Ping Identity adds controls for AI agents in businesses

Ping Identity has launched new identity management features for organisations using AI agents, extending its platform to cover governance and access control for machine-driven actors.

The update focuses on three areas: programmable identity tools for developers and AI systems; discovery and governance controls for AI agents across their lifecycle; and privileged access management for desktop-based agents such as coding tools and AI assistants.

Businesses are beginning to use AI agents not only as software entities that need system access, but also as operators that help configure and manage identity environments. That shift raises new questions for security teams around visibility, accountability and credential handling.

Rather than creating a separate identity framework for AI deployments, Ping is extending existing controls so organisations can govern human users, non-human identities and AI agents within a single architecture.

Programmable identity

A central part of the announcement is a move towards machine-oriented administration of identity systems. Ping is adding what it describes as AI-first headless interfaces, allowing identity functions to be accessed through command-line tools, application programming interfaces, MCP and workflows designed for software agents.

It is also introducing agent-ready skills to help AI agents carry out tasks such as configuring access, troubleshooting identity flows and applying governance controls within approved policy boundaries. These tools are aimed at teams increasingly relying on automation and code-based workflows rather than graphical administration consoles alone.

The shift reflects a broader trend in enterprise software, with IT operations and security management moving towards programmatic control. In identity management, that means systems must serve both human administrators and automated tools acting on behalf of users or development teams.

Agent governance

Another part of the rollout focuses on governing AI agents themselves. As more organisations deploy agents across internal systems, they need to know which agents exist, what resources they can access, who owns them and how their activity can be reviewed.

The new controls cover discovery, ownership assignment, policy enforcement, access reviews, auditability and decommissioning. Under this model, an AI agent is treated as a distinct identity linked to a human owner and subject to oversight during both development and runtime use.

The governance issue is becoming more pressing as AI tools move from experimentation into operational settings. Security specialists have warned that unmanaged agents could create blind spots if they are given broad permissions without clear accountability or consistent controls.

Andre Durand, Chief Executive Officer and Founder of Ping Identity, said the rise of AI agents is changing the role of identity systems inside large organisations.

"AI agents are fundamentally changing how enterprise systems operate," Durand said.

"As enterprises make applications consumable by AI agents, Ping is making identity programmable, agents visible and governable, and resource access trustworthy. Identity is evolving from authentication infrastructure into operational governance infrastructure for the agentic enterprise."

Desktop access

The third element addresses desktop agents, including coding agents and AI assistants that interact directly with enterprise applications, repositories and tools. These systems may need access to complete tasks, but exposing credentials or long-lived secrets can increase security risk.

Ping says it brokers access to enterprise resources without handing those secrets directly to the agents. For coding agents, it is also adding attribution for code commits so organisations can distinguish actions taken by an agent from those taken by a human user.

The issue is significant for companies adopting AI-assisted software development, where agents can write code, make changes to repositories and interact with internal systems. Tracking what an agent did, and under whose authority it acted, is becoming an important compliance and security requirement.

Peter Barker, Chief Product Officer at Ping Identity, said the focus is on helping organisations expand AI use without weakening oversight.

"AI agents are changing both how work gets done and how identity must operate," Barker said.

"Enterprises need AI agents to operate across systems and resources without creating new trust gaps. Ping helps organisations adopt AI faster while preserving governance, accountability and control."

The announcement reflects how identity vendors are adapting their products as AI agents take on a larger role in business systems. Rather than treating agent activity as a niche edge case, suppliers are starting to build it into core identity, access management and audit functions.

For enterprises, the challenge is no longer limited to authenticating staff and customers. It now includes determining how autonomous and semi-autonomous software should be recognised, governed and granted access without losing human accountability.