^[35] NITI Aayog, Roadmap for Job Creation in the AI Economy, 2025, https://niti.gov.in/sites/default/files/2025-10/Roadmap_for_Job_Creation_in_the_AI_Economy.pdf

^[36] “India’s Offshore Tech Hubs Hit US$98.4 Billion Revenue in FY26, Nearing 2030 Forecast,” Reuters, May 6, 2026, https://www.reuters.com/world/india/indias-offshore-tech-hubs-hit-984-bln-revenue-fy26-nearing-2030-forecast-report-2026-05-06/

^[37] International Labour Organization and Institute for Human Development, India Employment Report 2024: Youth Employment, Education and Skills, 2024, https://www.ilo.org/sites/default/files/2024-08/India%20Employment%20-%20web_8%20April.pdf

^[38] International Labour Organization and Institute for Human Development, India Employment Report 2024.

^[39] International Labour Organization and Institute for Human Development, India Employment Report 2024.

^[40]  Government of India, Economic Survey 2024–25, Ministry of Finance, 2025, https://www.indiabudget.gov.in/budget2025-26/economicsurvey/doc/echapter.pdf

^[41] International Labour Organization and Institute for Human Development, India Employment Report 2024.

^[42] International Labour Organization and Institute for Human Development, India Employment Report 2024.

^[43] Press Information Bureau, AI@Work: Driving Productivity, Jobs, and Innovation.

^[44] Jha and Seth, “Global Capability Centres.”

^[45]  Nipun Sharma, “Why Are So Many of India’s 1.5 Million Fresh Engineers Every Year Unemployable?” Forbes India, November 13, 2025, https://www.forbesindia.com/article/upfront/column/why-are-so-many-of-indias-1-5-million-fresh-engineers-every-year-unemployable/2988705/1

^[46] Brij Disa Centre for Data Science and AI, Labour-Force Perception About AI: A Study on Indian White-Collar Workers, Indian Institute of Management Ahmedabad, August 2024, https://www.iima.ac.in/sites/default/files/2024-09/Labour-force%20perception%20about%20AI%20Final%20version%20%281%29.pdf

^[47] World Bank, South Asia Development Update: Jobs, AI, and Trade, October 2025, https://openknowledge.worldbank.org/server/api/core/bitstreams/71109bfe-cb0e-47d6-b2c5-722341e42b99/content

^[48] NITI Aayog, Roadmap for Job Creation in the AI Economy.

^[49] Invest India, “Six Indian States Leading the Next Wave of IT and ITeS Growth,” https://www.investindia.gov.in/team-india-blogs/six-indian-states-leading-next-wave-it-and-ites-growth

^[50]  International Financial Services Centres Authority, IFSCA Fund Management Regulations, 2025, updated January 1, 2026, https://ifsca.gov.in/Common/PreviewPdf?id=38fea9cc5969551d78bf00e670e2eeee&fileName=268967_20260101_0255.pdf

^[51] PRS Legislative Research, “The Viksit Bharat Shiksha Adhishthan Bill, 2025,” https://prsindia.org/billtrack/the-viksit-bharat-shiksha-adhishthan-bill-2025

^[52] “IFSCA Relaxes Norms for Fund Management Entities under New GIC Rules,” Business Standard, December 23, 2025, https://www.business-standard.com/markets/news/ifsca-relaxes-norms-fund-management-entities-new-gic-rules-2025-125122301137_1.html

^[53] Press Information Bureau, Government of India, “Digital Personal Data Protection Rules, 2025,” https://www.pib.gov.in/PressReleasePage.aspx?PRID=2190655&&reg=3&lang=2

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.