As enterprises rapidly adopt autonomous AI agents such as Claude Cowork, security teams are facing a new blind spot: unmanaged AI activity occurring directly on endpoints. Employees increasingly ...