AIR Raises $50M to Keep AI Agents From Doing Harm at Your Company

AIR, a new security startup founded by two veterans of Israel's Unit 8200 intelligence corps, just emerged from stealth with $50 million in funding. The company's mission: help companies watch over the AI agents that now work inside their systems. AI agents are not just chatbots — they can take actions like reading databases, booking meetings, or ordering supplies. AIR treats them like new employees that need monitoring.

Why should you care? If your employer uses AI agents, those agents can now access sensitive information and interact with the internet. A bug or a malicious plug-in could cause them to leak data, make bad purchasing decisions, or get tricked by attackers. AIR says it can find every agent running in a company, continuously check the skills and add-ons they use, and block anything dangerous — like antivirus for AI.

The bigger issue is the software supply chain. Just like apps on your phone, AI agents rely on add-ons and "skills" built by third parties. Those add-ons can be updated or hijacked, turning a safe tool into a security risk. AIR maintains a whitelist of vetted components and filters out about 27% of what it finds online. It already has more than 20 customers, mostly in banking and pharma — industries where mistakes are costly.

The catch? This is a young field with many rivals, including funded startups like Zenity and Noma. Also, AIR's system only works if companies let it see everything their agents do. And no security tool is perfect — determined attackers will keep looking for gaps. Still, having a watchdog for AI agents is becoming essential as more companies let these digital workers loose.