Summary
As AI agents increasingly become part of the modern workforce, accessing critical systems, organizations are struggling to secure them with only 21% implementing specific governance controls. A framework is proposed to manage these non-human identities, starting with discovery, formal registration, and managing access with least privilege.