The numbers that describe India’s GCC story in 2026 are impressive.
More than 2,100 centres. About 2.36 million professionals. $98.4 billion in annual revenue. Overall hiring up 12 to 15 per cent in the first half of the year. New centres are opening.
Existing ones are expanding. The narrative of India as the world’s capability hub appears intact.
Look more carefully at what is happening inside those numbers and a different picture begins to form.
Recent Dell Technologies-Zinnov research estimates that around 55 per cent of routine work inside Indian GCCs is exposed to AI-led automation. Nearly 60 per cent of the workforce will require reskilling by 2030. Meanwhile, ANSR’s latest analysis finds that nearly 65 per cent of new GCC positions now require AI skills, while demand for AI-related talent has risen about 45 per cent year on year.
GCCs, in other words, are continuing to hire.
They are also acquiring the ability to need fewer people for some of the work they already do.
The growth story and the displacement story are happening simultaneously. The problem is that only one of them is easily visible.
What agentic AI changes
Previous waves of automation in GCCs largely eliminated repetitive tasks. Humans moved towards slightly more complex work while the organisational structure around them remained recognisable.
Agentic AI changes the equation.
An AI agent does not merely produce an answer to a prompt. It can act across systems, execute multi-step workflows, make decisions within defined parameters and complete sequences of work that previously passed between several people.
That distinction matters because jobs are bundles of tasks.
Automating one task makes an employee more productive. Automating enough tasks inside a job changes the number of people an organisation needs to perform it. That is where task automation becomes role compression. And eventually, role elimination.
The structural consequence is beginning to appear in how GCC teams are being redesigned. The old pyramid, with a large base of junior execution talent supporting progressively fewer senior people, is becoming more diamond-shaped. Teams are getting leaner and more senior, while AI systems absorb more of the execution work underneath them.
This is not simply a productivity story.
It is a redesign of the organisational unit.
“GCCs are continuing to hire. They are also acquiring the ability to need fewer people for some of the work they already do.”
The roles disappearing and the people in them
The jobs most exposed are not abstractions.
The L1 IT support engineer handling the first line of technical queries. The KYC analyst working through identity verification. The manual QA tester running software test cases. The compliance reviewer processing documentation. The finance operations associate handling transactions, reconciliations and basic reporting.
These are not unimportant jobs. They are the jobs through which hundreds of thousands of Indian professionals entered the technology and services economy. They helped establish India’s original GCC proposition: educated talent capable of performing complex execution work reliably, at scale and at a cost attractive to multinational companies.
Agentic AI has arrived at precisely this layer of the workforce.
What is already happening at specific institutions illustrates the direction. HSBC is assessing changes that could affect around 20,000 roles, roughly 10 per cent of its global workforce, over the next three to five years as the bank uses AI to shrink its middle and back-office operations. Non-client-facing roles, particularly within its global service centres, are expected to be among the most affected. The assessment remains at an early stage and no final decision has been made, but the potential reductions include precisely the mechanism that makes AI-led displacement difficult to see: employees who leave may simply not be replaced.
HSBC is one institution. What makes the example significant is the mechanism.
The elimination of a role does not always arrive as a layoff.
A ten-person team can become seven without three employees receiving termination letters. Someone leaves and is not replaced. An open position disappears. Graduate intake is reduced. A team is consolidated with another. Work that once required several people is redistributed between fewer employees and an AI system.
No redundancy announcement is required.
The organisation may even continue hiring elsewhere.
That is why aggregate GCC employment can grow while particular categories of employment contract underneath it. The disappearance is easier to see in the job than in the headline headcount.
The replacement paradox
India’s GCCs are projected to add around 120,000 specialised roles this year, according to Nasscom-Zinnov projections.
That sounds reassuring. But the more interesting question is: 120,000 jobs for whom?
Nearly 65 per cent of new GCC positions now require AI skills. Demand is moving towards AI and machine learning engineering, data engineering, model validation, AI governance, cybersecurity, cloud architecture and the design and supervision of increasingly automated workflows.
These are not necessarily replacement jobs for the people whose execution work is disappearing.
A manual QA tester does not automatically become an AI engineer. A KYC analyst does not naturally progress into model governance. A finance operations associate whose reconciliation work is automated does not become a data scientist because the organisation has created a vacancy for one.
The organisation can therefore eliminate one kind of job while simultaneously struggling to fill another.
Its total headcount may remain stable. It may even increase.
But the people on either side of that equation are not interchangeable.
That is the paradox hidden inside the GCC growth numbers.
“That is where task automation becomes role compression. And eventually, role elimination.”
The reskilling assumption
Organisations understand this problem well enough to talk constantly about reskilling.
GCCs are investing in AI-enabled productivity, data capabilities, automation skills and new technical competencies. That is real and necessary.
