Feedzai has broadened its Digital Trust offering so that banks can identify when AI agents are entering customer accounts, a move that tackles a growing blind spot as automated software begins to transact on behalf of consumers.
The upgrade rests on three pillars: more durable device persistence, a fraud score informed by Feedzai’s network-wide visibility and pooled intelligence, and cryptographic device binding that links the customer, the device and the bank in a single chain of trust.
The company says these elements also stand alone as useful tools, allowing lenders to drop needless step-up checks, lower authentication spending and harden defences against attempts to circumvent multi-factor authentication.
The core issue is that a correct password no longer proves a session is safe. AI agents can authenticate using legitimate credentials, while fraudsters lean on hijacked devices, account recovery loopholes and manipulation tactics to exploit genuine users.
Digital Trust draws together endpoint, device, behavioural and network data to judge risk both ahead of sign-in and across the full customer session.
To spot an agent, the platform pairs the information an agent discloses about itself with fingerprinting and behavioural indicators that expose automation. Persistent fingerprinting tracks repeat activity over time, and network-level scoring highlights warning signs that one institution alone might overlook. Device binding then connects each session to a known person and handset, revealing whom an agent represents.
Once an agent has been verified, banks can recognise familiar user-agent pairings and spare them repeat challenges where policy permits, while retaining authority over which agents are allowed and how consent is managed.
Source link







