Skip to main content

Press release · November 18, 2026 · globenewswire.com

HYPR launches AgentPass to link AI agents to human supervisors

Summary

HYPR, the Identity Assurance Company, has launched HYPR AgentPass to address an identity blind spot created by autonomous AI agents. The offering ties each agent to a named human supervisor, so agent activity can be traced to an accountable person.

Autonomous agents now handle daily tasks across corporate networks, most on borrowed credentials such as shared service accounts or personal developer API keys. Nothing links an agent's actions to the person it acts for, so security teams cannot verify which agent initiated an action or on whose behalf, or stop an unauthorized request. Closing this visibility gap is necessary before deployments scale.

AI gateways route agent traffic and prompt-inspection tools screen for dangerous inputs, yet AgentPass addresses a different question before each model or tool call: whether an agent acting for a particular person may perform a given action. The platform binds each agent to a named human supervisor who signs in through the organization's identity provider.

All agent connections run through one governed gateway, where agents start with zero access and see only the tools their supervisor's role permits. Short-lived, single-tool session tokens are issued from the gateway, so agents never hold tool credentials. The product also covers coding agents and their sub-agents on developer workstations, and it is available today.

Administrators write policies in plain English, test them on live traffic in Observe mode, then enforce them to allow, block, or hold actions for approval. High-risk actions go to a supervisor's separate device for FIDO2 passkey approval. Authorized users can disable a rogue agent, and its next request is refused everywhere, including in running sessions.

“Security teams cannot protect what they cannot govern,” said Bojan Simic, CEO and co-founder of HYPR. “Enterprises are moving to autonomous agent workflows, but the access models underneath were never built for it,” said David Reilly, Chief Development officer at WWT. Gartner predicts that through 2028, over 50% of enterprise AI initiatives will halt due to unresolved agentic identity challenges.

Our summary. The original keeps its own byline.

Read the original on globenewswire.com

Companies in this article

Topics