Paytm Turned Profitable and Now It Is Going All In on AI

Paytm's first full year of profitability has given the company something it has not had in a while, and that is confidence. It is now channelling that confidence into one clear direction...artificial intelligence (AI).

Speaking during the Q4 FY26 earnings call, Vijay Shekhar Sharma, CEO of One97 Communications, the parent company of Paytm, said that while the company would pursue inorganic growth opportunities, every new investment would be ring-fenced for AI.

With a cash balance of ₹13,315 crore as of March 2026, Paytm has the resources to back that commitment.

Point to note: Paytm's net profit for the January-March quarter came in at ₹183 crore, against a loss of ₹545 crore in the same period last year. Operating revenue climbed 18.4% year-on-year (YoY) to ₹2,264 crore.

Rent, Not Build

The company has no plans to build its own data centre, preferring instead to rent capacity and run its models on top of it.

"There is enough capex in the US…We do believe there is an opportunity for us to invest in AI—we can rent a data centre somewhere and then run our own model on top of it," Sharma said.

The broader ambition is to automate all services across the Paytm platform, from Paytm Money to Paytm Check-In, for both consumers and merchants.

Paytm is, however, not starting from scratch. According to its investor presentation for the March 2026 quarter, the company is already deploying AI across its operations, building applied models for payments intelligence, fraud prevention, merchant onboarding and collections. It also credited its AI-powered fraud detection and collections engine as a direct contributor to the FY26 profitability turnaround.

Internally, AI agents have been embedded across engineering functions — coding, testing and deployment — compressing development timelines and reducing costs.

For merchants, AI features on Soundbox devices now offer business insights, customer notifications and support. On the consumer side, the technology is being used to acquire higher-quality users at lower costs and drive better monetisation through personalisation.

Betting on Agentic AI

The most forward-looking piece of Paytm's AI strategy is its experiment with agentic interfaces. Its travel ticketing platform, Check-In, is currently serving as a test bed for this technology.

"I see the agentic interface as a rejuvenated new opportunity for Paytm to gain consumer share dramatically in a number of categories. This is our experiment category," Sharma said. He added that funnel conversion on agentic interfaces runs seven to eight times higher than on a standard interface.

Sharma also reflected on the broader AI moment, suggesting that the rapid pace of change levels the playing field. "I feel lucky that we did not dump a lot of money earlier, because in the AI world, everything resets," he said. "Anybody that is not investing in AI is an opportunity for us — that person's customer."

The strategic pivot to AI follows a period of regulatory turbulence. The Reserve Bank of India last month cancelled the banking licence of Paytm Payments Bank, citing management concerns, though the payments bank had already been barred from offering services since 2024, leaving it effectively non-operational.

Paytm has since restructured its payments business around UPI partnerships with Axis Bank, HDFC Bank, and SBI, operating under the third-party application provider model.