SINGAPORE (July 2026) — Companies are letting autonomous AI agents run wild on their systems without any supervision, creating a massive security gap that hackers are already looking for.
Keeper Security is stepping in to fix this by treating these digital workers exactly like human staff. This means they now have to follow strict security rules and go through the same checks as everyone else.
Why this matters for your data
Most companies are rushing to use AI but forgetting about the locks on the door. Research shows that 63 percent of organizations have zero AI governance policies in place. Even worse, almost all companies that got hit by an AI-related breach did not have proper access controls.
The new feature identifies the AI agents already running on workstations, like GitHub Copilot or Claude Code. It uses a special algorithm to score every app and flag anything that acts like an AI agent, even if it is not on a known list.
Watching the machines at work
While other tools only watch the “chat” layer, this new feature watches what the AI actually does on the machine itself. It can stop an AI from writing files or changing system settings without permission.
Darren Guccione, CEO and co-founder of Keeper Security, says AI agents are not just helpful assistants: they are active participants on the network. He warns that if businesses do not govern them with the same rigor as people, they leave a blind spot that attackers will find first.
Craig Lurey, CTO and co-founder of Keeper Security, adds that this update stops the emerging threat of autonomous AI in its tracks. This allows enterprises to adopt cutting-edge tools without opening the floodgates to catastrophic risk.
How to get it
Agentic AI governance is now available through the Keeper Endpoint Privilege Manager. It can be used as a standalone product or as part of the larger KeeperPAM platform. Interested users can find more details or request a demo on the official website.
Note: Philippine pricing is currently unavailable.
