{
  "question": "The Hong Kong Monetary Authority published a discussion paper on agentic AI in banking in November 2026. What records does it expect a bank to keep for each AI agent and for the actions the agent takes, and which agents would have to be reported to a proposed HKMA registry?",
  "slug": "what-records-does-the-hkma-expect-banks-to-keep-for-each-ai-agent",
  "url": "https://www.rightshold.com/answers/what-records-does-the-hkma-expect-banks-to-keep-for-each-ai-agent",
  "short_answer": "Two kinds: a record for every agent deployment, kept before it goes into production and updated through its whole life, and action logs for what each deployment does. The HKMA's Discussion Paper on Agentic A.I. Governance in the Banking Sector (published November 2, 2026 with the launch of Project SAGE) says the deployment record should name the agent's purpose and status; its accountable, business, technical and control owners and authorised users; its capabilities and access, including data leaving the bank and who receives it; its infrastructure and shared dependencies; its stop mechanism, fallback and log locations; its change and suspension history and what happens to memory, state and credentials at retirement; and its materiality assessment and next review date. Action logs should show what the agent did, what was affected, when, the instruction or event that prompted it, the data and tools used, the permissions and controls relied on, and any human approval or intervention, and be available to the HKMA on request. Only material agents would go to the proposed central registry. The paper is a consultation, not a binding rule; comments are due by February 1, 2027.",
  "checklist": [
    "List every agentic deployment, including pilots, and open a record for each before it reaches production. The paper treats unapproved tools and approved tools used outside their limits as shadow AI that the record should help you find.",
    "For each deployment, write down the approved business intent: what it may do, what it must escalate, its limits and approval conditions, who it serves, and when its authority should be reviewed or withdrawn.",
    "Name the accountable, business, technical and control owners and the authorised users. The paper says using outside models, cloud services or tools doesn't move accountability away from the bank.",
    "Give each deployment its own identity and credentials where you can. The paper warns that shared or borrowed accounts break the link between an action and the authority behind it.",
    "Log each action with: what it did, what it affected, when, the instruction or event that triggered it, the data and tools used, the permissions and controls relied on, and any human approval or intervention.",
    "Record suspensions, restorations and material changes to models, prompts, tools, data sources, workflows and permissions as they happen, not afterwards.",
    "At retirement, close the record and say what happened to the agent's memory, accumulated state, credentials and retained data.",
    "Run the paper's two-step materiality test (reach, then oversight) for each deployment before production and after every material change, and keep the evidence behind the conclusion."
  ],
  "sections": [
    {
      "heading": "What the HKMA paper asks banks to record",
      "points": [
        "The paper's practical test is whether a bank can explain any material action a deployment has taken, identify the authority behind it, name the components and dependencies involved, and produce enough evidence for independent assurance, incident investigation and remediation. It notes that when software acts, the log takes the place of a person's memory and signature, so attribution has to be designed in from the start because it may not be possible to rebuild later.",
        "The deployment record follows the agent from conception through design, testing, production, changes, suspension and retirement. The paper also expects component lineage: the models, data sources, tools, systems, interfaces, providers and credentials each deployment uses, so the cause and reach of an incident can be traced.",
        "Logs should be retained, accessible and dependable enough to support internal audit, investigation and reconstruction of material events. In discussing enforceable controls, the paper lists tamper-evident logs, approval records, access reviews and test results among the evidence a bank can use to show an agent's tools and permissions stayed within approved limits."
      ]
    },
    {
      "heading": "How the HKMA tests whether an agent is material",
      "points": [
        "Step one asks about reach. A deployment has reach if it can change a customer's or the bank's position (moving funds, changing accounts or data, making commitments, or deciding a customer's entitlement such as credit), discharge a duty owed to an authority or market (for example closing a regulatory alert), or support a critical operation. An agent that only drafts, researches or summarises has none, and the assessment stops there. Reach must also be significant, measured against thresholds the bank already uses, and a fast stream of small reversible actions can count if it outpaces the bank's ability to stop it.",
        "Step two asks whether genuine human approval of each action stands between the agent and the harm. If not, the deployment is material. If so, it is not material unless the action is irreversible, so a payment-release agent with a person approving every payment is still material because settled payments may not be reversible. A read-only agent that is the only automated fraud monitor is also material, because a critical operation depends on its accuracy.",
        "Deployments that share a model, platform, orchestrator or other critical dependency are also assessed together, since one failure could hit them all. The independent control owner endorses each assessment, and it is redone after changes, incidents or evidence that an approval gate isn't working."
