Аңдатпа
This article examines the design and practical implementation of an intelligent multi-agent system for monitoring educational achievements in the context of the digitalization of secondary education. The relevance of the study is обусловлена the need to shift from retrospective knowledge assessment to preventive management of individual learning trajectories based on the analysis of a student’s “digital footprint.” The authors propose a system architecture that includes three types of specialized agents: a Diagnostic Agent, a Methodological Agent, and a Predictive Agent. The mathematical framework of the system is based on hierarchical models for classifying learning risks according to four key criteria: academic performance, attendance, activity, and learning pace. The developed software, implemented in Python, follows the concept of Explainable Artificial Intelligence (XAI), ensuring transparency and interpretability of diagnostic outcomes for educators. During the system’s testing on a student sample, significant positive results were obtained: automation of the monitoring process reduced the time required for preparing class analytical reports by 70- 80%, while simultaneously improving the accuracy of final grade predictions. The system automatically generates personalized intervention scenarios tailored to the specific type of identified learning barrier. The discussion section outlines unresolved issues, including the dynamic adjustment of weighting coefficients in the utility function and the need to integrate the developed solution with national educational information systems for academic performance tracking.