But much of the reskilling conversation rests on an assumption that deserves more scrutiny: that somewhere above every disappearing job sits a more sophisticated job into which the employee can be trained.
Agentic AI may be breaking that assumption.
The problem is not only a skills gap. It is potentially a numbers gap.
If an automated workflow allows an organisation to perform the same work with substantially fewer people and AI systems, reskilling cannot by itself preserve every job. Even if every affected employee were willing and capable of learning new skills, the redesigned organisation would still need somewhere for those people to go.
And the destination roles are not necessarily adjacent to the jobs disappearing.
Reskilling a KYC analyst into an AI governance specialist is possible. But it is not the equivalent of teaching someone a new software tool. The cognitive distance is considerable, the training investment substantial, and the number of governance positions being created is unlikely to match the number of execution roles that automation can compress.
This is the conversation hidden underneath the reassuring language of workforce transformation.
Announcing a reskilling programme is not the same as providing a viable pathway.
The pyramid India built
There is a historical reason this matters particularly to India.
India’s GCC story was not built on AI engineering talent. It was built on the availability of a large, educated workforce capable of performing increasingly sophisticated execution work reliably and at scale. Cost mattered, certainly. But so did capability, English-language proficiency, technical education and the ability to build large teams quickly.
The pyramid that emerged was not an organisational accident.
A broad base of junior and mid-level employees performed execution, processing, testing and support work. Above them sat progressively smaller layers of specialist, managerial and strategic talent.
That structure did two things simultaneously. It made India’s GCC economics work. And it created employment at scale. It also created a career ladder. People entered through execution work, accumulated experience and moved upwards into more complex roles.
That model is now under structural pressure from two directions simultaneously. The execution work at the base is being absorbed by AI. The specialist work at the top is increasingly what GCCs are hiring for. The pyramid is not being supplemented with new skills at the top. Its architecture is changing.
The emerging diamond is more productive per employee. It is also a structure that does not require the same number of people. And that creates a problem reskilling alone cannot solve.
“The problem is not only a skills gap. It is potentially a numbers gap.”
What happens to the first rung
There is another consequence that deserves more attention.
If AI removes large amounts of entry-level execution work, organisations do not merely lose jobs. They may also remove the mechanism through which people historically acquired the experience required for more senior jobs.
The manual tester who eventually became a quality leader first learnt by testing.
The junior support engineer who later understood complex infrastructure began by solving simpler problems.
The finance professional who eventually exercised judgement over difficult exceptions first processed ordinary transactions.
Much of that work is repetitive. That is precisely why it is attractive to automation.
But repetitive work has also functioned as apprenticeship.
The GCC model did not simply employ people at the bottom of the pyramid. It developed people through it.
If agentic AI removes the first rung of that ladder, GCCs will eventually have to answer a different question: where will the experienced talent at the top come from?
Today’s productivity gain can become tomorrow’s capability problem.
What HR is not being asked
HR functions inside GCCs are already being asked to manage the talent transition.
Much of the attention naturally goes towards the acquisition side: where to find AI-skilled talent, how much to pay for it, how to compete for scarce specialists and how quickly existing employees can be trained.
The harder question sits on the other side of the transition.
What happens to the people whose jobs the redesigned organisation simply needs fewer of?
Not everyone can become an AI engineer. Not every execution role has an adjacent higher-value destination. And not every employee who successfully reskills will find that the organisation needs as many people in the destination role as it employed in the one being automated.
This makes workforce planning under agentic AI fundamentally different from running an upskilling programme.
HR will eventually have to distinguish between three populations: people whose jobs AI will augment, people who can realistically transition into genuinely different roles, and people for whom the redesigned organisation may simply have less work.
The third group is the one organisations have the least comfortable language for.
Yet it may be the group that determines whether the transition is managed honestly.
“The third group is the one organisations have the least comfortable language for. Yet it may be the group that determines whether the transition is managed honestly.”
The number that matters
India’s GCC story is not ending.
The number of centres can continue to grow. Revenue can continue to rise. India can become even more important to multinational companies as GCCs take on higher-value technology, product, analytics and strategic work.
None of that contradicts job elimination.
That is precisely the point.
An organisation can grow while employing fewer people to perform a particular kind of work.
It can open new roles while eliminating old ones. It can report record revenue while allowing vacancies in execution teams to disappear quietly. It can hire AI engineers in Bengaluru while deciding that a support team elsewhere no longer needs ten people.
The aggregate numbers will describe growth.
They will not necessarily describe what happened to the people whose jobs were removed from underneath that growth.
The most important question for India’s GCC sector may therefore no longer be how many jobs it creates.
It may be which jobs it no longer needs, what happens to the people who held them, and where the next generation will enter once the first rung of the pyramid begins to disappear.
GCCs are not shrinking.
The jobs that made them are.