[Deutsche Bank and Ant Financial Release White Paper: Agentic Payments in Wholesale Finance] According to DeepFlow TechFlow, on November 9th, Deutsche Bank and Ant Financial jointly released a white paper, "Agentic Payments in Wholesale Finance," discussing the application of AI agents in institutional finance and the governance framework required for responsible deployment. The white paper outlines three scenarios: treasury and liquidity management, where agents can autonomously predict liquidity gaps and surpluses, assess response plans, and execute internal transfers, investments, external financing, and foreign exchange operations; cross-border payment orchestration, which can balance exchange rates, payment channels, and execution timing to optimize costs, speed, availability, compliance, and settlement risks; and trade finance, which can interpret complex trade documents, identify data discrepancies, coordinate anomaly handling, and generate financing or payment instructions after conditions are met and approval is obtained. At the governance level, both institutions advocate for "bounded agent execution" rather than complete autonomy, adopting the IMF's three-layer reference architecture: intent and orchestration, control and authorization, and settlement, separating AI reasoning from transaction execution. Supporting requirements include explicit, limited, and revocable authorization limits, binding agent identities to accountable entities, and "Know Your Agent" (KYA) verification. The white paper also uses Ant Financial's agent platform, Anvita Flow, as a case study. Sabih Behzad, Head of Digital Assets and Currency Transformation at Deutsche Bank, stated in the foreword that value lies not only in automatically executing payments but also in improving how decisions are made, coordinated, and controlled. The president of Ant Financial emphasized in the foreword that agents need verifiable identities, explicit and revocable authorizations, and a separation of decision-making, control, and execution.