At The AI Conference in San Francisco, RSA launched Agent ID, a new agentic identity security platform built for critical infrastructure, healthcare, government, and finance sectors. According to MarkTechPost, the platform arrives as enterprises face significant risks from rogue AI usage, highlighted by instances where a single problematic prompt disrupted a company’s Salesforce instance. Agent ID introduces three core modules designed to manage both sanctioned and hidden AI tools across an organization.
The Architecture of Agent ID
The platform’s functionality is split across its three announced modules. The Discover module is built to track down both sanctioned and shadow AI agents, alongside Model Context Protocol (MCP) servers. The Secure module acts as an inline gateway, verifying every tool call against established company policies. Meanwhile, the Govern module maps compliance evidence directly to ten different regulatory frameworks. The Discover and Secure components are scheduled to ship on November 16, 2026.
Unknowns and Implementation Timeline
Despite the product unveiling, several operational details remain absent from public disclosures. The announcement did not specify pricing models for the platform, nor did it outline the technical mechanism by which the Discover module identifies the estimated 4,000 shadow AI agents operating within a typical enterprise. Additionally, the exact release date for the Govern module was not provided in the initial rollout details.
Organizations looking to manage enterprise agent sprawl will need to evaluate how the November rollout fits into their broader security architecture as further details emerge.
Lead image: Anete Lusina / Pexels
