Mimecast Agent Risk Center Preview: Runtime Data Security for AI Agents — Only 14% of Fortune 500 Have Full Security Approval for Active Agents

Help Net Security·medium signal

Mimecast expanded Incydr with the Agent Risk Center (Early Access September 2026), revealing that 80% of Fortune 500 companies now run active AI agents but only 14% have full security approval. The platform continuously scores both human users and AI agents on behavioral anomalies, maps which agents access which sensitive data categories (PII, source code, financial records), and enforces governance across commercial agents, MCP servers, and user-developed agents with sanctioned/unsanctioned classifications and department-level policy enforcement.