India's second tech boom will be AI implementation
Most companies using AI see no bottom-line impact yet. Closing that gap is implementation work, the kind India's $282 billion tech industry was built on.
McKinsey’s latest State of AI survey found that 78% of organizations now use AI in at least one business function. The same survey found that more than 80% of respondents see no tangible impact on enterprise-level EBIT (operating profit) from their use of generative AI. Closer home, Nasscom’s Strategic Review 2025 estimates that only 10 to 15% of enterprise GenAI proofs of concept make it to full-scale production.
So companies have the models. What they don’t have is the working system: the model wired into their data, their approvals, their legacy software and their people. Boardrooms everywhere are asking the same question: how do we actually make this work in our company?
That question is where India built its first tech empire, and I think it is where we build the second one.
The hard part is the people and process work
Boston Consulting Group surveyed 1,000 executives last October and found that 74% of companies had yet to show tangible value from AI. When BCG broke down where the trouble came from, about 70% was people and process issues, 20% was technology, and only 10% involved the AI algorithms themselves.
That split matters for India. The 10% is the part Google, Microsoft, OpenAI and Amazon are spending billions on. The other 90% is integration, process redesign, change management, testing and keeping things running after go-live. It is slow, specific, unglamorous work, and it doesn’t come in a box.
India has sold this work before
In the late 1990s, global companies had computers and software but needed someone to build, integrate and run systems for them. Indian firms took that work and kept it. The industry grew along two streams. IT services belonged to engineers, who built and maintained software for global clients. ITES and BPO opened the door to commerce, arts and science graduates who could communicate well and run a process reliably.
Today that industry is estimated at $282.6 billion in revenue for FY2025, with $224.4 billion of it from exports and 5.8 million people on the payroll. That scale was built on implementation and operations. India did not invent the relational database or the ERP. It made them work inside other people’s businesses.
Clients are already paying for the AI version of this. Accenture reported $1.4 billion in generative AI bookings in a single quarter, the three months to February 2025. That money is going to the people who make AI work inside companies.
This wave lets more people in
The first wave split neatly. Engineers built systems, and everyone else ran processes inside them. AI blurs that line because its interface is increasingly plain language. You still need serious engineers for integration, security and data. But a lot of the value now sits with people who understand a domain well enough to decide what an AI agent should do, what it must never do, and how to check its work.
A commerce graduate who knows how loans actually get approved can design and supervise an AI-assisted credit workflow. Someone who understands brand and audience can run an AI content pipeline with judgment. That is the BPO stream moving up, from executing a process to designing and supervising one.
The real risk is selling it the old way
The strongest objection to my argument is a serious one. AI is very good at exactly the work Indian services firms bill for: writing code, testing it, answering support tickets, processing documents. If a client can do with ten people and an AI what used to take a hundred, a business priced on headcount shrinks.
I agree with that. Labor arbitrage was the first wave’s pricing model, and it won’t carry the second. The opportunity is for firms willing to price on outcomes, invest in domain depth, and own the result in production. Firms that try to sell AI implementation by the seat will find their own product undercutting them.
That is also why I think this is a founder’s opportunity as much as an incumbent’s. Large firms have the trust and the client lists, but they also have the most headcount revenue to protect.
What “two trillion” means
I call this a two-trillion-dollar opportunity as a sense of scale, not a forecast. McKinsey estimated in 2023 that generative AI could add $2.6 trillion to $4.4 trillion a year to the global economy. None of that shows up on a company’s books until someone does the implementation work. Ten percent of the low end of that range is close to the size of India’s entire tech industry today.
India can be a buyer of AI tools made elsewhere, or it can be the place the world goes to make them work. We have done the second thing once already.
So the question for Indian founders is a practical one: which industry do you know well enough to make AI work inside it, end to end, and be paid for the result?