Vol. 7, Issue 2, Part B (2025)

Employee disengagement continues to undermine organisational effectiveness, leading to significant costs through reduced productivity, diminished innovation, and higher turnover rates. Proactive identification of disengagement risk is essential in today’s evolving workplace, yet conventional HR approaches often lack predictive capability and empirical transparency. This study presents and validates the DIS-SCAN Model (Disengagement Signals-Cognition-Action Nexus), a comprehensive predictive analytics framework designed to integrate diverse behavioural and perceptual data sources-such as attendance patterns, sentiment survey results, and learning management activity-for the early detection of at-risk employees.
A cross-sectional analysis was conducted on data from 480 employees representing manufacturing, technology, and healthcare sectors. The research utilized both Structural Equation Modelling (SEM) and Partial Least Squares (PLS) path analysis to assess the relationships among key predictors and disengagement outcomes. Findings demonstrate that attendance rate, survey sentiment, and learning participation together account for 61% of the variance in disengagement risk, with the DIS-SCAN Model achieving robust fit indices and high classification accuracy (AUC = 0.87). The application of the model enabled a notable 27% increase in the rate of timely HR interventions for employees flagged as high-risk.
The DIS-SCAN Model emphasizes interpretability, transparency, and ethical governance, aligning with data protection standards such as GDPR and India’s Personal Data Protection Act. Its architecture supports seamless integration into existing HRIS platforms and can be tailored to specific organizational contexts. By leveraging real-time, multi-source data and advanced analytics, the model offers a substantial improvement over traditional, reactive engagement management strategies.
This research establishes the DIS-SCAN Model as an effective and scalable tool for enhancing employee engagement through predictive analytics. Further investigation is encouraged to extend validation across various organizational settings and to refine the model for broader cultural applicability.