Cyberhaven deploys Agentic AI Security to address the biggest blind spot in enterprise AI governance
Mountain View, California, March 24, 2026 /PRNewswire/ — Cyberhaven, the global leader in AI-native data security, today announced the launch of Agentic AI Security to address the rapid growth of shadow agents. This modern solution combines visibility, observability, and control to help security teams proactively manage AI systems running on endpoints across the enterprise and unlock the potential of agent AI.
Endpoint AI agent usage and risk are growing rapidly
AI is rapidly moving from chat-based tools to autonomous agents running on endpoints. These agents are rapidly developed and widely adopted, increasing the level of access to data and systems. While this dramatically improves productivity, it also introduces new security challenges that are largely unaddressed. It’s a shadow agent, an AI system that operates outside of a company’s visibility and control.
While enterprises have begun to implement governance for chat-based AI tools, those efforts have not kept pace with the rapid rise of AI agents. According to a recent study by Cyberhaven Labs, companies Endpoint-based AI agents have grown 276% in the past year, more than three times the growth rate of GenAI SaaS toolsThis marks a rapid transition to autonomous systems operating outside of traditional security controls. Meanwhile, adoption of endpoint coding assistants will more than double in 2025, jumping from 20% to 50%.
“AI no longer just generates content, it performs work.” Cyberhaven CEO Nishant Doshi said: “These agents have access to data, tools, and systems, and operate with a level of autonomy the industry has never seen before. However, most governance programs are still focused on what users input into the AI, rather than what the AI agent is actually doing. Security cannot work after the fact. It must work in real time, once the AI starts working.”
This represents a fundamental change in the threat model.
Most existing AI security solutions are built to answer questions such as:
- Who is using ChatGPT?
- What data is shared with Google Gemini?
However, answering these questions alone is not enough, especially when endpoint-based models are gaining traction among users while primarily focusing on mainstream web-based models.
Security teams must now answer:
- What agents are running on the endpoints?
- What data are they accessing?
- What actions are they taking?
Until today, I couldn’t do that.
Introducing Cyberhaven’s Agentic AI Security
To address this change, Cyberhaven is extending its unified AI and data security platform to secure autonomous agents at the endpoint. In other words, a new control plane for running AI.
The platform introduces a new approach to AI security built around three pillars:
- visibility – Discover and inventory running connections on AI agents, MCP servers, and endpoints
- observability – Monitor agent behavior including data access, tool usage, and execution paths
- control – Apply real-time guardrails during execution to prevent data leakage and unsafe or unauthorized actions.
As agent AI adoption accelerates, endpoints are becoming the primary execution layer and primary risk surface for AI. Organizations that continue to rely on SaaS-only visibility and API monitoring miss a large portion of agent activity and risk.
Cyberhaven defines a new category of AI-native data security for autonomous systems, enabling organizations to securely deploy agent AI while maintaining control over how their data is accessed, used, and manipulated.
To learn more about Cyberhaven and how it helps organizations secure their agent AI, visit booth S-1355 at the RSA conference.
About Cyber Heaven
Cyberhaven protects sensitive data no matter where it resides, no matter where it resides. Built for the AI era, Cyberhaven’s unified data security platform combines DSPM, data loss prevention, insider risk management, and AI security with deep data lineage and agent AI. Cyberhaven helps organizations stop data loss, reduce insider risk, and securely enable AI deployments without slowing down the business. For more information, please visit www.cyberhaven.com.
Source: Cyberheaven
