India’s GCCs set to become agentic transformation engines by 2030: Dell-Zinnov report

BENGALURU: India’s Global Capability Centres (GCCs) are entering a new phase in which their ability to scale artificial intelligence into measurable business outcomes will matter more than the number of AI pilots they launch, according to a new Dell Technologies-Zinnov report.

Titled “India GCCs 2030: From Capability Centers to Agentic Transformation Engines”, the report was released at the Dell Technologies Forum 2026. It draws on surveys and interviews with more than 50 senior GCC leaders across banking, financial services and insurance (BFSI), retail, manufacturing and software sectors.

The report identifies a gap between AI ambition and execution. Nearly 70% of GCCs remain stuck at the pilot stage, with fragmented data, legacy systems, unclear governance, immature security controls and workforce models designed for a pre-AI environment identified as key constraints.

GCC ecosystem expands

India currently hosts more than 2,100 GCCs, employing around 2.36 million people and generating $98.4 billion in revenue in FY26, the report said.

Around 64% of GCC leaders now hold dual global mandates, while 70% have a defined AI roadmap or charter. Indian GCCs account for approximately 28% of global GCC AI talent, with more than 1,200 centres having established AI and machine learning capabilities.

The report also found that 66% of GCC leaders rank top-line business impact as a high priority for their enterprise AI strategy, indicating a shift beyond traditional cost and delivery objectives.

AI maturity curve is compressing

The report noted that GCC maturity is advancing faster than in previous cycles. About 27% of new GCCs now reach Portfolio Hub maturity within five years, compared with nearly a decade historically.

At the same time, AI mandates are arriving earlier in the GCC maturity journey, increasing the need for infrastructure, data and governance capabilities at an earlier stage.

The report identified structural challenges behind stalled AI pilots. Production data is often more complex than controlled test environments, governance is frequently introduced after development, and solutions built outside common enterprise platforms can be difficult to integrate at scale.

Agentic AI raises infrastructure demands

The economics of AI also change significantly when systems move into production. According to the report, agentic workflows can consume between 10,000 and 500,000 tokens per workflow, compared with around 1,000 to 2,000 tokens for a standard chat interaction.

The report therefore argues that GCC leaders need to make infrastructure decisions at the workload level early in the development process, taking compute, token consumption, tooling and reskilling costs into account.

Four forces reshaping GCCs

The report identifies four key levers for the next phase of GCC evolution: building functional AI capabilities, planning AI architecture ahead of production, taking ownership of markets and business outcomes, and redesigning the workforce.

It calls for AI-enabled workflows to move beyond isolated experiments and become embedded in core business functions. It also recommends treating data readiness, compute, security, governance and economics as interconnected infrastructure decisions.

Workforce transformation is another major focus. The report estimates that 55% of routine GCC work is already exposed to AI-driven automation, while 60% of the workforce will require reskilling by 2030. It argues that the shift will require redesigning roles around engineering, product development and business problem-solving rather than relying solely on incremental AI training.

Balancing owned and managed infrastructure

The report also proposes a framework for deciding which AI workloads should be operated on owned infrastructure and which can be handled through leased or managed environments.

Workloads involving sensitive data, regulatory exposure, business-critical processes or high and predictable usage may require greater control, while lower-risk exploratory workloads could use flexible leased models.

For regulated or proprietary data, the report introduces a “Sovereign Sandbox” model that allows GCCs to experiment in a contained environment before moving workloads into production.

Manish Gupta, President and Managing Director, Dell Technologies India, said the GCCs that build strong foundations in data, infrastructure and governance will shape how their organisations use AI globally.

Sidhant Rastogi, President, Zinnov, said the GCC model is moving from a focus on scale, talent and capability towards greater ownership of products, platforms, markets and measurable business outcomes.