Password manager 1Password has launched a new tool that lets AI agents log into personal accounts for users. The agents never get access to the user’s password.
The product is called 1Password for Agents. The company announced it on Thursday, October 22, 2026.
Concerns have grown about how much access people are giving to AI agents. Security companies are now working to keep up with that change.
How 1Password for Agents Works
The system creates a line of communication between three parties. These are the agent, the user, and the password manager.
When an agent tries to sign in for a user, a unique digital key is sent to 1Password. This key confirms the agent’s digital identity.
The user then gets an alert about the login attempt. The user can personally confirm the request before the login goes through.
Once the request is approved, 1Password signs the agent in by filling in the credentials itself. The agent can then log in and complete its task without ever holding the password.
Users can choose from several AI agents. These include Claude Duet from Anthropic, OpenAI’s Dots and ChatGPT Work, Grok Bot, and Meta’s Muse.
Agents from Town, an AI assistant startup, are also supported. 1Password said it is working with other companies to expand the service.
In a blog post, the company explained why it built the tool. It said handing a password or credential to an agent is a common shortcut, but it can leave sensitive information “in places they can no longer control as a copy gets saved for the next time.”
The company said the product gives people “a trusted way to delegate tasks to personal AI agents while keeping control of the passwords, secrets, and credentials those tasks require.”
Banks Report AI Agent Activity in Customer Accounts
The launch came on the same day Axios reported that U.S. banks have begun to detect AI agent activity inside customer accounts.
The report did not describe rogue agents stealing from accounts. Instead, it showed that customers are allowing automated services to act for them, including on financial matters.
The main problem is that most banks cannot tell whether agent activity is tied to a real customer request. That makes it hard for banks to know which actions were approved.
Bankers interviewed by Axios described the situation as “horrifying.”
A Chase spokesperson also spoke with the outlet. The spokesperson said customers “should not give their usernames and passwords to anyone.”
