Automation, Indian Services Employment and the Honest Forecast Nobody Wants to Give

Every year for roughly two decades, India's largest information technology services firms, Tata Consultancy Services, Infosys, Wipro and their peers, hired tens of thousands of fresh engineering graduates each quarter, trained them for months in standardised bootcamps, and put them to work on tasks that, while not glamorous, were reliably remunerative: writing routine code, testing software for defects, migrating legacy systems, and staffing help desks for global clients. This pipeline did more for Indian social mobility than almost any other single institution outside of government reservation policy, turning graduates from small-town engineering colleges with no family history of white-collar employment into first-generation members of a genuine urban middle class. The question now facing the sector, and by extension a meaningful share of India's aspirational workforce, is whether generative AI is quietly dismantling the bottom rungs of that ladder.

What the hiring data is already showing

The evidence, while not yet conclusive, points in a consistent direction. Campus hiring numbers at India's largest IT services firms have fallen substantially from their 2021-22 peaks, when pandemic-driven demand for digital transformation services pushed hiring to record highs; several major firms have since reported flat or declining fresher intake even as revenue has grown, a divergence between headcount and output that is precisely what productivity-enhancing automation should produce. Nasscom's own strategic review documents acknowledge that generative AI coding assistants, from GitHub Copilot to enterprise tools built on large language models, are measurably reducing the time required for routine coding and testing tasks, the exact category of work that formed the bulk of entry-level assignments. None of the major firms describe this publicly as job elimination; they describe it, accurately as far as it goes, as a shift toward higher-value consulting and AI implementation work. But a shift toward higher-value work by definition requires fewer people at the lower rungs, and the fresh graduates who would have filled those rungs are the ones absorbing the adjustment.

Why this is not a repeat of past automation scares

It is worth being disciplined about what makes this moment different from earlier automation anxieties in Indian IT, because the sector has weathered predictions of its demise before, from the Y2K-driven boom eventually normalising to fears that cloud computing would eliminate infrastructure management jobs, which instead created new categories of cloud engineering work. Generative AI differs because it targets cognitive, judgment-adjacent tasks rather than purely mechanical ones, and because the tools improve continuously without requiring the multi-year retraining cycles that earlier technology shifts demanded of the workforce. A coding assistant that writes serviceable boilerplate code today will very likely write more sophisticated code within eighteen months, on a trajectory that previous automation waves in Indian IT did not follow at comparable speed.

The counter-argument that deserves real weight

The strongest case against alarmism here is that India's IT services demand is not fixed; it expands with global enterprise digitisation, and firms that can deliver AI-augmented services more cheaply than competitors elsewhere may actually win more business, generating employment through volume even as headcount per project shrinks. This is a real historical pattern: productivity gains in IT services over the past two decades did not produce net job losses in the sector, because falling per-unit costs expanded the addressable market for outsourced digital work faster than automation reduced the labour needed per unit. It is entirely plausible that Indian IT services firms, armed with AI tools, will capture new categories of global enterprise work that were previously uneconomical to outsource, from AI model fine-tuning services to industry-specific AI agent deployment, categories that barely existed three years ago and that could absorb a meaningful share of displaced entry-level capacity into new mid-tier roles.

Why the honest forecast still leans toward disruption at the bottom

Even granting that counter-argument its full weight, the composition of the workforce likely to benefit from this expansion differs from the composition that filled entry-level rosters in the past. New AI-adjacent roles increasingly require specific skills, familiarity with prompt engineering, model evaluation, and AI system architecture, that are not uniformly distributed across India's roughly 1,500 engineering colleges, most of which continue to produce graduates trained on curricula that have not meaningfully updated to reflect this shift. The result is likely to be a bifurcation: engineers from top-tier institutions and those who proactively reskill will find abundant, well-paid opportunity in AI-augmented services, while graduates from the thousands of tier-three and tier-four colleges that fed the traditional entry-level pipeline will find that pipeline substantially narrower than it was for the cohort just five years ahead of them.

What policy has not yet reckoned with

This bifurcation carries consequences well beyond the IT sector itself, because engineering education expansion in India over the past fifteen years was substantially a bet on exactly the entry-level IT services demand now thinning, with state governments and private trusts building capacity on the assumption that a steady supply of coding and testing jobs would absorb graduates regardless of the quality of instruction. Neither the All India Council for Technical Education nor state higher education departments have yet produced a serious, publicly available plan for what happens to this capacity if that assumption weakens, and the political economy of engineering education, with thousands of colleges and the employment they represent, makes any admission of overcapacity institutionally difficult.

The honest forecast, then, sits between the two extremes usually offered in this debate. India's IT services sector will not collapse; its revenue base is diversifying and its largest firms are genuinely capable of repositioning around AI-augmented offerings. But the specific social function that sector performed for two decades, absorbing large numbers of moderately trained graduates into stable, upwardly mobile employment, is narrowing in ways that will not reverse. Treating this as a temporary blip to be waited out would be a more serious policy failure than treating it, cautiously and without panic, as a structural shift that Indian higher education and labour policy need to start planning around now.