Built for Change: Designing GCCs for 2030

Most discussions about AI begin with a forecast. That is precisely the problem.

In March 2025, the research group METR estimated that the length of software tasks AI agents could complete on its benchmark was doubling roughly every seven months. Less than a year later, its updated modelling put the doubling time at about four months. The benchmark measures self-contained technical tasks rather than entire jobs, but it highlights a more important reality: even careful attempts to measure the rate of change are themselves changing.

For GCC leaders, that is the planning challenge of this decade. The question is not what works a GCC will do in 2030. It is whether a GCC can keep changing what it does faster than anyone can predict what it will need to do.

From Arbitrage to Ownership

GCC work has already reinvented itself several times.

The first generation was built on labour arbitrage. Well-defined processes moved to India because they cost less to deliver there.

The second generation built capability. Engineering, analytics, product development and specialised expertise became central to the GCC value proposition.

The third generation, now well underway, is built on ownership. Increasingly, centres are being asked to own outcomes rather than simply deliver outputs.

Each generation reduced dependence on cost and increased dependence on judgement.

The shift is visible in the work itself. Zinnov data indicates that complex problem-solving now makes up about 38% of GCC work portfolios, and that cutting-edge R&D work doubled between 2015 and 2026. More than 1,200 India GCCs have embedded AI and machine learning capabilities, supported by roughly 250,000 AI professionals.

Previous transitions unfolded over years, allowing organisations time to plan, hire and adapt. The next transition may not offer the same runway.

What Is Already Changing

Agentic AI differs from previous automation waves because it changes who, or what, performs the work.

An agent does not simply execute the next instruction. It receives a goal, determines a sequence of actions, works across systems and evaluates its own output.

Nobody knows how far or how fast this capability will evolve. But several shifts are already becoming visible, and each has significant implications for GCCs.

The unit of work is changing. For decades, GCCs have been designed to absorb well-defined tasks. Agentic AI increasingly operates across multiple steps, systems and decisions. As a result, the unit of work is beginning to move from task to outcome, and teams may increasingly be defined not by the tickets they close but by the business result they are accountable for.

The pyramid is under pressure. The traditional GCC model depends on a broad base of early-career talent performing structured work. That layer is precisely where agentic systems are becoming most capable today. For decades, GCC economics have relied on scaling capability through talent pyramids. Agentic AI challenges that assumption because it directly affects the work through which future expertise has traditionally been developed.

The role of people is changing. As agents assume a greater share of execution, the role of people shifts toward judgement. Defining what good looks like, recognising exceptions, evaluating risk and accepting accountability become increasingly valuable capabilities.

The Forecasting Trap

Forecasts about agentic AI frequently point in different directions, sometimes within the same analysis.

In mid-2025, Gartner predicted that at least 15% of day-to-day work decisions would be made autonomously through agentic AI by 2028, up from zero in 2024. In the same analysis, it predicted that more than 40% of agentic AI projects would be cancelled by the end of 2027, driven by rising costs, unclear business value and weak risk controls.

Both outcomes can be true simultaneously.

Agentic AI will almost certainly be overestimated in some functions and underestimated in others. The challenge is that nobody can reliably identify which is which in advance. The skills picture is just as unstable: the World Economic Forum expects 39% of core job skills to change by 2030.

The most dangerous GCC strategy is a confident one. A five-year roadmap built around a single view of how work, skills and organisational structures will evolve risks locking in exactly the assumptions most likely to break.

Five Capabilities That Stay Valuable Under Any 2030

If future work cannot be predicted with confidence, competitive advantage shifts toward capabilities that remain valuable regardless of how the technology evolves.

Ownership of Context

Agents can execute tasks. They do not automatically understand why a process exists, which customer exception matters most or which rule reflects organisational history rather than logic. Distance from the business has long been a GCC's weak point. Centres that close that gap and develop deep business context will increasingly direct how agents operate rather than merely consume their output.

Agent Operations Discipline

Agentic systems require governance. Organisations need capabilities to design workflows, establish guardrails, monitor performance and determine when human intervention is required. GCCs are well positioned to build this discipline because they already operate significant portions of their parent organisations' technology estates.

Judgement with Accountability

Even when an agent makes a recommendation or decision, accountability remains human. The highest-value roles will increasingly sit where judgement, risk, regulation and customer impact intersect, and those roles need decision rights within the centre, not only at headquarters.

A Redesigned Talent Pipeline

If agentic systems absorb more entry-level work, the traditional route through which future leaders gain experience becomes narrower. GCCs must rethink how talent develops judgement through supervised agent work, rotational experiences and earlier exposure to business decisions.

Learning Velocity

Perhaps the most durable capability is the ability to learn faster than the environment changes.

In practice, that means reviewing the work portfolio every quarter rather than every budget cycle, ring-fencing capacity for experiments, retiring outdated processes as deliberately as new ones are started, and redeploying talent quickly. In an environment of accelerating change, adaptability becomes an operating capability rather than a cultural aspiration.

Built for Change, Not for a Forecast

No one can say with confidence what a GCC will be doing in 2030. The technology is moving too quickly, and the forecasts are moving with it.

What leaders can decide is how quickly their organisations can change course.

The first era of GCCs was built around labour.

The second around capability.

The third around ownership.

The next may be built around adaptability.

In a world shaped by agentic AI, the most valuable GCC will not be the one that predicts the future correctly. It will be the one that can change course before everyone else does.

‍