Non-human identities in Fortune 500 companies grew 480% in the last six months, according to executives at Cyera, and enterprises have no visibility into most of them. “The moment an AI agent is deployed, it inherits elevated human permissions and moves at machine speed, touching data in seconds that would take a human a career to reach,” said the company. And until now, there’s been little infrastructure purpose-built to govern what agents can see and do, leaving sensitive data, critical systems, and regulatory obligations increasingly at risk.
Yesterday at Black Hat 2026, Cyera launched Agent Guardian, built to make every AI agent as visible and accountable as a human employee. Agent Guardian goes beyond the prompt and the response, monitoring every tool call, database query, and reasoning step in between, said the company.
Alongside Agent Guardian, Cyera also introduced Cyera Endpoint, bringing AI guardrails directly to employee devices for local agents that operate outside corporate network controls, where they are increasingly doing their most sensitive work. Together, they extend the trust layer Cyera is building for the agentic enterprise, answering what an agent can reach and what it should be allowed to do with it.
“AI agents have become a new class of enterprise identity. But the fundamental shift isn’t that organizations have more identities; it’s that the gap between permission and execution has disappeared,” said Tamar Bar-Ilan, cofounder and CTO of Cyera. “Agent Guardian and Cyera Endpoint give organizations the visibility and runtime control to govern this new reality, enabling them to accelerate AI adoption without putting sensitive data at risk.”
Whether enterprises are deploying agents built on AWS Bedrock or managing developer agents running on employee devices or rolling out Microsoft 365 Copilot, Agent Guardian secures the entire agentic surface from one platform, giving security teams the visibility to see every agent, the precision to govern what it touches, and the runtime control to stop it before it causes harm, said the company.
Agent Guardian enforces policy wherever AI interacts with enterprise data, including through platform APIs, AI gateways, browser extension, agent framework hooks, and remote scans of employee devices. That last control point has become increasingly important as developers adopt local agents such as Claude Code and Cursor, where activity never crosses a network boundary and traditional security tools have little or no visibility.
To add runtime protection to where local activity happens, Cyera today also introduced Cyera Endpoint, a lightweight endpoint control that brings Agent Guardian’s policies directly to the device. It provides live visibility into sanctioned and shadow local agent activity, classifies and protects sensitive files before they’re accessed or leave the machine, and blocks data from reaching unauthorized personal AI accounts, enabling consistent governance across cloud, SaaS, and endpoint environments.
