Deutsche Bank, Ant Publish Agentic Payments Paper
A joint white paper outlines how AI agents could handle treasury, cross-border payments and trade finance under bounded authority.
Updated on Nov. 9, 2026 in Artificial Intelligence
Deutsche Bank and Ant Digital Technologies have jointly published a white paper titled Agentic Payments in Wholesale Finance. The paper argues for bounded agentic execution rather than unrestricted autonomy, using Ant's Anvita Flow platform as an implementation case study.
Why it matters
Wholesale financial institutions are weighing how much authority to delegate to AI agents that could move money. The paper proposes safeguards, including revocable authorization limits, agent identity binding and Know Your Agent verification, as a framework for that decision.
The white paper describes a three-layer architecture of intent and orchestration, control and authorization, and settlement that separates AI reasoning from transaction execution. It draws on a reference architecture proposed by the International Monetary Fund, but no performance specifications for the Anvita Flow platform are given in the source material.
The players
Deutsche Bank
Deutsche Bank is a German multinational investment bank and financial services company headquartered in Frankfurt with a large global wholesale banking business.
Ant Digital Technologies
Ant Digital Technologies is the technology arm of Ant Group, the Hangzhou-based fintech company that operates Alipay and related financial services.
The details
The paper names treasury and liquidity management, cross-border payment orchestration and trade finance as the areas where AI agents are most relevant. It recommends explicit, bounded and revocable authorization limits, robust agent identity binding and Know Your Agent verification protocols as safeguards against unrestricted autonomy.
Timeline
November 9, 2026: Deutsche Bank and Ant Digital Technologies release the white paper Agentic Payments in Wholesale Finance.
The Tech Race
The paper applies a three-layer reference architecture proposed by the International Monetary Fund to wholesale banking, extending that framework from policy discussion into a bank-facing design. Ant's Anvita Flow platform is presented as a working example, positioning the firm alongside large banks in a contest over agentic financial infrastructure.
For consumers, the practical effect of agentic payments would depend on whether banks adopt bounded authorization and identity checks before AI agents can initiate transfers. Until such controls are deployed, the paper's proposals remain design recommendations rather than changes to how payments currently work.
The takeaway
Autonomy in financial AI is being framed as a question of bounded, revocable permission rather than all-or-nothing trust. Institutions weighing agent deployment may find the paper's authorization, identity and verification safeguards a useful checklist.
Further reading
For broader context on how AI agents are being deployed across industries, see Artificial Intelligence coverage.
Source note: This article includes information reported by Macau Business.
