The low level of middle school students' numeracy literacy based on the National Assessment is a crucial challenge stemming from abstract mathematics learning detached from cultural contexts. This study aims to develop and test the effectiveness of an AI-Exploration-based Smart E-Module that transforms the traditional Dakonan game into a Realistic Mathematics Education (RME)-based Learning Trajectory. The research method used is Design Research of the Validation Studies type, consisting of three phases: Preparing for the Experiment, The Design Experiment, and Retrospective Analysis. The subjects involved eighth-grade students from a partner junior high school in Malang Regency (with a small pilot group of 6 students in Cycle 1 and an intact class of 32 students in Cycle 2 teaching experiment). Furthermore, qualitative interaction logs indicate that AI chatbot scaffolding assisted students in overcoming individual learning hurdles and sustained student engagement. Quantitative analysis revealed an increase in the mean numeracy score from 52.40 to 78.50 with an N-Gain value of 0.55 (moderate category) and significance p < 0.05. This study contributes to the formulation of local instructional theory based on ethnomathematics and adaptive technology.
Copyrights © 2026