India is preparing to launch a payments infrastructure that would allow AI agents to make small purchases through UPI without requiring the user to approve every transaction. According to sources cited by Reuters, the new Unified Agent Protocol could be unveiled as early as next week at the Global Fintech Fest in Mumbai. If implemented, the project would make India one of the first countries to build national infrastructure for agentic payments.
At the initial stage, AI systems would be able to pay for routine, low-value purchases, such as groceries. Later, they could be entrusted with more complex tasks: buying goods once a specified price is reached, or investing according to pre-set rules. Users would be able to define spending limits and payment conditions, while the system would retain transaction histories and verify identities. Banks currently cap such earmarked amounts at Rs10,000 ($105) for up to 90 days, though those parameters could be revised.
The potential market is vast. UPI is the world’s largest fast retail-payments system by transaction volume. In August alone, it processed 24.51bn transactions worth Rs29.82trn ($314bn).
For BRICS, such infrastructure is particularly significant because UPI is being considered as one of the potential national nodes in the bloc’s future cross-border payments architecture. As we reported earlier, BRICS countries are already discussing ways to connect fast-payment systems and central-bank digital currencies in order to reduce the cost of international transactions and increase the use of national currencies.
India is among the most active participants in this effort. The Reserve Bank of India has previously proposed putting CBDC interoperability on the BRICS agenda; Brazil is exploring ways to connect Pix to foreign fast-payment systems; and the United Arab Emirates is developing its national card scheme, Jaywan. Under such a model, individual countries would retain their own domestic payment platforms, while technological gateways would link them together.
If UPI learns to transact not only with people but also with AI agents, it could eventually add another layer to BRICS cross-border payments: automated transactions and settlements between digital agents operating across different national payment systems.