“Clients no longer want PoCs,” Srini Pallia, CEO and MD of Wipro, said during the company’s Q2 FY26 earnings call last week. Dropped casually, that line is more than a corporate soundbite; probably an inkling of a turning point in Indian IT’s AI story.
Proofs of concept won’t pay the bills anymore. The hype around agentic AI launches and generative models that propelled the industry over the last two years is giving way to the next phase. Enough experiments have led to a stipulation that clients meet production-ready solutions, beyond ideas and demos.
The market is insisting on results such that the next few quarters may separate the promoters from the deliverers.
HCLTech is the first one to report the revenue directly from AI solutions, and it has crossed $100 million for the firm in this quarter. It is expected that other IT firms would follow the same in the next quarter, as they did not reveal it this time.
Q3: A Reality Check for Industry
For the last 10 quarters, Indian IT has sold the promise of AI rather than the outcome. Hundreds of AI agents, shiny platforms like Infosys Topaz, HCLTech AI Foundry, and LTIMindtree BlueVerse, and now Wipro Intelligence are touted as evidence of capability. But, the conversation is shifting now with companies focusing on a simple question about ROI.
Wipro’s statement points to a broader reality. In a world flooded with PoCs and pilots, businesses are tired of paying for exploration. They want measurable impact on operations, revenue, or cost structures. And all this comes after the warning signs from the MIT study that said 95% of pilots fail in production.
Wipro’s Pallia has repeatedly stressed operational deployment over experimentation. The firm claims to have built over 200 AI agents for clients. But the challenge ahead is proving that these agents actually generate value.
TCS also said in Q3 FY25 that it would see more generative AI PoCs move into production in Q3 FY26, and would possibly report it by the December results. Though this time, the company has not reported any concrete generative AI pipeline or projects, but a $6 billion investment in data centres over the next five years.
Infosys did not make any big AI announcements this quarter, but said that clients are increasingly spending on AI and the company has been one of the trusted AI partners for many. CEO Salil Parekh said in the earnings call that the company is already tracking AI revenue internally, and will start reporting it at the “right time”, which ideally would be in the upcoming quarters.
The company in its last quarter had reported building 300 AI agents for clients, which are now moving into production.
The coming quarters, starting with Q3 FY26, will be a reality check for the industry. Analysts and clients alike will track pipeline-to-revenue conversion. That metric, ignored in the last two years when AI was all about positioning, will now dominate boardroom conversations.
For instance, Wipro reported only 1.8% revenue growth in Q2 FY26, despite booking $4.7 billion in new deals and touting its Wipro Intelligence platform. Large as these numbers are, they are dwarfed by the market’s expectations around AI.
The deals need to turn into measurable improvement in client business, not just pipeline wins. Back in Q3 FY25, Wipro cloud chief Jo Debecker had said that over 50% of its AI PoCs have moved into production. The case might be even better now, but a clear distinction of AI revenue was not reported this quarter.
Signs of Life?
A warning is unfurling in another corner of the IT and consulting world. Accenture CEO Julie Sweet has already said that the returns from all of its AI investments over the years have been underwhelming. But despite this, the company reported another $5.9 billion in generative AI bookings this quarter, showing that there is no going back.
Similarly, HCLTech’s confidence is a wake-up call amidst the Indian IT firms.
HCLTech started working on over 200 AI PoCs in Q1 FY25. The firm slowly moved them into production and is now reaping benefits from it. CEO C Vijayakumar termed this quarter as a “standout quarter on every front” and revealed the firm’s advanced AI accounted for approximately 3% of its revenue.
Even Tech Mahindra said that it is building a 1 trillion parameter model with the IndiaAI Mission, though direct monetary impact from it is yet to be understood.
But apart from that, CEO Mohit Joshi said that the company is actively shifting more and more of its PoCs to production with TechM Orion Marketplace hosting over 200 AI agents. The company might also start reporting AI revenue in the coming quarters, hopefully.
This is the shift that will reshape how Indian IT firms are looked at over the next few quarters, instead of just claiming to be “AI-first”, or if actually AI is making a mark on the company’s profit and loss statements.
The stakes are high. By Q3, or at least Q4, every IT firm will be judged on the ability to turn AI pipelines into revenue streams.
The post The Death of AI PoCs at Indian IT Will be a Reality Check in Q3 appeared first on Analytics India Magazine.
