At a startup event in Bangalore recently, everyone was talking excitedly about AI. People were discussing new tools, how work was becoming more productive, and all the creative things AI could do. But then, during one of the panel discussions, someone asked a straightforward question that made everyone uncomfortable: "If AI starts doing most of our work, how many of us are actually prepared for what will be left?" That moment reminded everyone that while we're all excited about AI, there's a problem India needs to face seriously.

A Skills Gap Too Big to Ignore

India's tech boom has been one of our biggest economic success stories. But it's hiding a serious problem! We don't have enough people with the advanced skills that the next phase of the digital economy will need. The Mercer | Mettl India Graduate Skill Index says that only 42.6% of Indian graduates who apply for jobs are actually employable according to industry standards.

The situation is even worse when it comes to advanced AI talent. According to NASSCOM and talent firm Xpheno, India has less than 2,000 senior engineers who can actually build core AI products. For a country that has one of the world's largest tech workforces, this number is shockingly small.

This skills gap really matters because of how our economy is structured. More than half of India's GDP comes from services - IT, BPO, financial services, and other data-driven sectors. These sectors are built on processing large amounts of information - exactly the kind of work that AI can automate quickly and accurately.

McKinsey's global research shows that up to half of all work activities can already be automated using today's technology. In India's service-heavy economy, this proportion is probably even higher. And unlike manufacturing automation, which needs expensive machines and years of setup, AI systems can be rolled out in just weeks through software updates and cloud deployment. This speed makes job displacement happen much faster and makes it harder to protect workers.

I heard this directly from a startup founder in Bangalore whose team went from fifteen people to just four after they leveraged the AI code generation tools. Another professional working in social media content moderation told me how AI took over tasks that once needed large teams - like flagging policy violations and scanning for harmful content - without needing to understand cultural context the way humans do. In both cases, the change was fast, final, and couldn't be reversed.

Where New Jobs Are Not Enough

It's true that AI is creating jobs as well as eliminating them. Bain & Company says that India could see more than 2.3 million AI-related job openings by 2027. But these new roles are not accessible to most people who lose their jobs as they need very different skills.

There's also the question of numbers. Many of the new AI jobs replace larger teams with smaller, highly skilled groups. One experienced AI engineer can now do work that previously needed ten or more people. This is great for productivity but terrible for employment numbers.

The challenge is not just retraining but bridging a huge skills gap. Many corporate "reskilling" programs only work for people who already have a strong technical background.

For a software tester working on maintenance projects or a BPO agent doing repetitive tasks, the jump to AI development is enormous. Without focused, long-term support, most people will not be able to make this transition.

The early effects of AI are most visible in Bangalore and other tech hubs. Routine coding, basic testing, and IT infrastructure support are already being automated with tools like GitHub Copilot and Cursor etc. BPOs are seeing AI chatbots and document-review systems replace large portions of their workforce. In finance, algorithms are handling credit scoring, fraud detection, and even personalized financial planning, cutting deep into analyst roles that people once thought were secure.

Manufacturing as a Balance and the "Trump Effect"

While AI dominates discussions in services, manufacturing is still an area where human involvement is harder to replace. Complex assembly, quality control, and solving problems on the factory floor are not yet fully automatable. The Indian government's "Make in India" program and Production Linked Incentive (PLI) schemes are bringing significant investment in electronics and automotive production. Official targets show electronics manufacturing output is on track to reach $300 billion by 2026, with large plants like Samsung's in Noida and Foxconn's in Chennai already employing tens of thousands of people.

This could be an important balance to AI-driven job losses in services, especially if supply chains continue to move away from China. But the political reality has already become challenging: Donald Trump has imposed 50% tariffs on Indian goods - starting with 25% in August 2025 and doubling it to 50% by late August due to India's purchase of Russian oil. India's textile industry alone expects $5 billion worth of business to move away from India in the next few months due to these tariffs. The tariffs affect tens of thousands of workers and could potentially shave billions off India's earnings. The bigger challenge is that Trump's focus has shifted from the earlier "China plus one" strategy to bringing manufacturing back to America itself. His administration has explicitly stated they want America to be "a hub for manufacturing and production," which means Indian manufacturing may not benefit even if companies move away from China - they may be (arguably) looking to set up in the US instead.

What's certain is that manufacturing won't be a direct replacement for the kinds of service jobs AI is automating. Moving a displaced IT worker into a modern factory role will need major retraining and, in some cases, completely rethinking career paths. But it does offer a way to diversify, something India urgently needs.

Moving Faster Than the Technology

The real urgency is taking action before the wave peaks. If the past twenty years were about scaling up the IT services model, the next decade has to focus on diversity - both in the economy and in workforce skills. That means investing not only in AI capabilities but also in the uniquely human skills AI can't easily copy. Creativity, complex problem-solving in unclear situations, emotional intelligence, and the adaptability needed in physical, real-world environments.

For tech professionals, understanding how AI actually works - not just how to use its tools - will be essential. For business leaders, the choice isn't simply whether to automate but how to use AI in ways that work alongside human strengths. For policymakers, the challenge is changing education and training systems to prepare millions for an AI-heavy economy, rather than focusing only on a small elite group.

In Bangalore, I've seen the full range in a single office, a small AI team building a model that replaces an entire business process, a handful of employees rushing to learn new programming skills and the majority quietly continuing their work, aware that changes are coming but unsure how to prepare.

For now, the most realistic view is that in the short to medium term, India will see a net loss of jobs to AI. That's not being negative - it's understanding how quickly automation is advancing in the specific sectors where we employ the most people. But we've beaten tough odds before. We built a $200 billion IT industry in two decades, created one of the world's most successful digital payments system, and connected a billion people to mobile networks. If we treat AI as an urgent national challenge - investing in skills, balancing innovation with worker protection, and betting on multiple growth paths - we can shape its impact rather than let it shape us.

The technology isn't slowing down. The question is whether we can move faster.