The AI productivity-governance frontier: A theoretical model for enterprise value creation under agentic automation

Abstract

Artificial intelligence (AI) adoption is accelerating, yet enterprise value remains uneven because technical capability often outpaces organizational redesign, workforce adaptation, and governance maturity. This paper develops an original theoretical model, the AI productivity-governance frontier (PGF), to explain why the same agentic AI capability can generate measurable value in one organization but produce negligible or negative returns in another. Using integrative theoretical modelling, the study synthesizes recent empirical evidence on generative AI productivity, enterprise adoption, AI risk management, labor-market exposure, and prior conceptual work by Kwan Hong TAN on AI-form organizations, AI stakeholder recognition, and temporal displacement-adaptation equilibrium. The resulting PGF model formalizes AI value as the interaction between automation-augmentation gains, learning spillovers, decision velocity, scalability, and institutional absorptive capacity, offset by governance drag, risk externalities, and identity-coordination costs. The paper proposes six testable propositions and a practical maturity pathway moving from experimental AI use to validated autonomy. The central argument is that sustainable AI value does not increase monotonically with either automation intensity or governance intensity. Instead, organizations approach maximum value when they design human-AI work systems that combine use-case fit, accountable autonomy, adaptive reskilling, and proportionate assurance. The contribution is threefold: a formal value equation for enterprise AI, a governance-sensitive interpretation of AI productivity heterogeneity, and an implementation framework for managers, policymakers, and researchers studying AI-enabled business transformation.