The main contribution of this paper is to advance the integration of human sustainability practices within software systems, particularly within a Learning Management System (LMS) to address the social dynamics of communities around software. We leverage cognitive computing, which seeks to emulate human cognitive functions, to enhance the early detection of student dropout at Universidad EAFIT. Early dropout risk alerts is currently a process based on measuring student absences and academic performance. Pre-conceptual Schema (PCS) are computing models for domain representation of software systems. We propose a representation using the PCS that incorporates cognitive functions: perception, reasoning, and action for including human sustainability practices as continuous feedback, improvement, monitoring, and observability to a LMS. PCS representation is an advance to the model to avoid the risk of student dropout by facilitating proactive intervention and complementing support to the current early warning systems.