Banks grapple with AI agents reshaping customer loyalty

Banks are growing cautious about AI agents, not only due to security threats but also because they could disrupt the financial incentives behind customer loyalty. The concern extends beyond cybersecurity to the potential for agents to instantly redirect deposits to accounts offering better yields with minimal user input.

The central challenge remains identity verification. On October 15, Proof introduced its agent identity wallet alongside an open x401 protocol, which ties AI agents to confirmed human identities. The system operates by requiring digitally signed authorizations for transactions, allowing banks to verify both the identity of the requester and the approved payment details.

For instance, if an AI agent requests a $25,000 transfer, the bank triggers an x401 challenge, demanding proof of identity alongside explicit confirmation of the amount, recipient, and account. The customer verifies and approves these details on their own device before the bank cross-checks the signature against its policies. This approach eliminates blind trust by generating digital records for every transaction.

Even with verification in place, banks are not automatically permitting access. Different actions, such as reading balances, suggesting new accounts, or moving funds, carry varying risk levels. A bank may reject a verified request, indicating concerns beyond security alone.

AI agents have the potential to alter deposit behavior fundamentally. A sophisticated agent does not merely seek the highest interest rate but optimizes decisions based on a customer’s preferences, fees, and brand loyalty. Banks must decide whether to compete on price, service, or convenience or attempt to delay change by increasing switching costs. While friction may slow progress, it cannot prevent it entirely.

The critical issue is no longer preventing agents from moving deposits but how banks will retain customers when financial management becomes effortless. If an AI can handle comparisons and executions, the primary reason to remain with a bank may reduce to trust in the institution itself.

Proof’s protocol addresses part of the challenge by establishing agent identities. However, the larger question is what comes after verification. Banks recognize that agents will emerge, but readiness remains uncertain.