Integrating Dialogic Agents in Student Self-Regulation Dashboards

Abstract

The integration of artificial intelligence (AI) and Learning Dashboards has opened new possibilities for supporting personalised and self-regulated learning (SRL). Zimmerman’s (2008) model of SRL, which outlines the phases of forethought, performance, and self-reflection, provides a useful framework for guiding the design of student-facing dashboards. Key features such as goal-setting, progress monitoring, and reflective evaluation can be effectively supported through learning analytics dashboards. However, the effectiveness of these tools relies on thoughtful design that aligns with SRL processes rather than simply presenting data. This paper proposes a hybrid dashboard model that integrates learning analytics visualizations with dialogic agents to scaffold metacognitive processes following an SRL model. Alongside this framework, we identify three core design principles for metacognitive awareness, transparent and dialogic engagement with learning data, and for reflective feedback across time. Dialogic agents play a crucial role in interpreting learning traces, initiating reflective conversations, and fostering goal-setting, monitoring, and evaluation. This conversational layer makes learning behaviors visible and meaningful, enhancing both learner trust and engagement. We also highlight how transparent visual feedback, when paired with personalized guidance, can empower students to make informed decisions about their learning. While this work is conceptual, it lays the groundwork for developing evidence-based, trustworthy analytics systems that actively promote student agency.