Enterprises Shift Toward Federated Control to Manage AI Autonomy

The report, authored by EMA President Dan Twing, highlights a significant disconnect between AI integration and operational reliability. While enterprises are embedding artificial intelligence into observability, cloud platforms, and service management, they are not granting total autonomy. Instead, companies are adopting an incremental approach, expanding AI execution only after verifying governance compliance and stability. This caution is well-founded: more than 75% of surveyed IT professionals reported needing to intervene, correct, or roll back AI-driven actions to maintain business continuity.

Rather than forcing convergence onto a single platform, the research suggests that the future of enterprise management lies in a "control plane" that coordinates across multiple, distributed operational domains. This framework functions as an orchestration layer, providing the necessary visibility and accountability to offset the risks of autonomous reasoning. By defining these boundaries, organizations aim to reconcile the need for high-speed automation with the strict requirements of enterprise-wide governance, ultimately ensuring that AI acts within predictable parameters.