CyberArk Launches Solution to Secure Agentic AI and Digital Identities

CyberArk, the global leader in identity security, has introduced the CyberArk Secure AI Agents Solution. This new solution allows organizations to adopt identity-first security for agentic AI using the CyberArk Identity Security Platform. The goal is to help mitigate new identity-centric risks as AI agents autonomously interact, access data, escalate privileges, and adapt behaviors to perform complex tasks.

According to Gartner, by 2028, 25% of enterprise breaches will stem from AI agent abuse, whether from external threats or internal bad actors. This emerging threat landscape arises from a new class of digital identities—AI agents that operate like humans in autonomy but scale like machines. Protecting these agents requires more than prompt security; it demands a layered identity security approach.

Identity Security Is Key to AI Resilience

“When millions of autonomous, adaptable AI agents gain privileged access, organizations must not let security lag behind innovation,” said Matt Cohen, CEO at CyberArk. “Basic identity and access controls are no longer sufficient. Agents must be secured from day one, blending human identity security principles with machine scalability. With CyberArk, organizations can adopt an identity-first model that secures agentic AI while maintaining trust and resilience.”

The Secure AI Agents Solution leverages intelligent privilege controls from the CyberArk Identity Security Platform. Each agent is treated as a privileged, autonomous identity subject to continuous discovery, oversight, and adaptive control. Key capabilities of the solution include:

• Discovery and context: Visibility into both known and shadow agents across SaaS, custom agents, and infrastructure.

• Privilege control: Enforcing least privilege, securing credentials such as secrets and certificates, and enabling behavioral monitoring for threat detection.

• Automated lifecycle management: Onboarding and offboarding agents to reduce stale access.

• Governance: Ensuring agent operations align with compliance and regulatory standards.

In support of this initiative, CyberArk has released an open-source toolset on its GitHub page. The CyberArk Labs AI Agent Tool Set helps developers understand agent behavior, assess risk, and provision credentials securely during development.

Additionally, CyberArk’s platform now includes CORA AI™, its embedded AI engine. CORA AI not only secures agentic AI but also improves overall platform security. It analyzes behaviors, detects threats, recommends automated responses, and allows natural language interaction for streamlined operations.

CyberArk’s new solution marks a significant step in helping enterprises unlock innovation with AI—without compromising control or security.