This study aims to develop a governance-based explanation of how agentic artificial intelligence (AI) can be embedded within lean six sigma systems without undermining statistical discipline. It introduces agentic operational capability to explain how firms design and supervise bounded autonomy within statistical control routines and identifies the conditions under which recursive adjustment stabilizes or destabilizes operational performance.
The paper adopts a conceptual theory building approach. It integrates lean six sigma principles, dynamic capabilities theory and research on algorithmic governance to construct a capability cascade model. Propositions specify how bounded operational autonomy influences orchestration, Operational Excellence 5.0 maturity and sustainability aligned performance under varying governance and institutional conditions.
When detection and corrective adjustment occur within the same operational loop, statistical control becomes partially endogenous. Sigma stability then depends not only on process variance but also on supervision of recalibration logic. Agentic operational capability enables delegation of decision authority within defined statistical limits. Its effects are conditional. Strong governance of thresholds and escalation rules can preserve stability and embed sustainability constraints within operational routines. Weak oversight or regulatory volatility may instead amplify variation and coordination instability.
As a conceptual study, the framework requires empirical validation. It reframes statistical control as a recursive governance problem and identifies moderators that shape stability outcomes under bounded autonomy.
Managers must treat AI integration as a control design challenge. Policymakers should support transparency and stable institutional environments.
The study positions governance design, rather than technological intensity, as central to Operational Excellence 5.0.
