The Quiet Rewiring: What AI Agents Actually Mean for India's Digital Workforce
A realistic look at the disruption ahead, beyond the hype and the panic.
For three decades, India built its economic ascent on a simple value proposition: a young, English-speaking, technically literate workforce that could do the world's digital work at a fraction of Western costs. From Bangalore's IT parks to Hyderabad's BPO campuses, millions of families rode this wave into the middle class. The call centre agent sending money home to Warangal, the junior developer in Pune paying off an education loan, the quality assurance tester in Chennai saving for a first apartment. These weren't just jobs. They were the visible proof that education plus ambition equalled upward mobility.
Now that equation is being rewritten. And the authors are AI agents.
The Numbers That Matter
India's IT-BPM sector employs approximately 5.8 million people directly and generates over $280 billion in annual revenue. It accounts for roughly 10% of GDP and more than half of the country's services exports. When NITI Aayog released its "Roadmap for Job Creation in the AI Economy" in October 2025, it put a number to the anxiety that had been building for years: in a worst-case scenario, the tech services workforce could shrink from 7.5 to 8 million to 6 million by 2031. The customer experience sector (call centres, support desks, chat operations) could contract from 2 to 2.5 million to 1.8 million.
But the same report carried a counterpoint: if India acts decisively, AI could create up to 4 million new jobs over the next five years. The question, as NITI Aayog CEO B.V.R. Subrahmanyam framed it, is stark: "Are you going to be minus 2 million, or are you going to go above 8 million to 10 to 12 million?"
Those aren't just employment figures. As Subrahmanyam pointed out, those 2 million jobs support an ecosystem of 20 to 30 million others, through household spending, remittances, local commerce, and the everyday economic activity that radiates outward from a monthly salary.
What's Actually Happening on the Ground
The disruption is not some future scenario being debated in conference rooms. It is underway.
At the India AI Impact Summit in February 2026, Vinod Khosla, the Indian-American venture capitalist, declared that IT services and BPO firms could "almost completely disappear" within five years. That timeline is almost certainly too aggressive, but the directional claim is harder to dismiss. A major Indian IT company has already reported nearly 20,000 job cuts in a single quarter, and industry analysts expect similar moves to repeat.
EY India's 2025 report found that call centre management faces an 80% potential productivity enhancement from AI, and software development around 61%. Content development, customer services, and sales and marketing all sit in the 40 to 45% range. When AI can handle the majority of volume work, the math changes. A US e-commerce brand paying $1.8 million annually for 300 offshore support agents can achieve equivalent or better resolution rates with an AI agent platform and 60 human escalation specialists. The CFO's decision is not agonising.
Yet fully autonomous operations remain, as NASSCOM's 2025 Technology Leadership Forum found, "more vision than reality for now." Human-in-the-loop systems are the current standard. AI handles the volume; humans handle the judgment calls, the emotionally charged interactions, the ambiguous queries that don't fit a pattern. The transition is not a cliff. It is a slope. But it is steepening.
The Domino Nobody Talks About
The first-order effect is the one that gets the headlines: fewer support agents, fewer junior developers, fewer data entry operators. The more consequential impact is second-order.
India's BPO and IT campuses are not islands. They are economic anchors for entire urban micro-economies. Electronic City in Bangalore, HITEC City in Hyderabad, Hinjewadi in Pune. Each of these tech corridors supports thousands of small businesses that exist because of the daily foot traffic and spending power of tech workers. The PG accommodations, the lunch delivery networks, the transport services, the phone repair shops, the retail outlets that line the road to an IT park. When 40% of a campus's headcount is replaced by software, none of these downstream businesses received a memo from the AI company. They just see fewer customers.
Then there is the geographic pipeline. Many of India's BPO workers come from Tier 2 and Tier 3 cities, places where a family invested in English-medium education and a bachelor's degree specifically because it led to a 3 to 4 lakh per year job in Bangalore or Hyderabad. When that job category contracts, the signal propagates backward: families recalculate, coaching centres lose students, the aspiration engine that powered a generation of social mobility starts to sputter.
This is not speculation. It is a pattern visible in every prior industrial transition. The difference is speed. The mechanisation of agriculture played out over decades. The shift from manufacturing to services took a generation. AI's disruption of knowledge work is compressing that timeline into years.
What the Optimists Get Right
But the panic narrative has its own blind spots.
First, India's BPO and IT companies are not passive victims waiting to be disrupted. Infosys has built Topaz, an AI platform processing 20 billion transactions with human oversight. TCS has developed AI-driven collaboration tools for its workforce. WNS Global Services, despite deep AI integration, grew revenue from $1.20 billion to $1.31 billion over three years while improving margins. The industry is reshaping, not collapsing. Companies that were selling labour arbitrage are pivoting to sell AI implementation, orchestration, and domain expertise. The job changes before it disappears.
Second, cheaper AI-powered support doesn't just replace existing jobs. It creates access for businesses that could never afford human support teams. The long tail of Indian SMBs, the D2C brand doing 5 crore in revenue, the Shopify merchant managing customer queries through WhatsApp. These businesses couldn't hire 10 agents. They can afford an AI agent. That new market creates its own ecosystem of implementation, customisation, and oversight roles.
