AI’s Measurable Impact on AML Operations

A webinar hosted with NICE Actimize examined where artificial intelligence is delivering measurable value across the anti-money laundering lifecycle. The session featured Nicola Eschenburg, EMEA AML Strategy Leader at NICE Actimize, and Ashley Bostel, Head of Economic Crime Analytics Strategy at NatWest Group, and was moderated by Teresa Connors. Panelists focused on practical applications and the transition from pilots to operational deployment.

Panelists identified specific use cases with measurable results. In customer onboarding and KYC, AI is being applied to adverse-media screening and automated customer-risk scoring, which speeds reviews and consolidates information from public sources into a single view. In transaction monitoring, machine-learning models are used to improve screening rules and lower false positives, reducing the volume of alerts that require manual review.

For investigations and case management, firms are applying automation and natural language processing to triage alerts, collate relevant documents and draft sections of suspicious activity reports (SARs). Examples mentioned included model- and rule-driven alert prioritisation, automated evidence gathering that shortens investigation time, and template-assisted SAR drafting that streamlines reporting to authorities.

Speakers described how linking outputs from screening, transaction data and third-party sources creates a fuller view of customer behaviour. When those signals are connected across teams and workflows, investigators and compliance officers receive more context with each alert, enabling faster decisions and more targeted follow-up.

Panelists addressed agentic AI — systems that carry out sequences of tasks across multiple systems — and described its role as a helper for routine investigative tasks rather than a replacement for human judgment. Agentic approaches were described as able to orchestrate queries across systems, assemble case files and suggest next steps, while final decisions remain with experienced compliance staff.

Governance and regulatory expectations were discussed as part of operational deployment. Panelists said AI projects should include explainability features, audit trails and model-risk controls, and should be aligned with existing compliance policies. The panel suggested that the choice to build, buy or blend AI capabilities should depend on in-house expertise, time to market and the level of control firms require over models; several participants recommended a hybrid approach in which vendors provide core tools that firms customise and govern internally.

Attendees identified technical and organisational barriers to scaling AI, including integrating data across legacy systems, maintaining model performance over time, documenting decision logic for regulators and mapping AI outputs into existing workflows. The panel urged prioritising use cases that can be measured with clear metrics and establishing cross-functional collaboration between compliance, analytics and business teams.

The webinar concluded with a focus on practical outcomes: selecting AI applications that deliver quantifiable gains in efficiency or detection, embedding controls for transparency and connecting risk signals across teams so automated tools support faster, more informed compliance decisions.