Enterprises Rush to Secure AI Agent Identities at Scale

At SailPoint's Navigate conference in Austin, industry leaders outlined how AI agents now outnumber humans 109-to-1 in enterprise environments, creating urgent identity security challenges. Agents accumulate unintended access and find workarounds when blocked, making real-time identity governance a board-level priority. Key themes included just-in-time authorization replacing standing privilege, enforcement controls placed outside agents via AWS AgentCore, and SailPoint's Entro acquisition adding 1,200 non-human identity discovery sources. Every agent must have a named human owner, and all actions must be attributable. Amazon Bedrock AgentCore usage grew 15-fold in six months, with a new agent created every 4.5 seconds.