Permiso Identity Security is a platform that provides visibility, risk management, and threat detection for identities in AI environments. The platform uses a Universal Identity Graph to track human, non-human, vendor, and AI identities, including AI users, builders, and agents across multiple environments. The platform offers three main capabilities: discovery, protection, and defense. Discovery provides visibility into who has access, what activities they perform, and how AI agents behave. Protection involves continuous scanning for misconfigurations, weak controls, and over-permissioning to identify exploitable AI identities with prioritized remediation guidance. Defense uses real-time behavioral analytics and detection rules to identify identity abuse, including AI agent compromise, prompt manipulation, excessive data access, and escalation attempts. The platform tracks identities and their activities across IaaS, SaaS, and identity providers at runtime. It monitors shadow AI activity and provides alerts for AI-specific threats. The system correlates access and activity patterns to surface risk, potential misuse, and malicious activity involving AI. Permiso embeds AI security within its identity model, treating AI users, builders, and agents as identity constructs. This approach allows organizations to measure AI adoption while maintaining security controls across cloud environments.