<![CDATA[Teacher-led development of contextualized mathematics materials may be supported by generative artificial intelligence (AI), but evidence of independent teacher autonomy and student learning requires careful distinction from workshop output. This participatory design-based study describes an AI-assisted instructional-material workflow integrating Epistemic Learning Patterns (ELP), a local-context STEAM+X design orientation, and the onlinemath.id platform in West Lombok, Indonesia. Twenty-eight teachers participated in a facilitated, one-day workshop. Evidence consisted of a two-rater quality appraisal of one purposively selected Grade V worksheet, post-workshop teacher questionnaires and reflections, and initial platform records. The platform summary reported 62 course uploads; 44 students enrolled in courses associated with workshop teachers, 34 of whom had verified identifiers. The selected triangle-area worksheet scored 62.5/64 (97.7%) on the study rubric. Post-workshop self-efficacy averaged 4.19/5, and 24/28 teachers (86%) expressed an intention to share the approach. Among verified students, 32/34 completed a Module 1 quiz, and 30/34 (88.2%) received full marks. These findings support the feasibility and perceived acceptability of a scaffolded design process and document immediate platform use. They do not establish independent teacher autonomy, causal learning benefits, or sustained conceptual understanding. Further research should evaluate systematically sampled products, independent teacher performance, verified cultural specificity, and validated student outcomes.]]>
