Agentic AI: Visibility Gaps Threaten Federal Security

Artificial intelligence agents can accelerate federal missions and automate complex tasks. However, agencies cannot safely expand agentic AI without understanding the environments where these agents operate. Brian "Stretch" Meyer, federal field chief technology officer at Axonius, highlights significant visibility gaps when agencies introduce AI agents. He notes that in the rush to deploy new technology, basics like understanding devices, users, vulnerabilities, and applications can be overlooked. The biggest gap often lies with permissions. AI access is frequently based on user permissions, and many organizations are unaware that their users are already overprovisioned. This means users could potentially access sensitive systems they shouldn't. To address this, agencies need continuous discovery to track where agents are deployed and what they can access. Traditional inventories are often just snapshots, but agentic AI requires a "living operational map" because agents can connect to new systems and inherit new permissions over time. This matters because without clear visibility, agencies risk significant security vulnerabilities.