The technology that powered India's payments boom is reaching a new inflection point, and artificial intelligence will define what comes next — that is the view of Dilip Asbe, managing director and chief executive of the National Payments Corporation of India (NPCI), the body that runs the Unified Payments Interface.
The case for AI
"AI will be used very effectively when we look at the next wave of UPI, and that includes all aspects, including reaching new users," Asbe said, as reported by TechCrunch. He pointed to four areas: drawing in the next cohort of users, detecting fraud and the "money mule" accounts used to move illicit funds, extending credit to people and merchants with a digital track record but no formal credit history, and building voice-based, multilingual tools that make payments easier for first-time users.
The ambition behind that is large. UPI already processes around 750 million transactions a day, and NPCI's stated goal is to pass one billion — a leap that means reaching hundreds of millions of people still outside the digital economy, many in rural areas or unfamiliar with smartphone apps.
A system of extraordinary scale
Launched in 2016, UPI now handles volumes that dwarf comparable systems elsewhere; India is regularly cited as accounting for a large share — by some measures the largest — of the world's real-time digital payments. Two private apps, PhonePe and Google Pay, together account for more than 80 percent of UPI volume, a concentration regulators have tried to ease; a 30 percent market-share cap per app is due to take effect at the end of 2026 unless deferred again.
What AI could do — and the risks
The most immediate use may be fraud detection, where machine-learning models can flag suspicious behavior faster than human review as volumes climb. On access, NPCI has already launched a voice assistant, Hello UPI, and a language model called FIMI for dispute resolution that it says serves more than a million users; Asbe sees these as templates for AI that can onboard and support users across India's many languages. Credit is a third frontier: the payment records of small merchants and gig workers could let AI underwrite loans that banks have avoided.
The optimism comes with genuine risks. Folding AI into payments raises the threat of deepfake-enabled fraud — synthetic voice or video used to impersonate account holders — a tactic scammers have already deployed elsewhere. It also intensifies data-privacy questions, since training models on payment behavior means pooling sensitive financial data; India's 2023 data-protection law, still being implemented, will shape what is permitted. Asbe also floated an opportunity for Indian firms to build small, payment-specific AI models that can run on low-end devices and weak networks — important for the rural users NPCI most wants to reach.
A model others watch
India's UPI has become a reference point for governments modernizing their own systems, with cross-border links to Singapore's PayNow and frequent comparisons to Brazil's Pix. Whether AI can carry the system from 750 million toward a billion daily transactions, and beyond, will be one of the more consequential experiments in global financial technology in the years ahead.
