Feedzai Unveils RiskFM, Aimed at a Frustrating Plague of Financial Fraud

Why It Matters

By widening data inputs, RiskFM could help banks cut fraud losses and chargebacks, addressing the industry’s top priority for 2025.

Key Takeaways

• RiskFM expands beyond card‑network data to holistic activity
• Analyzes $9 trillion in payments, 120 billion events yearly
• Early adopters include Lloyds Banking Group for real‑world testing
• Targets adaptive fraudsters amid rising U.S. bank losses
• Aims to become multi‑use‑case fraud prevention suite

Pulse Analysis

The surge in financial fraud has forced banks to rethink traditional defenses. In 2023, nearly three‑quarters of major U.S. banks reported higher fraud losses, prompting a shift from cost‑focused initiatives to aggressive fraud mitigation. This environment creates demand for solutions that can ingest diverse data streams and adapt in real time, a niche Feedzai targets with its new RiskFM platform.

RiskFM distinguishes itself by leveraging a massive, unified data set that spans onboarding signals, transaction histories, and digital behavior, covering about $9 trillion in payments and 120 billion events annually. The model treats fraud detection as an adversarial problem, using advanced machine‑learning techniques that continuously retrain against evolving attacker tactics. By moving away from the predictability of card‑network‑only models, the system promises higher detection rates while reducing false positives that hurt customer experience.

The market impact could be significant. Early adopters such as Lloyds Banking Group are testing the platform, signaling confidence from legacy institutions in newer AI‑driven tools. If RiskFM delivers on its promise, it may set a new benchmark for fraud‑prevention suites, pressuring competitors to broaden their data horizons. For banks, the technology offers a path to lower chargebacks, protect revenue, and meet regulatory expectations, all while maintaining a seamless customer journey in an increasingly digital payments landscape.