What happened
C1 announced (June 18, 2026) general availability of enterprise-managed authorization (EMA) support, issuing short-lived, scoped tokens for AI agents accessing MCP servers. Anthropic shipped the first implementation in Claude on the same day. C1's Access Gateway enforces governance across both MCP-compatible and legacy systems via a unified control plane.
Why it matters
EMA solves the critical non-human identity risk in agentic AI: unmanaged agent credentials enable lateral movement and unauthorized tool access. C1's approach treats AI agents like human users in access reviews, revocation workflows, and audit trails — essential as agents proliferate across enterprise tools.
Applicability
Enterprise security teams deploying AI agents should implement agentic identity management immediately; C1 is now production-ready, integrating with Claude and any MCP server.