Dilip Asbe, managing director and chief executive of the National Payments Corporation of India (NPCI), said artificial intelligence would be an essential part of the next phase of digital payments growth in the country, as the Unified Payments Interface (UPI) seeks to increase its daily transaction count from more than 750 million transactions to over one billion.
Asbe said during an interview with TechCrunch at Mumbai Tech Week 2026, held last month, that NPCI, the Reserve Bank of India and the government could collaborate to use artificial intelligence to reach an additional half a billion users. The uses he cited include simplifying user onboarding, protecting existing customers, detecting fraud and intermediary accounts used in illicit operations, as well as making credit available to users and merchants with a digital track record.
Voice and multiple languages
Asbe believes that voice solutions and multilingual interfaces could make enrollment in financial services easier, but he stressed that the technology remains in its early stages because voice models need greater accuracy. NPCI launched an interactive voice-assistant-based system in 2023, but its adoption has not made significant progress so far. According to Asbe, voice could become an important element of the payments ecosystem if suitable use cases were found for it.
AI needs a regulatory framework
The report cited a race among startups and public companies in the United States to add artificial intelligence to financial services: Coinbase and Robinhood allow agents to execute trades on behalf of users, while OpenAI allows personal account data to be uploaded to ChatGPT to obtain financial advice. NPCI also showcased prototypes for agent-based commerce and payments in collaboration with Razorpay last year, but some of these capabilities have not been launched on a wide scale.
Asbe said that adopting AI-backed finance in India requires regulations and a clear framework that provide sufficient user protection and limit risks. He added that the system should be able, when a problem occurs, to review the instructions and consent the user granted to the AI agent.
An opportunity for small language models
The NPCI chief believes that banks, fintech companies and the rest of India’s financial sector can develop specialized small language models, noting that model differentiation will depend on the datasets available to them. According to Asbe, India’s financial ecosystem possesses extensive data that can be used to build models that are more accurate, specific and predictable.
NPCI launched a model called FIMI last year to process user disputes. Asbe said the model serves more than one million users for canceling mandates and resolving problems, and that its scope is expanding rapidly.
Competition among UPI apps
NPCI has been seeking for some time to strengthen competition among UPI apps, but market data indicates that Walmart-owned PhonePe and Google Pay account for more than 80% of the market share. The 30% market-share cap is scheduled to take effect on December 31, 2026, unless the regulator postpones the date again.