Paper Type
Complete
Abstract
Agentic artificial intelligence (AI) systems are transforming enterprise workflows by shifting from predictive support to execution–capable agents operating under delegated authority. This shift introduces governance challenges that extend beyond model transparency to the architectural control of autonomous action, including error propagation, compliance exposure, and accountability ambiguity. To address this gap, we develop a kernel theory–grounded governance–by–design framework that conceptualizes governance configuration as a higher–order construct comprising autonomy calibration, human–agent teaming, and embedded machine learning guardrails. Drawing on delegation and organizational control theory, we position autonomy as a configurable design variable requiring structured oversight aligned with workflow risk. Using a design science research (DSR) approach, we design and comparatively evaluate five governance configurations, demonstrating how architectural alignment influences workflow performance, compliance adherence, appropriate reliance, and accountability clarity. This study advances a prescriptive, theoretically grounded framework for governing execution–capable AI systems in enterprise contexts.
Paper Number
1483
Recommended Citation
Mutale, Wilfred; Kumar, Avanish; and Sivasubramaniam, Nagaraj, "Governing Agentic AI in Enterprise Workflows: A Design Science Approach" (2026). AMCIS 2026 Proceedings. 5.
https://aisel.aisnet.org/amcis2026/sig_odis/sig_odis/5
Governing Agentic AI in Enterprise Workflows: A Design Science Approach
Agentic artificial intelligence (AI) systems are transforming enterprise workflows by shifting from predictive support to execution–capable agents operating under delegated authority. This shift introduces governance challenges that extend beyond model transparency to the architectural control of autonomous action, including error propagation, compliance exposure, and accountability ambiguity. To address this gap, we develop a kernel theory–grounded governance–by–design framework that conceptualizes governance configuration as a higher–order construct comprising autonomy calibration, human–agent teaming, and embedded machine learning guardrails. Drawing on delegation and organizational control theory, we position autonomy as a configurable design variable requiring structured oversight aligned with workflow risk. Using a design science research (DSR) approach, we design and comparatively evaluate five governance configurations, demonstrating how architectural alignment influences workflow performance, compliance adherence, appropriate reliance, and accountability clarity. This study advances a prescriptive, theoretically grounded framework for governing execution–capable AI systems in enterprise contexts.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIGODIS