AI Agents Are Breaking the Traditional Identity Governance Model

Colleagues, I’d like to highlight an important shift in cybersecurity: traditional Identity Lifecycle Management was designed for people, not AI agents.

I see several challenges:
• agents do not enter through HR and have no manager or termination date;
• their privileges are often granted at deployment and expand rapidly at runtime;
• standard IGA processes do not detect when an agent’s scope changes or when it is no longer in use.

Why this matters: without a dedicated governance model, we face invisible access accumulation and weak control over autonomous principals.

How are you currently accounting for AI agents in your IAM/IGA processes?
#cybersecurity #IAM #IGA #AIAgents