SAMPLE-02 · AI-assisted AML transaction monitoring
Generated 2026-10-31 14:52 UTC · Owner: Marcus O'Brien · Sponsor: Head of Financial Crime
A Tier-2 Australian bank proposes to replace its rules-based AML transaction-monitoring engine with an LLM-based agent that reads payment memos, customer notes and adverse-media results to score suspicious-matter reports. The agent will draft the SMR narrative and auto-lodge to AUSTRAC once a compliance analyst clicks approve. Model hosted on a US cloud region.
Initiative SAMPLE-02 (AI-assisted AML monitoring) is assessed at ARA overall-residual Medium, CALIBRE band High (GII 63.75/100). The analyst-in-the-loop design materially reduces risk versus fully autonomous SMR lodgement.
The main remaining risks are (1) hallucinated content in SMR narratives, which the grounding-verification control directly mitigates, and (2) cross-border data transfer, which requires an APP 8 compliant contract before go-live.
Approve with conditions. Do not go live until the four listed conditions are satisfied and evidence uploaded against each.
None yet attached.
None outstanding.
| ID | Action | Owner role | Priority | Status |
|---|---|---|---|---|
| SAMPLE-02-A-01 | Implement Analyst approval gate before AUSTRAC lodgement | Head of Financial Crime | High | Open |
| SAMPLE-02-A-02 | Implement Grounding verification | AI Product Owner | High | Open |
| SAMPLE-02-A-03 | Execute Cross-border data-transfer contract with vendor | Legal / Procurement | High | Open |
| SAMPLE-02-A-04 | Establish False-positive monitoring dashboard | Compliance Analytics | Medium | Open |
| SAMPLE-02-A-05 | Update customer collection notice for cross-border transfer | Privacy Office | Medium | Open |
Audit trail: full machine-readable finding history, model IDs, timestamps, and evidence citations are stored in the Governance Fabric under initiative SAMPLE-02. Regenerate this pack at any time from the initiative page.