Security
Feedzai Launches AI Agent Detection for Bank Account Security
Digital Trust expansion helps financial institutions distinguish legitimate customer AI agents from fraudulent automated access attempts.
Key takeaways
- Feedzai expanded Digital Trust to detect AI agents accessing bank accounts using device fingerprinting, behavioral analysis, and cryptographic device binding.
- The solution helps banks distinguish legitimate customer AI agents from fraudulent automated access attempts using compromised credentials.
- Banks can apply custom policies to allow, challenge, or block AI agent activities while recognizing trusted customer-agent relationships.
- Enhanced authentication capabilities reduce reliance on vulnerable SMS one-time passcodes and protect against MFA bypass attacks.
- Feedzai protects over one billion consumers globally and processes more than $19 trillion in transactions annually.
Fraud prevention specialist Feedzai has expanded its Digital Trust solution to address a growing challenge for banks: detecting and managing AI agents that access customer accounts on their behalf.
The New York-based company's enhanced platform combines device fingerprinting, behavioral analysis, network intelligence, and cryptographic device binding to distinguish authorized AI agents from potentially malicious automated activity. The expansion addresses scenarios where customers deploy third-party AI agents using valid credentials—activity that appears legitimate but may not be authorized or secure.
The authentication challenge
Banks face a complex risk landscape where valid credentials alone don't establish legitimacy. Criminals can exploit compromised devices, manipulate account recovery processes, or use social engineering to gain access to customer accounts. Even when customers intentionally use AI agents, banks need mechanisms to verify the relationship and enforce appropriate controls.
Feedzai's Digital Trust assesses risk both before login and throughout a banking session. The technology recognizes returning devices and applies fraud risk scores informed by intelligence shared across Feedzai's network, identifying suspicious patterns that individual banks might miss in isolation.
The platform's new cryptographic device binding capability creates a three-way link between customer, device, and bank. This helps institutions establish which customer an agent represents and determine whether its activity aligns with bank policies.
Policy-based control and authentication improvements
Once Digital Trust identifies an AI agent, banks can apply their own rules to allow, challenge, or block specific activities. The system also recognizes previously trusted customer-agent relationships, reducing repeated authentication requests while maintaining oversight of customer consent.
Beyond AI agent detection, the updated solution strengthens authentication for human customers. By confirming registered device presence and checking for compromise indicators, Digital Trust aims to reduce dependence on SMS one-time passcodes—a method vulnerable to SIM swapping, phishing, and social engineering attacks.
"A leading agent detection strategy starts with a foundation banks can trust: protected signals, cryptographic device recognition and network intelligence," said Pedro Barata, Feedzai's Chief Product Officer. "Personal and malicious agents can look very similar. Our latest advancements in Digital Trust help banks tell them apart."
Why it matters
As AI agents become more prevalent in consumer banking interactions, financial institutions need technical infrastructure to manage this new category of access without compromising security or creating friction for legitimate use cases. The challenge extends beyond traditional fraud detection: banks must balance customer convenience with regulatory compliance and risk management when automated systems act on behalf of account holders. Feedzai's approach addresses this gap by providing visibility and control mechanisms specifically designed for the agentic AI era.
Founded in 2011, Feedzai operates a risk operations platform covering identity verification, fraud prevention, and financial crime detection. The company reports protecting over one billion consumers across more than 190 countries and safeguarding over $19 trillion in transactions annually. Feedzai plans to expand its agentic detection capabilities in coming months.
These details were first reported by Finovate.