Governance, agency, and survival in the era of distributed intelligence: what Global Capability Center (GCC) leaders say India’s hubs must do now to shape the next five years, rather than be shaped by them.
The question is meant to provoke, and it does. India’s global capability centers now number 2,117, employ 2.36 million people, and generate close to $98.4 billion in annual revenue, a sector that has quietly become the operating spine of hundreds of the world’s largest enterprises. Much of the artificial intelligence now reshaping global work was, in one form or another, built, trained or run from these hubs. And yet the same technology India helped scale is the one now being pointed back at the country’s own headcount.
That is the tension GCC leaders are living inside. At a recent closed-door panel, an Operations Leader, a Group Chief Operating Officer, and a veteran International Board Director set out, in unusually candid terms, what they believe the next five years demand. Their collective message
was not reassurance. It was a choice.
India’s GCC scale
all GCCs
AI/ML capabilities
of excellence
Source: Nasscom-Zinnov India GCC Landscape Report 2026; ANSR 2025
The choice that runs through everything
When headquarters is under pressure to take headcount out, and hubs are now central to their operations, the hub has two postures available. The first is defensive. “You can be defensive and bring in the regional aspect and say, we are cost-optimised, don’t cut us,” as one hub strategy leader put it. The second is to get into the room and change what the conversation is about: “If you want to cut intelligently, partner with us. Because we know the most current view of what’s actually happening on the ground.”
“If you’re not involved in this, if you aren’t shaping the cut, it’s very likely that you will get cut.”
— Hub strategy-and-operations leader
This gets to one of the central tensions in the discussion. The broader question underneath it is how a hub earns the right to shape the decision rather than absorb it.
Why "don't cut us" fails fast
The defensive posture fails not because it is wrong on the facts, but because of how it reads. To a global executive weighing options, a hub is, on the spreadsheet, simply critical mass in one location. “Org charts… are very clean,” the strategy leader noted, with clear structures, vertical lines, functional lines and matrix structures. Argue for protection in that frame and you have already lost, because you are debating the number on its own terms.
The more effective move is to change what is being measured. Rather than defend roles, a hub can enter the conversation “with a view on what roles are really required to be consolidated, where we can optimise workflow, where we are carrying hidden risks, where we can redesign instead of just cutting certain roles.” Do that and the hub stops being an interested party pleading its case and becomes the party improving the quality of the decision. That earns strategic leverage. Offence, in other words, begins well before the automation question.
The balance sheet headquarters can't see
There is a reason hub leaders are uniquely equipped to shape the cut: they can see an entire asset base that never appears in a headcount report. “Things get done when you have relationships,” one leader said. “How you ensure there is resiliency, how you ensure there are backup plans, how you ensure you’re able to scale when things get hectic, those things don’t show up in metrics.”
- This is the hub’s real inventory: who is genuinely productive, where attrition risk actually sits, where the bottlenecks are, who is carrying too much weight, and the web of relationships that makes work happen at all. Global leadership, the panel argued, consistently underestimates these intangibles, “it’s just unknown to them, invisible to them” precisely because they are invisible on the instruments headquarters uses to steer.
- The corollary is uncomfortable but clarifying: if the hub does not make this balance sheet visible, no one else will, and decisions will be made as if it did not exist. Making the invisible legible is not corporate storytelling. It is the first act of governance.
From cost per seat to capability value
Underneath the whole debate sits a measurement problem. “Right now all we measure in GCCs is often cost per seat,” a hub leader observed. “It has to shift to capability value. Once we start thinking in those terms, we will start investing in building those capabilities.” As long as the only number that travels up to the board is cost per seat, no one, in the hub or at headquarters, has a reason to fund the capabilities that determine the hub’s future.
The market has already moved further than the metric. By EY’s 2025 GCC Pulse Survey, 92 percent of GCC leaders say their centers now deliver value beyond cost arbitrage, 87 percent run end-to-end global processes, and 96 percent of centers set up since 2021 launched with a product or portfolio mandate rather than a support one. The centers have become capability owners; the spreadsheet that governs them is often still a 2018 cost sheet. AI widens that gap by design: it collapses the ratio of headcount to output, so the unit of value shifts from seats to outcomes almost overnight. A hub that cannot reframe its own business case away from cost per seat will keep being managed as the very thing it has stopped being.
The automation paradox: offence is the best defence
The most counter-intuitive conclusion of the panel was about automation itself. The instinct to shield the hub’s work from AI, to protect the teams, is precisely the instinct that destroys the hub. “To safeguard anything from automation? You’ll lose credibility very quickly,” one leader warned. “You cannot come across as anti-automation. That’s not going to take you anywhere.”
The credible move is the opposite, and it echoes how GCCs earned their mandate in the first place: “go back to first principles, they started taking problems the enterprise wanted to solve at scale.” Applied to AI, that means the hub should identify the automatable work itself, run the pilots, prove the productivity gains, and lead from the front. Crucially, this reframes what the hub is protecting. Not jobs, but service continuity, controls, client trust and capability. The one caution the panel flagged was against theatre, automation “that just gives a false sense of efficiency without actually delivering measurable gains.” Lead the automation and you own the story; resist it and you become the story.
