Beyond Identity: Governing the Agentic Enterprise
Akamai Technologies, Tuesday, September 22nd, 2026
Akamai argues enterprises must govern AI agent behavior in real time, not just verify identity.
This Akamai post argues that as AI agents gain autonomy to act across enterprise systems, verifying who or what has access is no longer sufficient.
It highlights risks including an expanding attack surface from the Model Context Protocol, unmanaged AI browser extensions holding sensitive data, and prompt injection vulnerabilities.
The piece recommends moving beyond traditional identity management toward real-time behavioral governance, edge-based API protection, and clear policies on employee AI tool usage. It positions these steps as essential to securing what it calls the 'agentic enterprise.'