Designing a Comprehensive Intelligent Human Resource Management Model with a Data-Driven Approach in Iranian Public Organizations

Abstract

The rapid advancements in modern technologies, particularly artificial intelligence, have introduced new challenges and opportunities for human resource management in public organizations. On one hand, the increasing volume of HR data and the complexity of decision-making in areas such as recruitment, training, retention, and performance evaluation emphasize the necessity of adopting data-driven approaches. On the other hand, bureaucratic constraints and traditional organizational structures in the public sector make the implementation of innovative models more complex. This study aims to design a comprehensive intelligent human resource management model tailored for Iranian public organizations. To achieve this, the research first reviews the theoretical literature and prior studies on digital HRM and AI applications. Then, through a mixed-method approach, empirical data were collected and analyzed using data analytics and structural modeling algorithms. The findings indicate that the proposed model encompasses four major dimensions: (1) talent identification and recruitment through resume mining and professional network analytics, (2) employee training and development via AI-based adaptive learning systems, (3) performance evaluation and optimization using predictive analytics, and (4) retention and employee satisfaction enhancement through sentiment analysis and behavioral data insights. The empirical validation of the model in three selected Iranian public organizations demonstrated a 25% increase in HR decision-making accuracy and a 30% reduction in administrative process time. From a scientific perspective, this study contributes by presenting a novel data-driven framework for HRM in public organizations. Practically, it provides actionable guidance for policymakers and HR managers to design and implement evidence-based HR strategies. Finally, recommendations for future development of the model and the integration of deep learning tools are presented.