As financial crime teams accelerate AI adoption, many organisations are grappling with a critical question: where do traditional rules-based approaches still outperform large language models — and where do LLMs unlock entirely new capabilities?

This session explores how Santander is combining explainable, scalable rules-based AI with targeted LLM use cases to improve efficiency, reduce false positives and accelerate investigations across financial crime operations.

• Reducing false positives through automated triage in customer screening using rules-based AI approaches
• Using LLMs to automate materiality assessments in adverse media investigations and accelerate case handling workflows
• Understanding how the architectural, governance and operational requirements differ between rules-based AI and LLM deployments in financial crime environments