This study aimed to develop AI-assisted authentic assessment templates to support the monitoring of early childhood development. Preliminary findings from five early childhood education institutions in Bantul Regency, Special Region of Yogyakarta, Indonesia, revealed that teachers faced difficulties in implementing authentic assessment due to fragmented documentation and challenges in interpreting and narrating children’s developmental records. To address these issues, this study employed a Research and Development (R&D) approach using the ADDIE model, involving teachers and school principals throughout the Analysis, Design, Development, Implementation, and Evaluation phases. The developed product consisted of structured, authentic assessment templates, including work assessment templates, observation checklists, anecdotal notes, and assessment rubrics integrated with NotebookLM to support AI-assisted analysis of children’s developmental data. Expert validation yielded an average score of 4.4 out of 5.0 (88%), indicating that the developed assessment templates were feasible for implementation. The implementation results showed that the templates enabled teachers to document children’s development more systematically, organize assessment data, and generate individualized narrative reports with AI assistance. Teachers remained responsible for reviewing, validating, and interpreting AI-generated outputs before using them to make decisions about appropriate stimulation strategies for children at school and at home.