Investigators approve the agent's alert dispositions and case notes. Senior investigators approve restrictions on accounts. The decision to report suspicion to the authorities stays with the nominated officer or a named deputy, never with the agent. Approval rights depend on a current certificate, and no one closes an alert on a customer they manage.
What is the financial crime approver rule, and who owns it?
Financial crime teams work alerts from transaction monitoring and screening, investigate cases and decide whether to report suspicion. Agents help by gathering transactions, KYC data and prior alerts into one view, drafting the disposition narrative and preparing reports. The FATF Recommendations, the international standard for anti-money laundering, expect a risk-based approach, customer due diligence and suspicious transaction reporting; those duties stay with the firm and its people. The approver rule says who releases each held action.
The money laundering reporting officer or nominated officer owns the rule. Internal audit reviews it. The agent owner configures it. Our explainer on what an agent approver is covers the role, and our guide to AI agent skills by role covers what investigators need to know.
What are the steps?
- Nominated officer: lists the agent's actions by depth. Depth 1: reads alerts, transactions, KYC files, screening hits and prior cases. Depth 2: drafts disposition narratives, requests for information and draft suspicious activity reports. Depth 3: enriches cases, requests information from relationship teams, and routes alerts to queues. Depth 4: restricts an account, recommends exit, or submits a report.
- Nominated officer: assigns alert dispositions to investigators. Closing an alert as not suspicious is a decision with the investigator's name on it.
- Nominated officer: assigns restrictions to senior investigators, with a time limit and a review date.
- Nominated officer: keeps report submission with themselves and named deputies only.
- Nominated officer: ties each role to a current certificate and to the firm's own AML training.
- Agent owner: configures the holds and files the rule.
| Step | The agent does | Depth | Stays with a person |
|---|---|---|---|
| Gather case | Pulls transactions, KYC, prior alerts | 1 | Nothing |
| Request information | Asks the relationship team | 3 | Nothing within templates |
| Draft disposition | Writes the narrative | 2 | Investigator decides and approves |
| Restrict account | Prepares and holds | 4 | Senior investigator |
| Draft report | Drafts from the case | 2 | Nominated officer or deputy decides |
| Submit report | Prepares and holds | 4 | Nominated officer or deputy |
What must the agent never do in a financial crime case?
Two things. First, decide that something is or is not suspicious on its own. Suspicion is a judgement the law places on people, and the firm's reporting decision belongs to its nominated officer. The agent drafts; people decide.
Second, tip off. The agent must never tell a customer, or anyone who could tell the customer, that an investigation or report exists. That bound should be in its brief and tested in certification. Messages to customers from other agents, such as a support agent explaining why a payment is delayed, need the same bound. A customer asking "why is my account restricted?" gets an approved holding response, not an explanation.
The rule should also cover conflicts. An investigator who manages the customer relationship, or who referred the alert, does not dispose of it. The same goes for pressure from the business: a relationship manager asking for an alert on a valuable client to be closed quickly is a reason for more care, not less. The agent should record such requests on the case, so the investigator and the quality reviewer can see them, and the investigator decides on the evidence alone.
Which controls make the rule work, and what evidence do they leave?
Named investigator on every disposition. Evidence: approver names on each closed alert in the verdict record.
Reports decided by the nominated officer or deputy. Evidence: submitter names on each report.
Tipping-off bound tested. Evidence: certification results on the refusal and scope gates for disclosure attempts, and refused verdicts in live use.
Restrictions time-limited. Evidence: restriction review dates and outcomes.
Rights lapse with certificates. Evidence: certificate status on each approval date, checked against the approver authority matrix for the financial crime team.
What should you measure?
- Alerts reopened after closure, traced to the narrative or the decision.
- Quality assurance findings on sampled dispositions.
- Time from alert to disposition, by risk level.
- Customer messages refused for disclosure risk, as evidence the bound is working.
- Investigator coverage, using the competency matrix for the financial crime team.
Where does it usually go wrong?
Narratives become boilerplate. The agent's well-formed narratives start to look alike and investigators approve them without reading the transactions. Quality assurance should test the decision, not the prose.
Screening hits are closed in bulk. Approving a batch of name-match dismissals at once removes the individual judgement the rule exists for.
Restrictions are never reviewed. An account restricted pending investigation stays restricted for months.
Rules borrowed from credit or risk. Compare the approver rule for the credit team and the approver rule for the risk team: neither involves reporting duties or tipping-off.
Where the Academy fits
Pryme Intelligence Academy's Operator track covers human review, teaching the agent and when something goes wrong, with module tests set in the learner's own industry and role. Its practicum requires no missed red line, which suits investigators whose red lines include tipping-off. A firm can make disposition and restriction rights depend on a current Certified Pryme Operator certificate.
Questions people ask
Can the agent close obvious false positives?
Some firms let an agent recommend closure of clearly mismatched screening hits, with an investigator approving in small batches and quality assurance sampling. Full closure without a person is a policy decision for the nominated officer.
Who decides whether to exit a customer?
A senior person under your exit policy, often with the nominated officer. The agent prepares the case and the exit letter for approval.
Does the agent need access to all past cases?
It needs what your policy says investigators may see for that case. Scope its sources the same way you scope a new investigator's access.