Verification of Payee (VOP) is the overlay service for credit transfers in which the input beneficiary is checked against the actual account holder. It has been our bread and butter from day one. Every EU bank had to implement VOP, the regulation made the business case for us. When VOP became mandatory, adoption was guaranteed. Fraud prevention isn't. This session shows how we're building a Fraud Risk Intelligence platform on top of existing VOP infrastructure, reusing verified account data, adding third-party sources, and using AWS-based ML to turn a compliance layer into real risk intelligence banks choose to adopt.

Rather than build from zero, we started from what already exists and layered AI on top of it. VOP gave us verified account data, integration pipelines, and a trusted relationship with hundreds of PSPs. Our Fraud Risk Intelligence platform builds directly on that foundation: the same data-sharing rails, enriched with account activity signals, business registry data, and transaction history, all consolidated into a growing data lake that forms the foundation for model training and feeds a real-time A2A risk scoring engine built on AWS SageMaker.

Under the hood, this means heavy use of data science and machine learning: continuous drift detection, automated retraining, and a dynamic rules engine that adapts as fraud patterns shift. Feedback loops are central to this, every fraud label and outcome flows back into the data lake, continuously sharpening the models and keeping the platform effective against tactics that keep evolving. All of this had to be built for sub-second response times at scale, without ever disrupting the VOP service running underneath it.

In this talk, we'll share what changes when your business case rests on value instead of regulation, how we're using existing infrastructure and data to accelerate a fraud platform, and lessons learned so far.