With the rapid rise of autonomous AI agents capable of executing complex tasks and interacting with critical infrastructure, the cybersecurity landscape is facing unprecedented identity challenges. In response, CyberArk , the global leader in identity security, has introduced the CyberArk Secure AI Agents Solution, designed to deliver identity-first security for the new class of agentic AI. This solution helps organizations manage and secure the privileged access, identity lifecycles, and behavioral orchestration of AI agents—treating them not just as digital tools, but as privileged, autonomous identities requiring defense-in-depth protection.
Subtopics and Highlights:
1. Addressing a New Class of Digital Identities
As AI agents increasingly mimic human autonomy while scaling like machines, they introduce a growing attack surface. According to Gartner, by 2028, 25% of enterprise breaches will stem from AI agent abuse, driven by both external threats and malicious insiders.
The Secure AI Agents Solution acknowledges this paradigm shift, recognizing AI agents as new privileged identities requiring proactive oversight, access controls, and governance from day one.
“Relying solely on basic IAM will leave organizations vulnerable,” said Matt Cohen, CEO of CyberArk. “We must secure AI agents with the same rigor we apply to human identities—while scaling for the future.”
2. Comprehensive Identity-First Security for AI Agents
CyberArk’s solution leverages its Identity Security Platform to secure agentic AI across multiple fronts:
- Discovery & Context: Provides observability into both known and shadow AI agents across SaaS, custom, and off-the-shelf tools
- Privilege Control – Access Management: Enforces least privilege and manages credentials like secrets and certificates
- Privilege Control – Threat Detection: Offers real-time behavioral monitoring and drift detection to prevent misuse
- Lifecycle Automation: Enables secure onboarding, offboarding, and rightsizing of agent permissions
- Governance & Compliance: Ensures AI agents operate within policy and regulatory frameworks
This holistic approach aligns with the pace of innovation in agentic AI, ensuring that identity security does not lag behind operational scale.
3. Open Source Toolset for Secure AI Agent Development
To support developers, CyberArk also introduced the CyberArk Labs AI Agent Tool Set, an open-source collection on GitHub. This toolkit:
- Visualizes how AI agents interact and communicate
- Identifies potential security risks during development
- Includes just-in-time credential provisioning for enhanced security and developer efficiency
By engaging the open-source community, CyberArk fosters secure-by-design principles in the AI agent development lifecycle.
4. CORA AI™: Embedded Intelligence for Adaptive Security
The new solution is complemented by CyberArk CORA AI™, the platform’s embedded AI engine. CORA AI enhances security across the board by:
- Analyzing agent and user behavior
- Detecting anomalies and emerging threats
- Recommending automated mitigation actions
- Enabling natural language interactions with the platform
CORA AI not only secures agentic systems but also elevates the overall security posture of enterprises through smart automation.
As organizations embrace agentic AI to accelerate innovation, CyberArk ensures that security scales with autonomy. The Secure AI Agents Solution positions CyberArk at the forefront of identity-first security in the AI era—helping enterprises adopt agent-driven infrastructure with trust, control, and resilience.
By addressing agentic AI’s unique risks through intelligent privilege management, automated lifecycle controls, and developer-ready tools, CyberArk is setting a new standard for securing digital autonomy.