      ]
    },
    {
      "heading": "When the HKMA counts human approval as genuine",
      "points": [
        "Approval counts only if the approver sees the substance of the action, has the authority and practical ability to reject it, rejection is accepted as a legitimate outcome, and the action stays pending until a decision is made. If any of these is missing, the paper treats the deployment as ungated.",
        "A bank claiming a gate needs continuing evidence that it still works. In practice that means keeping each approval with what the approver was shown, who decided, the decision and the time, and tracking how often approvers reject, so a gate that has become a rubber stamp shows up in the data.",
        "Material deployments would be notified to the HKMA, either before production or soon after; the paper asks which. A notification would draw on the internal record and evidence: owners and who may instruct the agent, how each action is tied to the agent and the human authority behind it, how the audit trail is retained, and the stop mechanism with evidence it has been tested."
      ]
    }
  ],
  "checks": [],
  "ask_rightshold": {
    "start": "https://www.rightshold.com/start?ref=seo",
    "mcp_endpoint": "https://www.rightshold.com/mcp",
    "ask": "Our payment-release agent staged a supplier payment and a treasury approver released it. Record the action against the agent's deployment record with the instruction that triggered it, the tools and permission used, and the approval, with no account numbers or customer details, so we can reconstruct it later.",
    "mcp_call": {
      "tool": "record_action",
      "arguments": {
        "action_type": "agent.payment_release",
        "outcome": "success",
        "occurred_at": "2026-11-13T09:40:00Z",
        "client_event_id": "dep-treasury-007:pay-2026-11-13-0042",
        "summary": "Deployment dep-treasury-007 (record v12) staged supplier payment batch PB-0042 from instruction INV-RUN-1113; tools: erp.read_invoices, bank.stage_payment; permission role treasury-agent-stage-only; released by approver role treasury-approver-2 after viewing amount and payee; approval ref APR-88213"
      }
    },
    "also": [
      "Needs your API key (or accept_terms:true once to open a free account; the key comes back once). It returns a ledger-committed receipt showing what was recorded and when. It proves the entry was submitted, not that the payment or approval happened as described.",
      "Keep account numbers, customer names and staff names out of summary. Hash the full internal log line or approval record and send it as action_sha256 if you need to tie the receipt to it.",
      "Use one client_event_id per action so a retry doesn't create a second entry.",
      "It doesn't keep your deployment record, run the materiality test or notify the HKMA."
    ],
    "scope": "Rightshold isn't affiliated with the HKMA and isn't a bank, regulator or compliance adviser. A recorded entry isn't evidence that a deployment meets any supervisory expectation. Details are from the HKMA's Discussion Paper on Agentic A.I. Governance in the Banking Sector (November 2026) and its November 2, 2026 press release, read November 13, 2026. The paper doesn't itself create binding requirements and the HKMA may change its approach after consultation, so check the HKMA's site for later guidance. Not legal advice."
  },
  "sources": [
    {
      "title": "HKMA: Discussion Paper on Agentic A.I. Governance in the Banking Sector (November 2026)",
      "url": "https://www.hkma.gov.hk/media/eng/doc/key-information/press-release/2026/20261102e3a1.pdf"
    },
    {
      "title": "HKMA press release: HKMA launches Project SAGE and establishes Expert Group for agentic finance (November 2, 2026)",
      "url": "https://www.hkma.gov.hk/eng/news-and-media/press-releases/2026/11/20261102-4/"
    },
    {
      "title": "Verified AI Runs on Rightshold",
      "url": "https://www.rightshold.com/verified-ai-runs"
    }
  ],
  "related": [
    {
      "title": "How to log every AI agent tool call for audit",
      "url": "https://www.rightshold.com/answers/how-to-log-every-ai-agent-tool-call-for-audit"
    },
    {
      "title": "How to audit which human authorized an AI agent action",
      "url": "https://www.rightshold.com/answers/how-to-audit-which-human-authorized-an-ai-agent-action"
    },
    {
      "title": "How to show a customer or regulator that a human reviewed an AI decision before it went out",
      "url": "https://www.rightshold.com/answers/how-to-show-a-customer-or-regulator-that-a-human-reviewed-an-ai-decision-before-it-went-out"
    },
    {
      "title": "How to build a kill switch for an autonomous AI agent",
      "url": "https://www.rightshold.com/answers/how-to-build-a-kill-switch-for-an-autonomous-ai-agent"
    },
    {
      "title": "How long do you have to keep AI agent logs under the EU AI Act?",
      "url": "https://www.rightshold.com/answers/how-long-do-you-have-to-keep-ai-agent-logs-under-the-eu-ai-act"
    }
  ],
  "computed_from": "Curated guidance; license verdicts computed from Rightshold's license profiles at render time. Not legal advice.",
  "updated": "2026-11-13"
}