The Monetary Authority of Singapore (MAS) is preparing a controlled environment where artificial intelligence agents can carry out corporate treasury tasks using tokenized deposits, with limits set by companies and banks. MAS Managing Director Chia Der Jiun presented the Agentic Money Sandbox at the Singapore FinTech Festival on November 19, while MAS and the Future of Finance Institute had announced on November 18 that the institute would build the sandbox.
The initiative combines AI agents with programmable money across two use cases. In corporate treasury management, a company grants the mandate, and participating banks set controls and acceptance terms. The agent monitors liquidity positions and due payments, decides where funds should move, and uses tokenized deposits to execute permitted actions. In the payments track, customer agents would execute payment instructions with digital money such as tokenized deposits or stablecoins, while the bank or payment provider verifies the agent’s identity and authority and applies risk controls before execution. Launch partners include DBS, OCBC, Standard Chartered and UOB.
Chia also linked the sandbox to the Safeguards for Agentic Finance at Runtime, or SAFR, protocol. The next phase would give financial institutions four pieces of information: who the agent is and who it is acting for, its authority, its mandate, and assurance about its origin and reliability. He cautioned against giving personal agents login credentials to financial accounts and raised questions about liability for incorrect agent transactions. Separately, Visa, Mastercard and Ant International are developing a Know-Your-Agent interoperability framework through BuildFin.ai. MAS has also issued S$100 million in tokenized MAS Bills settled against a Singapore dollar wholesale CBDC, and the SGD Testnet is set to become the SG Digital Ledger utility. Under MAS BLOOM, StraitsX, OCBC and Standard Chartered completed US dollar payments across two separate tokenized deposit networks in minutes.