Yetisensoy, Okan2026-09-012026-09-0120261049-48201744-5191http://dx.doi.org/10.1080/10494820.2026.2637614https://hdl.handle.net/20.500.12403/8506Formative assessment practices have been a significant focus in Social Studies education for many years. One commonly used technique in these practices is the diagnostic branched tree, which is designed to assess students' understanding through a branching structure based on their responses. While this technique shows promise for identifying students' learning status and gaps throughout the ongoing process, it presents several structural limitations. This research aims to address these limitations by integrating diagnostic branched trees with natural language understanding, a key AI technique in educational assessment, thereby enhancing their diagnostic capabilities and making them smarter tools in educational settings. Furthermore, this research seeks to evaluate the impact of this revised system in middle schools within a practical setting. Following the development phase, the results of the experimental study demonstrated that the revised system had a significantly positive impact on students' learning performance. Qualitative data, meanwhile, suggested that this success was largely due to the system's ability to provide formative feedback and personalized guidance, which fostered metacognition and self-regulated learning. At this point, considering the positive transformations brought about by technology, it is recommended that innovative technologies be developed to enhance formative assessment across various levels of K-12 education.eninfo:eu-repo/semantics/closedAccessArtificial IntelligenceDiagnostic Branched TreeFormative AssessmentNatural Language UnderstandingSocial Studies EducationArticle10.1080/10494820.2026.26376142-s2.0-105033979512Q1WOS:001726754700001Q1