International Journal of Science, Engineering and TechnologyAuthors: Marco Alvarez, Kenji Nakamura, Daniel Whitaker, Dr. Priya Raman, Dr. Elena Petrov, Robert Anderson
Abstract: Modern organizations manage large volumes of workforce data across multiple Human Capital Management modules, yet valuable relationships among employees, skills, roles, learning activities, and performance outcomes often remain hidden within fragmented enterprise data structures. Platforms such as SAP Success Factors generate interconnected data across recruiting, performance management, learning, succession, and employee central modules, but traditional relational analytics and rule-based reporting methods provide limited capability for uncovering deeper structural relationships within this ecosystem. This study proposes a Graph Neural Network (GNN) framework for cross-module talent relationship mining using enterprise workforce data integrated through SAP HANA Cloud. The framework models employees, skills, roles, training activities, and organizational hierarchies as nodes in a heterogeneous workforce graph, while relationships such as reporting structures, competency associations, learning participation, and performance interactions form edges that capture organizational connectivity. By applying graph neural learning and message-passing mechanisms, the proposed architecture identifies latent talent networks, skill adjacency patterns, collaboration clusters, and internal mobility pathways that are difficult to detect using conventional analytics. The approach demonstrates how graph-based machine learning can enhance enterprise talent intelligence by enabling deeper workforce insights, improving succession planning visibility, supporting data-driven career development strategies, and strengthening cross-module analytical capabilities within integrated cloud HR systems.
DOI: https://doi.org/10.5281/zenodo.19128428
