Identity-based security requires us to clearly define who did what, making it difficult to overlook details. If an agent modifies infrastructure, moves data, or grants access, you need to trace the action back to a specific identity, under a given policy, with an approval or constraint chain. Governance isn’t optional; it’s the essential control framework for autonomous operations.
Cloud finops and cultural shifts
If you think cloud bills were unpredictable before, wait until you unleash systems that can decide to use more resources in pursuit of a goal. Agentic AI changes how cloud resources are leveraged by making consumption far more elastic, exploratory, and continuous. Agents will spin up ephemeral environments, run iterative experiments, call paid APIs, generate and store large artifacts, and repeat tasks until they converge—sometimes without a natural stopping point.
The old finops playbook of tagging, showback, and monthly optimization is not enough on its own. You need near-real-time cost visibility and automated guardrails that stop waste as it happens, because “later” can mean “after the budget is gone.” Put differently, the unit economics of agentic systems must be designed, measured, and controlled like any other production system, ideally more aggressively because the feedback loop is faster.
