Ping Identity launches AI agent access control for enterprise Claude deployments

Ping Identity has launched Enterprise Personal Agent Access, an end-to-end security approach designed to give enterprises visibility and control over personal AI agents - including shadow AI - without slowing adoption.

Delivered through PingOne Privilege, the offering discovers personal AI agents running across an enterprise environment, associates each session with the user and device behind it, and enforces access and action policies at runtime.

Supported environments include desktop assistants such as Claude and coding agents such as Claude Code.

For developer workflows specifically, Ping Identity said agents can commit code and reach approved resources without holding long-lived credentials. Every action is attributed to either the developer or the agent itself - not just the user account - giving security teams an auditable record of what the agent did and when.

Policy enforcement covers whether to allow, deny or log an action, require human approval for sensitive activity, or revoke access in real time. The company said agents can reach MCP servers, code repositories, internal APIs, Kubernetes clusters, databases and cloud infrastructure, making enforcement at the point of action a priority.

"The security challenge is establishing a common way to see and govern all of them," said Andre Durand, founder and chief executive at Ping Identity.

"The question isn't whether an agent is intelligent enough to act, but whether the enterprise can see and control its action when it does."

Enterprise Personal Agent Access is available now and is already being piloted with global enterprises, Ping Identity said.