As more businesses and employees let AI agents loose on their work, 1Password has released a new product to keep those agents reined in.
The Toronto-based cybersecurity software company launched its Unified Access Platform on Tuesday to help companies securely deploy AI agents, allowing autonomous work without sacrificing company control over authentication credentials.
While traditional cybersecurity tools only had to keep one human in check at a time, those humans are now deploying agents to work on their behalf. To do that, AI agents are using security credentials that could potentially expose a company’s production systems and sensitive data.
How agents gain access to systems and what they are doing with this access can be a black box for enterprise security teams.
The Unified Access platform promises to discover existing AI agents and credentials, secure them, then continuously authorize access and audit every action made by humans and AI agents. The platform tracks AI tools and agent activity across company systems, identifies and secures exposed credentials and secrets (like unencrypted security keys), and maps AI usage to specific users and devices.
Imagine that an AI agent is an unaccountable night-shift office worker who completes their work by using passwords left on sticky notes, and opening doors with copied keycards. Unified Access is a security officer who finds and locks that sensitive information in a briefcase, follows the agent around, and only opens that briefcase when necessary.
1Password said that Unified Access will soon have an audit capability, which will provide clear records of which credential was used, when, by which identity, and under whose authority. The platform is launching with collaborations across the AI and developer ecosystem, including giants like Anthropic and OpenAI, who are enabling the use of 1Password for credential security in their platforms.
RELATED: Q&A: Moltbook, OpenClaw, and the security risks of the new agentic-AI era
How agents gain access to systems and what they are doing with this access can be a black box for enterprise security teams, which has alarmed some cybersecurity experts. In an interview with BetaKit last month, Ian Paterson, CEO of Victoria-based cybersecurity company Plurilock, equated a hacked AI agent to hacking a person’s brain.
“What could they do with that? The answer is limitless,” Paterson said.
Founded in 2005 and formally known as AgileBits, 1Password is one of Canada’s most valuable tech companies. It sells identity security and access management software that helps individuals and corporate clients secure sign-ins to apps and websites. In recent years, 1Password has evolved from a consumer-facing password manager to a broader digital identity security platform for businesses, including more solutions for agentic AI security.
This past January, 1Password promoted Nancy Wang to CTO to drive the cybersecurity software company’s engineering efforts and AI plans.
Feature image courtesy 1Password.