Where AI Is Delivering in the AML Lifecycle

A recent online webinar hosted by NICE Actimize and a financial publisher examined practical uses of artificial intelligence across the anti-money laundering lifecycle. The session focused on how firms can move beyond pilots to operational deployments and measure outcomes.

Nicola Eschenburg, EMEA AML Strategy Leader at NICE Actimize, took part on the panel alongside other industry experts, with a contributing editor moderating the discussion.

Panelists outlined specific use cases where AI is already applied. These included adverse media research to identify negative information, automated customer risk assessment to support onboarding and KYC decisions, screening optimisation to reduce false positives, alert triage to prioritise investigations, and automation of investigation steps and suspicious activity report (SAR) preparation.

Speakers described machine learning models as tools for detecting transaction patterns and unusual behaviour. Large language models were framed as assistants for summarising case facts and drafting narrative components of reports. Agentic systems were discussed as mechanisms to automate multi-step investigative workflows under supervised conditions.

The webinar addressed the industry shift from experimentation to production. Panelists identified obstacles that limit scaling, including siloed systems, inconsistent data across teams, and disconnected workflows that block the flow of risk signals. They said integration of systems and data consistency are prerequisites for end-to-end AI benefits.

Regulatory and governance issues were a focus. Panelists noted regulators are asking for transparency, explainability and controls around AI-driven decisions. Firms were advised to document model development and testing, keep audit trails for automated actions, and set up oversight and validation processes that align AI outputs with compliance requirements.

Discussion covered build, buy or blend choices for AI capabilities and the governance each path requires. The panel highlighted measurable performance indicators firms can track when moving to production, including reductions in false positives, shorter investigation times and lower resource costs. They recommended human-in-the-loop controls for high-risk decisions involving agentic AI.

The panel outlined where AI is delivering measurable improvements across customer onboarding, transaction monitoring, investigations and SAR preparation, and identified operational and governance steps firms should take to scale those improvements.