Why Just-in-Time Authorization for AI Agents Depends on Intent

Over the past year, most of our conversations with security leaders about AI agents started with discovery. Teams wanted to know which agents were operating in their environment, which MCP servers