Agentic AI-based human resource systems and organizational performance: The sequential mediating role of data-driven decision making and workforce adaptability
Abstract
The rapid development of artificial intelligence is driving human resource management to transform from its traditional administrative support function into an intelligent strategic system capable of strengthening an organization's core capabilities. A new generation of agentic artificial intelligence (Agentic AI) human resource systems with independent adaptive capabilities has gradually been implemented in enterprise application scenarios. Currently, various types of AI-enabled human resource systems have become widespread in the market, but both academic and industry communities still lack empirical research on the internal mechanisms through which these agentic AI human resource systems impact organizational performance. This study is conducted specifically to fill this gap. This study adopts a quantitative research design. Using SmartPLS 4 software, it analyzes cross-sectional survey data through partial least squares structural equation modeling (PLS-SEM). Its core variables include the independent variable agentic AI human resource system, the mediating variables data-driven decision-making and workforce adaptability, and the dependent variable organizational performance. This study relies on dynamic capabilities theory, information processing theory, and the knowledge-based view as its core theoretical supports. Empirical findings show that the core independent variable has a significant direct positive effect on the dependent variable, and can also produce a significant indirect positive effect transmitted through two mediating variables. Among these mediators, employee adaptability exerts a stronger impact on organizational performance than data-driven decision-making. In subsequent work, this paper will sort out the study's theoretical contributions, put forward practical suggestions for corporate practice, and clarify potential directions for future research expansion.