Third, India's track record of adaptation is strong. After the 2008 recession hit IT hiring, engineering college enrolments didn't drop. They shifted toward new specialisations. The Indian workforce has historically been resourceful, mobile, and willing to retool. The 5,000 workers displaced from a BPO don't sit idle for years. Many will find adjacent work, move into AI operations, shift to sales or digital marketing, or enter the gig economy. The transition is painful, but it is not permanent paralysis.
Fourth, the NITI Aayog report itself outlines emerging roles that barely existed three years ago: AI DevOps engineers, sentiment intelligence analysts, ethical AI specialists, data centre controllers, quantum ML engineers. NASSCOM projects that India's AI talent pool needs to roughly double, from around 650,000 to 1.25 million, by 2027. That is a massive hiring wave, even if it demands different skills.
The Real Risk: Speed vs. Adaptation
The core question is not whether the transition happens. It is whether the creation of new work outpaces the destruction of old work, and whether the people displaced are the same people who can access the new roles.
History suggests that technological transitions ultimately create more jobs than they destroy. The printing press, the power loom, computerisation: each triggered panic, each ultimately expanded employment. But history has never seen a technology that targets cognitive and language-based work at this speed and across this many sectors simultaneously. The printing press took centuries to fully reshape the labour market. AI agents are doing it in a single product cycle.
The specific vulnerability of India's position is that its competitive advantage was built on the very tasks AI is best at automating: repetitive, language-based, rule-following, high-volume digital work. Countries whose economies are built on physical goods production or natural resources face a different, slower disruption timeline. India's exposure is front-loaded.
And the skills gap is real. Only 3% of Indian enterprises have sufficient in-house AI talent, according to EY India. Only 5.8% of white-collar job listings require AI expertise, per a World Bank analysis, and the vast majority are concentrated in the southern tech corridor of Bangalore, Hyderabad, and Chennai. The 23-year-old with a BCom degree from a Tier 3 city who was a perfect fit for a BPO role is not an obvious candidate for an AI engineering position. The reskilling gap is not just about content. It is about the speed and scale at which retraining needs to happen.
What Needs to Be True
For India to emerge stronger from this transition, several things need to happen in parallel. And they need to happen faster than is comfortable.
Education reform must move beyond adding "AI modules" to existing curricula. The fundamental model of training millions of students for scripted, process-following roles needs to shift toward problem-solving, judgment, and human-AI collaboration. This is easy to say and extraordinarily difficult to execute across a country of 1.4 billion people with wildly uneven educational infrastructure.
The IT services industry needs to complete its pivot from selling headcount to selling outcomes. The companies that survive will be those that can deploy smaller, more skilled teams augmented by AI to deliver better results than the old 300-person engagement model. This means accepting lower revenue per contract while pursuing higher margins, a transition that publicly traded companies with quarterly earnings pressure find structurally difficult.
Government policy needs to be calibrated rather than reactive. AI taxes and mandatory human staffing requirements would slow adoption and push work to other countries. But doing nothing while hoping the market self-corrects ignores the real human cost of displacement. The NITI Aayog's proposed National AI Talent Mission, coordinating reskilling across academia, industry, and government, is the right structural response. Whether it can move at the speed required is an open question.
And perhaps most importantly, the aspiration pipeline needs a new destination. For 30 years, the implicit promise to middle-class Indian families was: invest in education, learn English, get a tech job, join the modern economy. That specific pathway is narrowing. But the broader promise, that skill and effort lead to economic mobility, needs to be sustained through new pathways. AI implementation, digital entrepreneurship, India's growing domestic tech market, and the country's potential as a global AI services hub are all credible successors. But they are not yet visible enough, accessible enough, or trusted enough to replace what came before.
The Honest Forecast
India will not collapse under AI disruption. The apocalyptic predictions of BPO "disappearing by 2030" underestimate the inertia of enterprise adoption, the complexity of real-world deployment, and the resilience of the Indian workforce. Revenue in the sector continues to grow. Hiring hasn't stopped. It has shifted.
But neither will the transition be smooth or painless. A realistic estimate is that 1.5 to 2 million routine digital jobs will be automated over the next five years, with significant downstream effects on urban economies and aspirational pathways. Some of those displaced workers will move into higher-value roles. Many will face a difficult period of adjustment, lower earnings, and uncertainty. The downstream impact on families and small-town economies will be real, even if it doesn't make the headlines.
The difference between a manageable transition and a destabilising one comes down to velocity and intentionality. If AI adoption accelerates faster than reskilling programmes can absorb (which is the current trajectory) India will face a period of genuine social strain. Not catastrophe, but the kind of widespread anxiety and disruption that reshapes politics, migration patterns, and family decisions for a generation.
The 250 million young people entering India's workforce over the coming decades deserve honesty about this. The old playbook worked brilliantly. The new one is still being written, and the authors are racing against a deadline that keeps moving closer.
This article draws on data from NASSCOM's Strategic Review 2025, NITI Aayog's "Roadmap for Job Creation in the AI Economy" (October 2025), EY India's "The AIdea of India" report, and reporting from the India AI Impact Summit 2026.