Name: Shashikumar R, Keerthi H K
Country: India

Year: 2026
Volume: Volume-11, Issue-1, January-June 2026
Page Number: 980-995
DOI: https://doi.org/10.5281/zenodo.22829383

Organizations increasingly make decisions in settings where employees, customers, suppliers, technologies, alliances and other stakeholders are densely interconnected, so that an action affecting one entity can propagate consequences across a wider relational system. Graph theory offers a natural language for representing such interdependence, and over recent decades its principles have diffused into network analysis, business analytics, graph databases, knowledge graphs, graph machine learning and graph neural networks (GNNs). More recently, the emergence of explainable artificial intelligence (XAI) and decision intelligence has opened a further possibility: moving graph-based analysis beyond describing relationships and predicting outcomes toward explaining, contextualising and informing managerial choice. The literature supporting this trajectory, however, remains fragmented across mathematics, computer science, information systems, operations research and management, with graph prediction and graph explanation developing considerably faster than research connecting these outputs to constrained, context-sensitive managerial decisions. This review synthesises that fragmented landscape through a critical, integrative methodology and proposes Graph-Based Explainable Decision Intelligence (GEDI) as an organizational capability comprising relational sensing, graph analytics, predictive intelligence, explainable intelligence, decision integration and adaptive reconfiguration. Rather than treating graph theory, GNNs or XAI as novel in themselves, the review's contribution lies in integrating these previously separate streams into a management-oriented architecture and in identifying where the evidence base is mature and where it remains conceptual. The review closes with a research agenda organised around dynamic, causal, human-governed and prospectively agentic graph-based decision intelligence, together with guidance for management, information systems and decision-support scholarship.

Keywords: graph theory; graph neural networks; explainable artificial intelligence; decision intelligence; organizational networks; dynamic capabilities; management research