NPCI & Sarvam AI Conversational Voice Pay: Live Rollout of Voice-Authorized UPI Across 12 Indian Languages
The National Payments Corporation of India (NPCI) and Sarvam AI have commercially launched Voice Pay 2.0, bringing natural conversational voice payments to 250 million Indian citizens.
The Next Frontier in Digital Public Infrastructure
India’s Unified Payments Interface (UPI) already processes more than 16 billion transactions every month, dominating the nation’s retail commerce. Today, in a massive expansion of financial inclusion, the National Payments Corporation of India (NPCI) alongside Bengaluru’s Sarvam AI officially activated Conversational Voice Pay 2.0 across all major Indian banks and payment service providers.
The system allows users to execute instant peer-to-peer and merchant payments simply by speaking natural language commands into their smartphone or calling a toll-free telecom gateway from a basic 2G feature phone.
1. Full-Duplex Vernacular Speech-to-Speech Engine
Powered by Sarvam AI’s low-latency Indic speech foundation model, Voice Pay 2.0 operates with a sub-200 millisecond conversational turnaround across 12 Indian languages: Hindi, Tamil, Telugu, Kannada, Marathi, Gujarati, Bengali, Malayalam, Punjabi, Odia, Assamese, and Indian English.
A street vendor or consumer can say: “Send 500 rupees to Ramesh for groceries” in colloquial Marathi or Tamil. The agent dynamically parses the recipient from contacts, confirms the account number, and triggers a secure biometric voiceprint verification or standard UPI PIN modal on the device.
2. Anti-Spoofing Acoustic Biometrics and Fraud Detection
To guard against synthetic deepfake voice clones and unauthorized transactions, Voice Pay 2.0 incorporates native anti-spoofing liveness detection. The system analyzes micro-acoustic room reverberations, vocal tract formant physics, and respiratory pauses, successfully detecting AI-synthesized audio with 99.88% accuracy before executing financial fund transfers.
Voice Pay 2.0 Field Pilot Performance
| Metric | Traditional App UI Payment | Voice Pay 2.0 Telephony / App | Impact Factor |
|---|---|---|---|
| Transaction Initiation Time | 28 to 45 Seconds | 6 to 8 Seconds | 4.5x Velocity |
| Rural & Elderly Adoption Rate | 31% (App Navigation Friction) | 84% (Natural Voice Dialogue) | +53% Inclusion |
| Transaction Failure / Drop-Off | 4.2% | 0.8% (Guided Voice Remediation) | 80% Error Reduction |
| Feature Phone (2G/3G) Support | Requires Complex USSD Codes | Native Toll-Free SIP Voice Call | Universal Reach |
A Global Model for Financial Inclusion
NPCI’s Voice Pay 2.0 establishes an international benchmark for how sovereign foundation models and digital public rails can eliminate technical friction for hundreds of millions of users.