Ping Identity says its new “Identity for AI” offering will become generally available globally by March 31, 2026, as enterprises increase use of autonomous AI agents and face new governance and control challenges.
In a statement, the company argued that traditional identity and access management approaches focused on login and authentication are insufficient for autonomous agents, where risk can emerge from what agents do after access is granted. Ping framed the issue as a need for continuous and contextual enforcement during runtime, not just identity lifecycle management.
The product comprises three components: Agent IAM Core, Agent Gateway, and Agent Detection. Ping said the tools are intended to establish an agent’s identity, enforce delegated authority at runtime, and detect “agentic” activity.
Ping Identity CEO and founder Andre Durand said organisations will need “continuous verification and enforcement at every decision” as agents operate at scale against systems of record. The company also said it is aiming to govern human and non-human identities under a single framework.
Partner organisations referenced in the release echoed the need for stronger controls around autonomous agents. Deloitte’s Chad Veldhuizen said agent autonomy is “only as safe as the identity and access controls behind it,” calling for auditability across the identity lifecycle. Cloudflare’s Kyle Krum said AI agents introduce risks tied to actions after deployment, and highlighted guardrails and least-privilege enforcement as key to protecting sensitive data.
Ping said Identity for AI is designed to support delegated, scoped tokens rather than human impersonation, and to evaluate agent requests in real time. It also said the Agent Gateway can provide a runtime enforcement layer for agent-to-service interactions and centralised monitoring and audit, while Agent Detection (via PingOne Protect) can identify personal and external AI agents using behavioural signals and bot authentication protocols.
The release also pointed to support for emerging AI integration frameworks including the Model Context Protocol (MCP), which it said is intended to enable secure agent-to-system interactions without requiring changes to existing services.
Ping said a webinar on the offering is scheduled for April 7, 2026.
