AI reshapes anti-money-laundering programs

NICE Actimize and an industry events organiser will host a webinar where experts will examine how artificial intelligence is delivering measurable value across anti-money laundering (AML) processes. The event will address use cases that have moved beyond pilot stages and are in operational or near-operational deployment.

Panelists will outline specific applications of AI across customer onboarding, know-your-customer (KYC) checks, transaction monitoring, case investigations and suspicious activity report (SAR) preparation. Topics scheduled for discussion include automated adverse media searches, refined customer risk scoring, optimisation of screening rules to cut false positives, alert triage to prioritise higher-risk cases, faster investigations through intelligent summarisation, and automated drafting support for regulatory filings.

The webinar will separate the roles of different technologies. Machine learning models will be presented as tools for scoring, anomaly detection and pattern recognition in transaction data. Large language models will be shown in tests for interpreting unstructured text, extracting facts from documents and compiling narrative elements of reports. Agentic AI, which can carry out a sequence of actions to complete tasks, will be discussed in the context of workflow orchestration and repeatable investigation steps, along with the extra governance such systems require.

Speakers will address operational hurdles that keep many projects at pilot stage. They will discuss technical integration between models and case-management systems, the need for consistent and cleaned data feeds, routine model validation and performance metrics, and clear escalation paths between analytics teams and investigators. The session will also cover organisational choices about building in-house models, licensing vendor solutions, or combining both approaches, considering each option’s impact on deployment time, cost and regulatory oversight.

Regulatory and governance expectations will be part of the agenda. Panelists will describe documentation requirements for model design and testing, expectations for explainability when automated outputs affect customer outcomes, the need to keep audit trails for data and decisions, and policies to limit reliance on opaque outputs without human review.

The webinar will present practical examples and operational steps for compliance officers and technology leaders seeking to move AI from experimentation into routine AML work. The discussion will focus on measurable outcomes, integration tasks, model controls and governance practices that align with supervisory expectations.