Dev.to AI πŸ€– Ai πŸ‘ 0 πŸ“– 2 min read

HKMA Launches Project SAGE, a Governance Framework for Agentic AI in Finance

πŸ“… 02 November 2026 β€’ Hong Kong Monetary Authority The Hong Kong Monetary Authority (HKMA) has launched Project SAGE β€” Securing Agentic Finance with Governance and Enablement β€” to support the responsible development and

HKMA Launches Project SAGE, a Governance Framework for Agentic AI in Finance

Infographic cover

πŸ“… 02 November 2026 β€’ Hong Kong Monetary Authority

The Hong Kong Monetary Authority (HKMA) has launched Project SAGE β€” Securing Agentic Finance with Governance and Enablement β€” to support the responsible development and adoption of agentic finance in Hong Kong. The launch was announced on 2 November 2026 by Chief Executive Eddie Yue during Hong Kong FinTech Week. Agentic finance means financial services in which AI agents execute tasks and transactions on behalf of users, and the regulator notes that the industry is moving from internal adoption of agentic AI to customer-facing services β€” a shift that puts identity, delegated authority and accountability at the centre of the supervisory agenda.

Four focus areas

Project SAGE is built around four areas. The first is market development and use-case exploration, covering B2C agentic commerce and everyday payments, B2B procurement and treasury management, and machine-to-machine (M2M) usage-based payments for digital services. The second is Agent Identity: Know Your Agent (KYA) practices, meaning verification of who an AI agent is and what authority it carries, together with authentication mechanisms to confirm agents' identities and delegated authority. The third is interoperability standards and protocols for secure interactions across the ecosystem. The fourth covers legal and regulatory frameworks, including governance arrangements, liability and consumer protection.

Expert group and controlled testing

A cross-sector Agentic Finance Expert Group has been established, bringing together financial institutions, technology service providers, professional service firms and academia. Use-case testing in controlled settings, including sandbox environments, is flagged as part of the project's approach.

Supervisory track for banks

In parallel, the HKMA published a discussion paper on banks' use of agentic AI. It seeks industry feedback on supervisory considerations including governance models, responsibility arrangements and a possible central register of trusted agents. Feedback is open via the discussion paper and no closing deadline has been announced. Speaking at the launch, Eddie Yue said the purpose of the project is to promote the application of agentic finance while paying attention to governance and safeguarding the stability of the financial system.

Key figures

IMPACT

Who is affected: banks and other financial institutions operating in Hong Kong, along with fintech firms, technology service providers and their compliance teams. What changes: agent identity verification under KYA, governance and liability arrangements and consumer protection expectations are being formalised, and a central register of trusted agents is on the agenda. Timing: the expert group is in place, controlled use-case testing is part of the project and bank feedback via the discussion paper is open now, with no deadline set. Risk of inaction: firms that delay may face unresolved liability for agent-driven transactions, weaker controls over delegated authority and limited influence over the interoperability standards and trusted-agent frameworks now taking shape.

Sources: hkma.gov.hk, news.rthk.hk, caixinglobal.com, opengovasia.com, tokenpost.com

IMPACT checklist

Sources:

  • h
  • k
  • m
  • a
  • .
  • g
  • o
  • v
  • .
  • h
  • k

- ,

  • n
  • e
  • w
  • s
  • .
  • r
  • t
  • h
  • k
  • .
  • h
  • k

- ,

  • c
  • a
  • i
  • x
  • i
  • n
  • g
  • l
  • o
  • b
  • a
  • l
  • .
  • c
  • o
  • m

- ,

  • o
  • p
  • e
  • n
  • g
  • o
  • v
  • a
  • s
  • i
  • a
  • .
  • c
  • o
  • m

- ,

  • t
  • o
  • k
  • e
  • n
  • p
  • o
  • s
  • t
  • .
  • c
  • o
  • m

Continue the conversation: https://t.me/ComplianceExplained Β· Bluesky

Standards and Signals

πŸ“° Read the original article on Dev.to AI

Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes β€” full credit and traffic to the original publisher.