Generative AI chatbot for teachers’ data-informed decision-making: Effects and insights

By prof. Jeongmin Lee
Department of Educational Technology
jeongmin@ewha.ac.kr

Jeongmin Lee, Professor in the Department of Educational Technology, recently published a study in Educational Technology & Society entitled “Generative AI chatbot for teachers’ data-informed decision-making: Effects and insights.” The study highlights the transformative potential of generative AI in strengthening teachers’ professional judgment through dialogue and reflection.

As data-informed decision-making becomes a central expectation in education systems worldwide, teachers are increasingly required to interpret student data and translate evidence into instructional action. While assessment results, learning analytics, and digital learning traces have dramatically expanded the volume of available data, many teachers continue to struggle with making meaningful sense of these resources. Existing data tools often prioritize numerical summaries and visualizations, offering limited support for the cognitive and interpretive work that underpins high-quality instructional decision-making.

This study addresses this gap by introducing a generative AI–based chatbot designed not as a prescriptive decision-making system, but as a reflective, human-centered professional support tool. Grounded in research on teacher professional learning and human–AI interaction, the chatbot is conceptualized as a dialogic partner that supports teachers’ sense-making processes. Through natural language interaction, the system encourages teachers to articulate instructional challenges, interpret student performance data, explore alternative explanations, and justify pedagogical decisions.

A defining feature of the chatbot is its commitment to preserving teachers’ professional agency. Rather than automating instructional decisions, the system provides adaptive cognitive scaffolding that promotes reflection, justification, and deliberate reasoning. This design clearly differentiates the chatbot from conventional dashboards and analytics platforms, which often emphasize efficiency and information delivery over professional judgment.

Findings from a mixed-methods empirical study demonstrate the effectiveness of this approach. Teachers who engaged with the chatbot showed significant improvements in key dimensions of data-informed decision-making, particularly in evidence-based interpretation and instructional justification. They became more capable of grounding decisions in data, articulating coherent rationales for their instructional choices, and systematically considering alternative interpretations of student learning patterns.

Qualitative analyses of interaction logs and teacher reflections further reveal how the chatbot shaped teachers’ thinking. The conversational format prompted teachers to make assumptions explicit, revisit initial interpretations, and engage more deeply with data than when using traditional analytic reports. Many participants reported that the dialogue-based interaction supported slower, more thoughtful reasoning and increased confidence in their instructional decisions.

Overall, this research positions generative AI as a scalable, on-demand professional learning companion for teachers. By reframing AI’s role from automation to dialogic cognitive support, the study offers a compelling model for the future of AI-supported teacher professional development. When designed responsibly, generative AI can strengthen—rather than replace—teachers’ professional judgment, contributing to more reflective practice and advancing research on human–AI collaboration in education

* Related Article
Jiwon Lee, Jeongmin Lee, Generative AI chatbot for teachers’ data-informed decision-making: Effects and insights, Educational Technology & Society, 28(3), 298–317, 2025.