AI security startup Cyata Security has closed an $8.5 million seed funding round. The investment, led by TLV Partners, will fund the development of its platform for managing autonomous AI agents. The round also included participation from former executives of Cellebrite DI Ltd.
The company addresses a critical gap in enterprise cybersecurity. AI agents, designed to automate tasks, often gain unrestricted access to critical systems and data. Consequently, they operate outside of traditional identity access management frameworks, creating significant risk without proper oversight or governance.
Controlling Autonomous AI Risks
Cyata has developed an agentic control plane to provide visibility and control over these systems. Its platform offers a new approach to **AI agent security** by focusing on the agents themselves, not the large language models that power them.
The technology automatically discovers AI agents operating within a company’s cloud and SaaS environments. It analyzes behavioral patterns to identify agent activity. Furthermore, the system can lock down unauthorized agents and enforce least-privilege access for approved ones, preventing potential data leaks or malicious actions. This is a necessary evolution, as tools from providers like Okta or CyberArk were designed for human users, not dynamic AI.
Cyata also provides detailed forensic observability. This creates an audit trail of an agent's actions within environments like Microsoft Azure, capturing its intent in real time.
This specialized approach to **AI agent security** is crucial. The rapid, autonomous, and often temporary nature of AI agents makes them invisible to conventional cloud security tools. The new funding reflects investor belief that agentic AI adoption will increase dramatically, creating an urgent need for solutions that can prevent catastrophic breaches. Therefore, Cyata aims to define a new category of **AI agent security**.
