What happened
At the end of 2024, Bank Negara Malaysia (BNM) supervised 258 money services businesses (MSBs), most of them money changers. MSBs send BNM their transaction data every month. Each quarter, BNM's MSB Customer Anomaly Detection tool looks for customers with outlier transaction behavior. It combines machine learning models with rule-based parameters. FATF and Asia/Pacific Group on Money Laundering (APG) assessors visited Malaysia from 3 to 21 February 2025. Their mutual evaluation report of December 2025 describes BNM's supervisory technology in a box. There the report states that the tool flagged two MSBs in 2022 for facilitating abnormal customer transactions. On-site examinations then uncovered weaknesses in their ongoing due diligence. It also states there that BNM's technology solutions flagged 292 MSB licensees from 2020 to 2024. In the main text, the assessors judged that BNM's use of supervisory technology to automate risk ratings had been particularly effective for the MSB sector. The report does not say how the assessors checked the results in the box or what followed the examinations.
Sources
-
Descriptive report, by authors with no tie to the system.
FATF and APG. Mutual Evaluation Report of Malaysia: Anti-money laundering and countering the financing of terrorism and proliferation financing measures. FATF, Paris, and APG, Sydney, December 2025. Chapter 3 (Immediate Outcome 3), para 234 and Box 3.3 'Technology solutions for supervision of MSB and DPI issuers'.
In 2022, the “MSB Customer Anomaly Detection” tool flagged two MSBs for facilitating abnormal customer transactions; on-site examinations uncovered weaknesses in ongoing due diligence processes.
Details
AI system
- Workflow
-
MSBs submitted their transaction data to BNM every month. Each quarter, machine learning models and rule-based parameters flagged MSB customers with outlier transaction behavior. Supervisors followed the flags with on-site examinations of the MSBs concerned.
Follow-up
- Conditions for replication
-
BNM must keep access to detailed transaction data from MSBs, which the assessors named as the reason the approach suited this sector. BNM must have supervisors to examine the flagged MSBs, and the report counts 52 supervisors who examine its higher-risk firms on a three-to-five-year cycle.
Entry details
- Sector
-
Financial supervision: anti-money laundering.
- Purpose
-
Bank Negara Malaysia built the MSB Customer Anomaly Detection tool to find MSB customers with outlier transaction behavior. Supervisors use its results to focus on the MSBs concerned and to refer suspicions to Malaysia's financial intelligence unit.
31 more questions not yet checked
model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). MSB Customer Anomaly Detection points supervisors to two weak firms (AIBD-1041). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1041.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)