CerberusAI has raised $1.5M in pre-seed funding, betting that enterprises will need a dedicated security layer as AI agents multiply across their systems.
The startup is building what it calls a "trust layer" for the agentic future-a platform that watches over AI agents themselves, not just the data they access. It combines behavioral analysis with cross-agent attribution to spot when an agent goes rogue, while giving security teams a single inventory of every agent and tool in operation. The platform tracks what data each agent touches and flags drift or anomalous behavior in near-real time, with explainable alerts and flexible remediation options. The pitch: companies can adopt AI more aggressively if they have an independent oversight layer that catches problems before they become breaches.



.png&w=3840&q=75)












