From Voluntary Practice Towards Actionable Learning Analytics: Architecture, Feature-Level Use, and Design Implications for an AI-Supported Learning Platform
TL;DR
Structured practice merits refinement; teacher-facing analytics require evaluation for actionability; Structured practice merits refinement; teacher-facing analytics require evaluation for actionability.
Abstract
INTRODUCTION: Voluntary use of AI-supported learning platforms may differ across features, while aggregate logs cannot show whether teachers act on analytics.
Objectives
To evaluate AI Sensei’s architecture, logged activity across platform features, self-reported feature use and student feedback.
Methods
An observational pilot analysed logs for 205 students and an anonymous questionnaire (N = 100). Training quizzes and feedback were programmed; chat used generative AI.
Results
All 205 students submitted training quizzes and self-check tests (1,999 and 481 submissions, respectively); 164 (80.0%) sent at least one chat message. In the separate anonymous questionnaire, 91% selected quizzes/tests and 17% the chat assistant; 94% agreed that combining teacher support with AI Sensei was more beneficial than teaching without the platform.
Conclusion
Structured practice merits refinement; teacher-facing analytics require evaluation for actionability.