GFF 2025: This year, we will work in AI: Paytm’s Vijay Shekhar Sharma
Paytm founder Vijay Shekhar Sharma said the entity will roll out its SLM catering to merchants and businesses on its network

From a payments company to now working on Artificial Intelligence (AI)-led services, Paytm founder Vijay Shekhar Sharma said the future for the company will be focussed on “building in AI for 2025”.
Comparing the Indian AI ecosystem to a mild breeze compared to the whirlwind of AI-centred investments in the Bay Area in US, Sharma said India needs to build AI use cases for the country and should not be handing over its AI sovereignty to providers from other countries.
“There are deals worth $10 billion to $100 billion in AI (in the Silicon Valley) and none of us are bothered about it,” said Sharma, contrasting it with the mega fundraises in India by private companies, averaging at $1 billion.
Calling it an eye-opener, Sharma added that, going ahead, he will build out a separate entity to house AI innovations being developed by the company.
“In 2005, I worked in the telecom operator sector. In 2015, I worked in payments. In 2025, we will work in AI. Because, in this country, if someone is not working in AI, then that person is not working for the future,” said Sharma during a fireside chat with Yes Bank CEO Prashant Kumar at the Global Fintech Fest 2025 in Mumbai on Tuesday.
Also Read: GFF 2025: Why AI & credit infra could be key risks for Indian finance
“We are working on a Small Language Model (SLM) for businesses we work with. Not every business at the level of a kirana or a standalone store has the ability to hire a CFO who can advise them on how they can grow their business or what measures to take. The SLM will answer these queries for our merchant partners in their native language,” Sharma said on the sidelines of the event.
The merchant-facing app of Paytm will also roll out a voice-based AI agent to respond to their business-related queries. Sharma added that, beyond payments, Paytm’s AI brand will focus on other use cases such as embedded systems and devices, without mentioning details.
Use of SLMs in education, healthcare, agriculture advisories address sector specific issues. SLMs can be trained in different native languages, solving for access to AI services.
As part of the India AI Mission, the government of India has selected 12 startups to build indigenous LLMs and sector-focussed SLMs to cater to specific sectors and close the digital divide. These include the likes of IIT-Bombay consortium’s BharatGen, focussed on creating open-source language models across 11 Indic languages, CoRover’s BharatGPT Lite, which supports virtual assistants across 14 Indian languages and Gnani.ai, offering voice-first SLMs for Indian languages.
First Published: Oct 08, 2025, 11:21
Subscribe Now