RSA and SailPoint ship guardrails for AI agents

RSA launched Agent ID, routing every agent call through a policy gateway with authorization down to the tool and argument, while high-risk actions like wire transfers and payments require approval from a named human via a separate phishing-resistant channel. SailPoint expanded its Agentic Fabric to find unregistered shadow agents and disable them at runtime. SailPoint research shows 79% of enterprises already run AI agents in production, but only about 2% have identity security tools to govern them.

• RSA Agent ID routes every agent call through a policy gateway
• High-risk actions wait for human approval via a secure channel
• SailPoint Agentic Fabric finds and disables shadow agents
• 79% of enterprises run agents, only 2% govern their identity