How AI Is Changing Anti-Money-Laundering Work

Banks and other financial institutions are using artificial intelligence to detect, prioritize and investigate potential money laundering. The focus is shifting from producing more alerts to helping compliance teams assess them faster and with greater consistency.

AI is being used in transaction monitoring, customer-risk scoring and suspicious activity report reviews. Machine-learning systems can examine patterns across accounts, payments, locations and counterparties, then flag activity that differs from a customer’s usual behavior.

The systems can also reduce the number of alerts investigators review manually. Banks can use earlier case decisions and customer information to rank alerts by risk, group related transactions and identify links between accounts that may not be visible in a single payment record.

Generative AI is being tested for tasks such as summarizing case files, locating information in internal records and preparing draft investigation notes. Human investigators remain responsible for checking the evidence, deciding whether activity is suspicious and submitting reports to regulators or financial intelligence units.

AI systems can produce inaccurate results when customer records are incomplete, transaction data is poorly structured or models are trained on limited examples. Banks must record how an alert was generated and why an investigator accepted or rejected it.

Financial institutions are combining AI tools with rules-based monitoring systems. Rules can identify known risk indicators, including transactions involving sanctioned entities or activity above a set threshold. AI can examine broader patterns and relationships across large volumes of data.

The technology is changing how anti-money-laundering teams allocate their time. Investigators can spend less time collecting and sorting information and more time assessing unusual activity. Compliance departments are reviewing model performance, testing for bias and setting procedures for human oversight.

Regulators in major financial markets permit the use of advanced analytics in financial crime controls while requiring firms to manage model risk and maintain accountability. Banks must demonstrate that their monitoring programs address relevant risks and that their reports are based on documented evidence.

Adoption varies by institution. Larger banks generally have more transaction data, technical staff and funding for model development. Smaller firms may rely on external technology providers. In both cases, system performance depends on data quality, control design and the judgment of trained compliance staff.