CyberArk Launches Secure AI Agents Solution to Address Risks

CyberArk has announced the release of the CyberArk Secure AI Agents Solution, developed to help organisations implement identity-first security for agentic AI environments. Built on the CyberArk Identity Security Platform, the solution is designed to mitigate evolving identity-centric risks introduced by AI agents operating autonomously within enterprise environments.

Addressing Risks in Autonomous AI Agent Interactions

AI agents are increasingly used to perform complex tasks, including accessing sensitive data, escalating privileges, interacting with infrastructure, and modifying behaviors without human intervention. As these agents communicate with one another and operate independently, they introduce new challenges in identity and access management.

CyberArk's new solution focuses on securing these agents by managing their privileged access, lifecycles, and orchestration. This goes beyond prompt-level security and requires a comprehensive identity security strategy.

AI Agent Abuse: A Growing Threat Surface

Gartner has projected that by 2028, 25% of enterprise breaches will be traced to the abuse of AI agents, involving both external attackers and internal threat actors. This emerging threat landscape is linked to the rise of a new category of digital identities—AI agents that exhibit human-like autonomy but machine-like scalability.

CyberArk's approach provides a defence-in-depth framework to help organisations secure these identities and reduce the risk of misuse or compromise.

"When millions of autonomous, adaptable, and interactive AI agents gain privileged access to resources and services, organisations must not find themselves in a situation where security has lagged innovation. Relying solely on basic identity and access management controls will leave organisations vulnerable to breaches they won't see coming," said Matt Cohen, CEO at CyberArk, "Agents must be secured on day one by combining the principles of human identity security with the scalability and automation of machine identity security. With CyberArk, organisations can plan for an identity-first model to secure the future of agentic AI, unlocking innovation while maintaining control, trust and resilience."

CyberArk has introduced new capabilities through its Secure AI Agents Solution, which leverages the CyberArk Identity Security Platform to secure agentic AI environments. The platform applies intelligent privilege controls across all environments, treating each AI agent as a privileged, autonomous identity subject to continuous discovery, oversight, and adaptive control. These capabilities are integrated natively to align with the evolving nature of AI agent usage in enterprise settings.

Key Capabilities of the CyberArk Secure AI Agents Solution

The solution includes the following core functions to support secure deployment and operation of AI agents:

• Discovery and Context: Offers visibility into known and shadow agents across SaaS platforms, off-the-shelf and custom-built agents, and agent-related infrastructure.

• Privilege Control – Access Management: Enforces least privilege principles, manages credentials such as secrets and certificates, and secures access for agents with elevated privileges.

• Privilege Control – Threat Detection and Response: Enables real-time behavioral monitoring to identify anomalies and prevent misuse.

• Automated Lifecycle Management: Supports secure onboarding and offboarding of AI agents, helping eliminate unused or excessive access rights.

• Governance: Ensures agent activity aligns with organisational and regulatory compliance requirements.

Open Source Tools for AI Agent Developers

To support development efforts, CyberArk has released an open-source AI Agent Tool Set via its GitHub repository. This toolkit provides visibility into agent communication patterns, highlights potential risks, and includes features such as just-in-time credential provisioning to enhance security during development.

Role of CORA AI in Agent Security

CyberArk’s embedded AI engine, CORA AI, further strengthens the platform by improving overall security through behavioral analysis of users and agents. CORA AI enables:

• Threat detection through pattern recognition.

• Automated security response recommendations.

• Natural language interaction for platform administration.

These enhancements help streamline operations and improve response time, particularly in complex agentic AI environments.

Read More:

India as a Fast Growing Market is Important for Business Growth

Servers and Data Storage both are the Pillars

CAIT and DGST Host Conference Marking 7 Years of GST

We Emphasise on Interdependence of AI and Cloud technologies