Feedzai Launches AI Agent Detection to Combat Automated Financial Crime and MFA Bypass

Feedzai has expanded its Digital Trust solution to enable banks to distinguish between human users and AI agents accessing accounts. By implementing cryptographic device binding and persistent fingerprinting, the platform aims to secure the growing "agentic" economy while reducing the friction and costs associated with traditional multi-factor authentication methods like SMS passcodes.

What was announced

Feedzai has introduced new capabilities to its Digital Trust platform specifically designed to identify and manage AI agents at scale. As automated agents increasingly perform tasks on behalf of consumers, financial institutions face the challenge of distinguishing between authorized personal agents and malicious automated bots. The solution utilizes persistent device fingerprinting and network signals to recognize returning activity, while cryptographic device binding creates a secure link between the customer, their device, and the bank. This "triangle of trust" allows banks to verify who an agent is acting for and whether the session is legitimate.

The update addresses the limitations of traditional login security, where valid credentials no longer guarantee a session's safety due to compromised devices or social engineering. Feedzai’s risk engine now draws on a broader network view and shared intelligence to surface threats that might be invisible to a single institution. Beyond agent detection, the release targets the inefficiencies of legacy multi-factor authentication (MFA). By using the device itself as a possession factor, the system has demonstrated a 41% reduction in the use of SMS one-time passcodes (OTPs). Performance data indicates that the shared signals identify 44% more fraud within the highest-risk 0.1% of sessions without increasing false positives. Additionally, the system removes approximately 6,000 unnecessary authentication challenges per million logins compared to competing technologies.

The companies involved

Feedzai is a specialist in AI-native financial crime prevention, providing platforms designed to manage risk across the customer lifecycle. The company’s technology is used by global financial institutions to combat money laundering, fraud, and other forms of financial crime by analyzing massive volumes of data in real-time. Feedzai has established itself as a significant player in the regtech and fraud prevention space, focusing on the integration of machine learning to stay ahead of evolving criminal tactics.

The company’s recent focus has shifted toward the "agentic" shift in technology, where AI agents—rather than humans—increasingly interact with banking interfaces. This requires a transition from simple behavioral biometrics to more complex identity verification that accounts for automated but authorized activity. Feedzai’s market position is built on its ability to provide a unified view of risk, often leveraging shared intelligence across its network of clients to improve detection rates for all participants. The firm continues to operate as a private entity, maintaining a global footprint with a significant presence in major financial hubs.

What FF News has reported before

FF News has extensively tracked Feedzai’s expansion into agentic AI and real-time payment security. In late 2026, the company released the Feedzai Index Reveals 13-Point Gap Between Bank Confidence and Actual AI Fraud Readiness, highlighting a disconnect in the industry’s preparedness for AI-driven threats. This followed the launch of its own agentic tool, Feedzai Launches Farol: How Agentic AI for Fraud is Automating Bank Risk Operations, which focused on streamlining internal bank operations.

The company has also been active in regional infrastructure projects, notably when Peoples Group and Feedzai Partner to Secure Canada’s Real-Time Rail Payments. Additionally, the firm has strengthened its European footprint through a collaboration with IT Smart Systems and Feedzai Partner to Secure Digital Banking with AI-Native Fraud Prevention.

What this means

This announcement signals a critical pivot in the fraud prevention landscape: the industry is moving beyond simple bot detection toward a nuanced "agent management" model. As consumers begin delegating financial tasks to AI agents, banks can no longer rely on traditional behavioral biometrics that flag all non-human activity as suspicious. The pressure is now on legacy authentication providers whose reliance on SMS OTPs is increasingly viewed as a liability rather than a security measure. This shift raises significant questions about liability and consent; if an authorized agent makes a mistake, the "triangle of trust" established here will be the primary evidence used to settle disputes.