What to build, and what to borrow
Leading on automation raises an immediate question: build the capability or buy it? The panel’s answer was a disciplined core-versus-context line. “What is core? What differentiates us, what is the IP we really want to invest heavily in? That’s what we build in-house,” said the strategy leader. “Where we can scale faster by partnering, by buying an open-source component, or through a deeper vendor partnership, that’s where we pivot.”
The group COO sharpened it with a distinction between domain capability unique to an industry and generic functional capability that runs across sectors. For functions like finance, HR or generic technology, the deepest AI investment will come from specialists who transform those functions “for a living”; no single enterprise will out-invest them. But for the capability at the heart of the business (in his case, how a bet is priced and the models trained around it), he was blunt: “nobody’s going to do that better than us… if we let others become the best in the world at the thing that’s core to our business, we risk not having a future.” For a GCC, the implication is precise: win the mandate for what is genuinely core to the parent, and partner aggressively everywhere else. Owning generic work at scale is not strategic authority. Owning the core is.
The third wave of the India story
Zoom out, and this moment fits a pattern the veteran board director traced across two decades. The first wave, in the early 2000s, made India an offshore delivery center. The second, accelerated through the pandemic, moved genuine leadership and management ownership to the country. “I think we’re about to see the third wave now, which is very much around architecting solutions, showing much more intellectual leadership in how these solutions are implemented.”
The missing piece, he argued, is for GCCs “to move beyond optimisation of skills into the capabilities required to deliver enterprise solutions all over the world.” The raw material is already there: more than 1,200 GCCs in India have embedded AI/ML capabilities, over 250 run dedicated AI centers of excellence, and the country’s AI talent concentration is now roughly twice the global average (Nasscom-Zinnov 2026; ANSR 2025).
Asked whether India would be the architect of the AI era or a spectator to it, his answer was unhesitating: “I have every confidence that it will be the architect.” But architecture is a choice a hub makes, not a status it inherits.
When the hub owns the technology, who is accountable?
The third wave surfaces a governance question the panel treated as urgent rather than academic. As ownership of technology shifts to the hub while people-decisions often stay at headquarters, accountability splits. If the hub owns the technology but HQ still owns people-decisions, who is actually accountable for enterprise outcomes? And as workflows turn agentic, the question of where AI-workforce risk should sit, audit committee, compensation committee, or the full board, stops being theoretical.
The wider market is not ready for this. By EY’s 2025 reading, 72 percent of executives have deployed AI across most initiatives but only 33 percent have proper responsible-AI protocols; Deloitte’s 2025 boardroom research found AI on 69 percent of board agendas even as two-thirds of directors admit limited-to-no AI experience. That gap is a vacuum, and it is exactly the vacuum a mature hub is positioned to fill. A GCC that can hold the controls, the model governance and the responsible-AI discipline for a global enterprise is no longer executing someone else’s strategy. It has become the place where the enterprise’s AI risk is actually managed.
The governance gap
include AI discussion
2/3
of board directors admit limited-to-no AI experience
Source: EY 2025; Deloitte Global Boardroom Program 2025
Working backwards from 2030
How should a hub plan for any of this when no one can forecast the technology? The board director’s answer was a planning discipline, not a prediction. Rather than running projects forward and absorbing the consequences, hubs should, as he put it, “work backwards.” He offered a concrete illustration: take an organisation of 5,000 people that will become 4,000, where “12 job families are going to go to seven job families, including three new ones. Getting ahead of that and working back from that answer is the way to look at this.”
That reframes the next five years as a redesign, not a retrofit, and as a period explicitly for disciplined experimentation. “You need to experiment, to fail, to get evidence of success… we have to think about redesigning more than just retrofitting.” Several panellists converged on the need for continuous review: experiment, gather evidence of what works and what does not, and keep revisiting the plan as the technology evolves. The workforce of 2030 will not be reached by extrapolating from 2025. It has to be designed backwards from where the enterprise needs to be.
The real risk
The sharpest line of the session was also its simplest. The danger is not AI. The danger is running a 2026 strategy on a 2018 operating model. Hubs that get ahead of the operating-model change lead; hubs that wait to be told get optimised. And on the technology itself, the panel landed somewhere unexpected for a conversation ostensibly about AI: “I would put the talent and human capital side as important as the technology.”
Which returns us to the question we started with, though the panel answered it with more optimism than the title suggests. India built much of the AI now in production, and the leaders in the room were clear that this is India’s phase to architect rather than sit out. What it turns on is a choice each hub gets to make: shape the next operating model, or have it shaped for you. On the evidence of this conversation, the leaders closest to the work already know which one they intend to do.
About this piece: This article synthesises perspectives shared at a closed GCC leadership panel convened in June 2026. In keeping with the Chatham House Rule, no participant is named or identified. The views are those of the leaders who took part, brought together here into a single point of view; they are not the position of Kingsley Gate.
Data sources: Nasscom-Zinnov India GCC Landscape Report 2026; EY GCC Pulse Survey 2025; ANSR AI Talent Trends in India’s GCCs 2025; Deloitte Global Boardroom Program 2025; EY responsible-AI / AI-adoption survey 2025.
English 
Español de México