jsonscraper

Wells Fargo sets rules for AI agents with access to bank accounts

Starting November 24, the bank plans to treat actions by an agent connected through a customer’s representative as the customer’s actions, and make the customer responsible for supervising them. The agreement clarifies the customer’s responsibilities while preserving the bank’s control over access.

Editorial Policy Report an error

When an AI agent can interact with a bank account, it matters not only what access it has but also who is responsible for its actions. Wells Fargo sets out rules for this scenario in a new version of its online access agreement, which takes effect on November 24, 2026.

Article: Wells Fargo sets rules for AI agents with access to bank accounts
Summary: An update to Wells Fargo’s agreement, effective November 24, 2026, outlines the responsibilities of customers who use AI agents through authorized representatives or delegates. The bank treats unwanted transactions made by an agent as customer-authorized, puts the customer in charge of supervision, and reserves the right to end access.
Image role: cover — the central idea
Section: 
Visual subject: Close-up
© jsonscraper · AI-generated illustration

In the updated Wells Fargo agreement, agents are addressed in the context of actions by a customer’s authorized representatives and delegates. The document extends responsibility to transactions carried out by their AI agents: the bank may treat as customer-authorized even transactions the customer did not want or intend to make. Customers also agree to supervise the actions of their representatives and their agents.

The agreement also covers data: its terms make customers responsible for protecting authorized persons’ login credentials. The practical effect depends on how the customer sets up access and what permissions the representative or delegate receives.

How this relates to bank practices

The update comes amid reports of third-party AI agents accessing accounts at U.S. banks. An October 22 Axios report, citing people familiar with the matter, says banks had detected access by such agents, and that some actions appeared independent of users’ instructions. The publication also reported that some institutions had restricted access. These reports describe an industry-wide situation; Wells Fargo’s terms are set out by the bank in its agreement.

Wells Fargo reserves the right to end a delegate’s access to a service, account, or financial product if the bank sees a security risk or a breach of the agreement. A customer’s permission therefore does not prevent the bank from managing access.

What to check before connecting an agent

Before granting access, find out what data the agent can read, what actions it can perform, and how to revoke permission. If the task is just to review spending, choose read-only access when available; for transfers and payments, limit permissions to what is necessary. After connecting, check the transaction history and confirm that access has been revoked when requested.

For developers of financial agents, these terms underscore the value of separating read access from the ability to carry out transactions. For customers, the takeaway is practical: read the terms from both the bank and the agent provider, and find out what permissions you are actually granting through an authorized representative or delegate.

The updated Wells Fargo agreement takes effect on November 24, 2026; at the time of publication, it is not yet in force. Customers can review the terms in advance and decide whether the proposed way for an agent to access an account suits them.

People

No people listed for this article yet.

Keep readingUS Banks Restrict AI Agents’ Access to Customer Accounts
Read the next article

Turn what you read into a working integration

Explore jsonscraper's social-data APIs, test requests and build your next workflow.

Explore APIs