Enterprises have spent the past two years deploying AI agents faster than they can keep track. Pilots have multiplied across procurement, customer service, coding, and IT operations, often built on different models, cloud platforms, and vendor stacks.
The result, increasingly, is agent sprawl—pockets of automation nobody can fully see, govern, or shut down if something goes wrong.
That governance gap has created an entirely new product category among IT services and platform vendors—the ‘agentic control tower’. It is a platform-agnostic layer that discovers every AI agent running across an organisation's estate, observes what it is doing, enforces permissions, and can intervene or shut it down if it misbehaves.
What was once framed as a compliance afterthought is now being sold, and increasingly built in-house, as core infrastructure. Nowhere is this shift clearer than in India's Global Capability Centre (GCC) ecosystem, where captive units run by global banks, insurers, and enterprises are being asked to become the AI control layer for their parent organisations.
GCCs Turn Enterprise Nerve Centres
India's GCCs, long seen as cost-arbitrage back offices, are now being explicitly rebuilt around this control-tower logic. A joint Nasscom-Zinnov report on India's GCC landscape found the sector, now numbering 2,117 centres employing roughly 2.36 million people, is shifting from being a pure delivery engine towards what the report calls an "enterprise nerve centre."
“The GCCs best placed to benefit are those that move beyond executing tasks to owning platforms, governance, and outcomes on behalf of their global parent, rather than simply adding headcount,” comments Pari Natarajan, Co-founder and CEO of Zinnov, in a conversation with AIM.
He highlights that over half of the work traditionally handled inside Indian GCCs sits in categories exposed to automation, meaning centres that fail to move up into governance and oversight roles risk becoming smaller, not larger, as AI scales.
Governance for Legacy Estate
DXC Technology, a US-based IT services and consulting company, has positioned its Oasis platform squarely within this category.
Rob Le Busque, President of Asia Pacific & Japan at DXC Technology, explained that Oasis is designed as a platform-agnostic "agentic control tower" that can sit not only atop new digital transformation programmes, but directly on top of legacy managed-services environments.
The pitch is aimed squarely at DXC's core client base: large, security-conscious organisations in public sector, banking, insurance, aerospace, and defence that cannot replace legacy infrastructure outright. Rather than forcing a costly re-platforming exercise, Oasis is meant to layer agentic workflow, observability, and control on top of what already exists, giving clients a way to extract AI-driven efficiency from old systems.
“Oasis is designed as an agentic control tower, which is platform agnostic, and has a really sophisticated capability to not only sit at the heart of a digital transformation journey for clients, but also to sit on top of legacy applications and legacy managed service environments,” notes Rob Le Busque, President of Asia Pacific and Japan at DXC Technology.
Finance GCCs Combat Tighter Oversight
Banking, financial services, and insurance is one sector where this transition is most concrete.
Manoj Marwah, Partner and GCC Market Leader for financial services at EY, observes that global financial institutions are actively looking for simpler models and tighter oversight, allowing for AI, data, and risk teams to work in sync.
“This description matches the control-tower blueprint closely rather than scattering AI governance across business units. Banks are increasingly centralising discovery, monitoring and risk sign-off for their AI agents inside their India GCC, treating it as the oversight layer for AI deployed globally, not merely a delivery location,” he remarks.
EY's broader research into agentic AI-first GCCs found a shift from linear, approval-based workflows towards event-driven models where autonomous agents handle routine execution and escalate only genuine exceptions requiring human judgement, with the GCC providing the governance, security, and audit infrastructure.
A Converging Bet
What unites DXC's Oasis and the reinvention of India's GCCs is a shared wager that the most defensible, monetisable layer of the AI stack will not be the underlying models themselves but the orchestration and governance layer that sits above them.
Global software players such as ServiceNow have reached the same conclusion, building their own "AI Control Tower" with discovery, observability, governance, security, and cost-measurement capabilities spanning multiple cloud and agent ecosystems.
For GCCs and India-headquartered engineering hubs of global services firms like DXC, the opportunity is to translate complex, regulated, and legacy-heavy enterprise environments into central governance capability centres, at exactly the moment enterprises most need one.
The risk, of course, is execution. Zinnov's research has found that more than 70% of GCCs still lack a structured framework to measure the return on their AI investments, even as they position themselves as control towers to oversee enterprise AI risk. Whether that risk is curbed remains to be seen.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.