Ping Identity launches platform to secure & manage AI agents

Ping Identity has unveiled a new solution, 'Identity for AI'. It is aimed at helping organisations introduce identity-first accountability in the growing domain of AI agents. The solution seeks to provide tools for managing and securing AI agent interactions, addressing visibility, governance, oversight, and threat protection as companies explore new forms of agentic automation.

Agent management

Identity for AI introduces a centralised platform that enables organisations to register, manage, and monitor AI agents across the enterprise. The platform offers control over both human and non-human identities, helping organisations oversee how agents interact within digital environments and with enterprise resources.

Features include agent authentication, authorisation, and the allocation of access privileges. These controls are designed to facilitate minimum necessary access while supporting clear oversight pathways. According to the company, this framework is intended to provide enterprises with confidence as they begin to rely more on agent-driven workflows.

"AI agents are changing how business gets done. With Identity for AI, we give organisations the guardrails to innovate responsibly and with confidence through enterprise-grade identity management. Identity is becoming the universal language of accountability-for humans and agents alike. It's how every AI decision can be verified and trusted," said Andre Durand, CEO and founder, Ping Identity.

Visibility and oversight

Identity for AI is structured to provide real-time monitoring, offering discovery of all agents deployed within digital estates. This visibility supports audit trails and data loss protection, including session recording features that allow for scrutiny of agent activity-enabling organisations to respond to any unexpected use or threat from AI agents.

Human oversight features ensure that sensitive actions taken by AI agents can be subject to review and require explicit consent. This design is intended to maintain a level of human involvement and accountability, particularly for decisions or operations with higher risk profiles.

Threat detection

The new solution includes measures for detecting and mitigating tampering or impersonation by unauthorised or rogue agents. Integrated gateways and security layers enforce policy-based restrictions and inject context-specific credentials only when required. These measures reduce reliance on static credentials, aiming to curb potential credential abuse or exposure.

The product's support for integrating with third-party credential vaults also enables organisations to use existing security infrastructure, while implementing secretless identity options for dynamic access management.

Deployment

Identity for AI is scheduled for general availability in early 2026. The initial feature set will include agent registration, session recording, human delegation features, and privilege management controls, among others. The company noted that further capabilities spanning agent visibility, governance, threat detection and privileged-access management are planned for rollout beyond the initial launch window.

Durand said, "Identity for AI will help enterprises engage the agentic commerce channel, secure the autonomous workforce, and protect against adversarial AI threats."