Leadership Agility and Innovation Management in Industry 5.0 Organizations

Abstract

Leadership agility and innovation management have become essential for enabling human-centric, intelligent and sustainable organizational transformation in Industry 5.0 through effective integration of Artificial Intelligence (AI), digital technologies and organizational intelligence. Current leadership models, however, perform poorly in the area of decision making, lack collaboration analysis, poor innovation prediction and lack of explainability, resulting in lower adaptability of the organization in changing business environments. To overcome these challenges, the Human-centric Graph Transformer Optimization Network (HGTON-Net) has been developed to learn the optimization strategies for human agility in leadership and innovation management, intelligent, adaptive and explainable decision support. The framework integrates the Transformer Encoder (TE) to learn from the context, the Graph Attention Network (GAT) to capture relationships, the Temporal Fusion Transformer (TFT) to forecast trends in innovations, the Soft Actor-Critic (SAC) algorithm to fine-tune the leadership approach and the SHapley Additive exPlanations (SHAP) to interpret the choices. The integrated workflow preprocesses heterogeneous organizational data, learns contextual and graph representations, predicts future innovation demands, optimizes strategic decisions and generates explainable recommendations. Experimental evaluation demonstrates the effectiveness of HGTON-Net, achieving a Leadership Decision Accuracy of 96.12%, confirming superior performance over existing approaches. Overall, HGTON-Net provides an accurate, adaptive and trustworthy AI framework for sustainable leadership and innovation management in Industry 5.0 organizations.