India’s UPI bets on AI to bridge a billion-user gap
MUMBAI — The next half-billion users of India’s Unified Payment Interface may never swipe a card or type a bank account number. They might never even touch a screen. That is the bet Dilip Asbe, the head of the National Payments Corporation of India, laid out in a recent interview.
Speaking at Mumbai Tech Week 2026 last month, Asbe said artificial intelligence would drive the next wave of growth for UPI, a network that already processes more than 750 million transactions every single day. The target is a billion daily transactions.
The gap is real. And the people needed to close it are not the urban smartphone crowd already on the network. They are rural, often less literate, and harder to reach through conventional banking channels.
Asbe framed AI as the tool for both sides of the same problem: pulling new users in while keeping fraud out. “AI will be used very effectively when we look at the next wave of UPI, and that includes all aspects, including reaching new users,” he told TechCrunch.
He added that NPCI, India’s central bank, and the government are working together on this. The stakes are concrete. UPI has become the backbone of retail payments in India.
Street vendors take payments through it. Bill payments run on it. The network, launched by a consortium of banks, now handles transactions that would have been cash-only a decade ago.
Cheap smartphones and cheap data fueled that growth. But those easy gains are mostly exhausted.
Rural India is a different terrain. Literacy rates dip. Digital fluency drops.
Fraudsters have already begun targeting the system. Asbe said AI must be used “to protect our current citizens, to find fraud, and to fi” — his words cut off in the published interview, but the direction is clear.
Fraud is not an abstract risk. It is the fastest way to kill trust in a payment system. If new users — especially those with limited experience — lose money to scams, they will not come back.
They will revert to cash. That would stall UPI’s growth and undermine years of policy work by the government and the central bank. So AI is being asked to do two jobs at once.
On the onboarding side, it could help build voice-based interfaces, local language support, and simpler authentication for users who cannot read a screen. On the protection side, it could flag unusual transaction patterns in real time, block suspicious activity, and reduce false declines that annoy legitimate users.
NPCI oversees the entire network. Asbe’s comments signal a strategic shift. The organization is no longer just scaling up merchant adoption or adding bank links.
It is looking at AI as the core infrastructure for the next phase. The numbers back the urgency.
At 750 million daily transactions, UPI is already one of the world’s busiest payment networks. Getting to a billion means adding 250 million transactions a day — roughly the entire daily volume of some developed countries’ payment systems. That growth cannot happen without technology that adapts to users rather than forcing users to adapt to it.
India’s digital payment share has risen steadily. But the next half-billion users will not come from the same places as the first half-billion. They are in smaller towns, in villages, in households where the only smartphone is shared among family members.
They are people for whom a failed transaction is not a minor inconvenience but a real loss. Asbe’s interview made plain what is at risk.
UPI has become a national project. If AI can bridge the gap to a billion daily transactions, the system becomes nearly universal. If it cannot, the network risks plateauing — and leaving millions outside the digital economy.
