UBS’s $20 Million AML Fine Exposes a Blind Spot in Automated Monitoring

On August 3, 2026, the Financial Industry Regulatory Authority (FINRA) imposed a $20 million fine on UBS Financial Services Inc. for anti-money laundering (AML) failures that had occurred between January 2019 and June 2023. The key issue was not the absence of monitoring, but rather that an automated monitoring system failed to identify a significant portion of transactions it was supposed to review. Automation shifted the risk from human oversight to data completeness.

In 2018, UBS had already faced a $4.5 million fine from FINRA for inadequate monitoring of foreign-currency wires. From January 2019 to January 2021, the bank continued using a legacy process of quarterly manual review of a report containing thousands of wires, often lacking critical geographic data. The new automated monitoring tool was introduced in February 2021, aiming to address the limitations of manual review.

However, an incomplete data file and a change in labeling caused the tool to overlook approximately 33% of foreign-currency wires in retail accounts approved for spot transactions. From January 2019 to June 2023, UBS failed to reasonably monitor over 60,000 transactions, totaling $10 billion.

The case highlights that the failure was not solely due to the software, but rather a remediation program that did not ensure end-to-end assurance that the relevant transaction population was being monitored. Additionally, FINRA found weaknesses in UBS's customer due diligence program and delays in detecting and reporting suspicious money movements.

The distinction between detection accuracy and coverage completeness is crucial. While detection accuracy assesses how well a system evaluates the transactions it receives, coverage completeness determines whether the system receives every transaction it is required to evaluate. Success in one does not guarantee success in the other.

Automation has its advantages, but it requires separate control measures. Four important practices for an effective AML program are: population reconciliation, data-lineage validation, exception and feed monitoring, and change-triggered revalidation. Monitoring the monitoring system is essential, as failures can occur at various stages, including upstream feeds, data mislabeling, or misclassification.

A comprehensive approach that combines automation with independent monitoring and human oversight is necessary to ensure accurate and complete AML coverage.