With the use of merchant data, the firm is creating its own lightweight AI model that will provide customized business insights.
AI-Powered Merchant Assistance
Vijay Shekhar Sharma, the founder and CEO of Paytm, stated that the fintech giant is preparing to introduce an AI-powered point-of-sale (PoS) device throughout its merchant network.
Sharma said that the new gadgets would serve as AI assistants for small firms, ready to provide merchants with real-time insights and recommendations based on their own transaction data, during his remarks at the Global Fintech Fest 2025 on October 7.
Sharma explained how the future gadgets would operate, saying, “If I own a fashion shop, I can ask, ‘How is my company going?’ or ‘Did I get money from the bank?’ and get responses in my local language.”
Lightweight On-Device AI Model
The new PoS machines will make this technology available to millions of retailers, he said, adding that the firm has been developing AI-powered agents that can communicate with users in normal language. “This AI will be the teammate or co-operator that every founder needs,” Sharma stated.
Sharma claims that the business is creating a smaller, more efficient language model of its own that will be trained using local merchant data on the device. According to him, this method will enable the gadget to use AI features without requiring expensive inference tokens or an internet connection.
“We are optimizing the size of large models and using merchant data from the device to create a small model that we have trained,” Sharma said.
AI as Paytm’s Growth Engine
The founder of Paytm went on to say that the company’s several verticals—from payments to devices—will be enabled by AI and be self-sufficient, much as Amazon and Amazon Prime are separate but linked ecosystems.
In light of regulatory obstacles, the fintech company is aiming to restructure and diversify its operations at the time of the relocation. Sharma emphasized the company’s shift to artificial intelligence as its next growth engine when he said, “I have worked in telecom, I have worked in payments, and now I will work in AI.”