Technology
Transaction monitoring and anomaly-detection systems
Transaction monitoring and anomaly-detection systems are the bank-side software that continuously analyses customer transactions for patterns indicating money laundering, terrorist financing, or sanctions evasion, generating the alerts from which suspicious activity reports are filed. Built as anti-money-laundering compliance infrastructure, the technology has been progressively repurposed as a collection layer for state targeting: the private banking system running pattern detection at scale on behalf of financial intelligence units.
Function
Monitoring engines apply detection logic to account activity, often after execution, distinct from the pre-execution blocking of watchlist screening. Rules encode typologies such as structuring, rapid movement, funnel patterns, round-tripping and activity inconsistent with a customer's profile. Models may score anomalies and prioritise queues. An alert is only a prompt for human triage. Investigation, escalation and the decision to file a suspicious-activity report are separate stages, and a filing does not prove criminal conduct. The United States framework rests on the Bank Secrecy Act as amended by the Anti-Money Laundering Act of 2020; FATF, EBA and FFIEC materials guide risk-based design but do not demonstrate operational effectiveness. FATF revised parts of its standards in 2025, so current recommendations, not a superseded sector guide, are the controlling reference for the general framework.
Strategic significance
The monitoring layer is regulated private infrastructure. Article 1 of the definitional spine lists transactional and behavioural pattern recognition, funnel transactions, repetitive payment patterns, layering and crypto off-ramps among the FININT workflow's analytic techniques. Regulation requires covered institutions to operate controls and deliver selected reporting through the SAR or STR pipeline. Typology updates flow the other way, with advisories and FATF publications steering detection towards identified risks. This public-private arrangement extends state analytic reach, but an alert remains an institutional risk signal. It is not proof of a sanctions breach, criminal conduct or state direction.
Effectiveness and contestation
The technology's effectiveness is contested on the evidence. A 2024 US Government Accountability Office review found that interviewed banks and technology providers reported potential benefits from artificial intelligence and machine learning, including improved alert quality, but also data, explainability, privacy and supervisory challenges. Ronald Pol's system-wide critique argues that anti-money-laundering controls have imposed large costs while disrupting only a small share of criminal finance. Neither source supplies a universal alert-level false-positive rate. Defenders answer that monitoring also creates leads for FININT, raises evasion costs and can reveal relationships that static list screening misses. Machine-learning detection sharpens the trade-off rather than resolving it: a model tuned to reduce routine alerts may improve triage, but novel patterns still require investigation and governed human judgement.
Evaluation must specify the unit of analysis. Alert precision, investigator workload, report quality, prosecution, asset recovery and disruption of a network are different outcomes. A lower alert count can reflect better prioritisation or a blind spot, while a higher filing count can reflect improved detection or defensive reporting. The GAO found that fragmented outcome data constrain government-wide effectiveness assessment, which is itself a material limit on strong causal claims.
See also
Suspicious Activity Report (SAR) and STR systems · Sanctions list and watchlist screening technology · Bank Secrecy Act and United States financial reporting architecture (1970-present) · Financial intelligence (FININT) · Economic statecraft
Sources
Recommended citation
Cite this entry
Tennant, James J., ed. 'Transaction monitoring and anomaly-detection systems.' The Encyclopedia of Economic Statecraft, version 2.0, last reviewed 30 July 2026. https://jamesjtennant.com/entries/transaction-monitoring-and-anomaly-detection-systems/.
Suggest an edit