Across selected UK, Irish and US enforcement actions, monitoring failure repeatedly appears as an end-to-end control-chain problem — not simply a weak rule or a poorly tuned threshold.
The selected evidence shows multiple, distinct ways for monitoring to fail: relevant activity may never enter the monitoring perimeter; scenarios may not cover known risks; parameters may make detection ineffective; transaction meaning may be wrong; customer context may be insufficient; or alerts and suspicious activity may not be investigated and reported effectively.
This page synthesises primary-source findings across individual cases. It does not assert that every institution or every enforcement action exhibits these patterns, nor does it attribute the cross-case synthesis itself to any regulator.
Relevant customers or transaction classes are excluded, filtered or never delivered to monitoring.
The platform runs, but the scenario set does not cover known or changing money-laundering risks.
Rules exist but thresholds, parameters or configuration make them insufficiently risk-sensitive.
Transactions or customer context reach monitoring in a form that does not support the intended risk decision.
Detection output does not translate reliably into timely investigation, escalation and suspicious activity reporting.
Without reconciliation, coverage testing, scenario-risk review and post-change assurance, a long-running control gap can remain invisible.
Official finding. The FCA found that a data-feed error meant transactions around account opening were not monitored, and that Metro did not have an effective mechanism consistently checking that all relevant transactions were fed into the monitoring system until December 2020.
Official finding. The FCA identified weaknesses in scenario risk coverage, testing and updating of system parameters, and checks over the accuracy and completeness of data feeding and held within monitoring systems.
Official finding. The Central Bank of Ireland found that historic filters in Danske's automated system erroneously excluded certain customer categories from transaction monitoring.
Official finding. DOJ states that TD Bank excluded domestic ACH, most check activity and other transaction types from automated monitoring, and added no new monitoring scenarios from 2014 through 2022 despite known deficiencies, emerging risks and new products and services.
Official finding. The FCA states that some cash deposits were incorrectly recognised by NatWest's automated monitoring system as cheque deposits and therefore were not subjected to cash-specific monitoring rules; the FCA also described calibration and escalation failures.
Official finding. The FCA found ineffective systems for verifying customer information about intended business activity and failures to properly monitor expected account activity against what was actually occurring.
Official finding. The FCA's Final Notice states that insufficient customer data reduced Monzo's ability to assess whether transactions were consistent with expected activity, alongside weaknesses in transaction-monitoring processes, procedural guidance and staff capability.
Official finding. The Central Bank of Ireland states that Coinbase Europe failed to fully and properly monitor more than 30 million transactions and was required to conduct additional monitoring over a further transaction population.
FCRisk interpretation: a defensible TM framework needs evidence across population completeness, risk-to-scenario coverage, parameter effectiveness, data integrity, change control, alert handling and outcome testing.
Risk coverage → in-scope population → source-to-control reconciliation → scenario inventory → parameter testing → data correctness → change/migration testing → alert/investigation QA → reporting outcome → sustained control evidence.
The FCA's current wholesale-bank supervision material separately emphasises end-to-end testing, data lineage, completeness controls, reconciliation and alerts for stale or incomplete feeds.
Go deeper into the upstream population, mapping and reference-data dependencies that can cause a functioning control to operate on the wrong scope.
Examine why known control weaknesses can persist even after fixes, programmes and formal remediation activity begin.