Financial crime doesn't operate in silos, yet most fraud detection tools do. Today's landscape is filled with point solutions: KYC platforms that stop at onboarding, AML engines focused on transaction monitoring, and network graph tools that excel technically. Fraud signals are everywhere, but data remains fragmented. AI Fraud Surveillance unifies financial, qualitative, and network signals into a single view, enabling fraud detection before it happens, not after.

Delinquency patterns shift faster than models can be retrained, and labeled examples of emerging risks are scarce. This session shows how a Tabular Foundation Model purpose-built for lending detected approximately 33% more early-delinquency cases than a quarterly-retrained ML model, without retraining. The speakers will explain how the model, relevant historical examples and an AI agent work together to support lending decisions, share lessons from development, and demonstrate the practical workflow.

Stablecoins can move money across borders in minutes, but businesses still need payments to land locally, reconcile cleanly and stay compliant. A payments provider and a stablecoin issuer share what they learned from connecting through a stablecoin payment network: how network access, local payout and operational controls fit together, and what it really takes to integrate, onboard clients and run it every day.