Enhancing dashboards using generative artificial intelligence

Purpose This study aims to enhance the interpretability and actionability of learning analytics (LA) dashboards for teachers by integrating generative artificial intelligence (GenAI) to produce tailored narratives. Design/methodology/approach The research employs a mixed-methods approach, involving the design and implementation of GenAI-driven narrative features within existing dashboards. Prompt engineering techniques were used to generate narratives based on real course data. The approach was evaluated using questionnaires, interviews and structural content analysis of generated narratives. Findings The integration of GenAI-generated narratives improved teachers’ understanding and interpretation of dashboard data, supported decision-making and increased willing of adoption of the enhanced dashboards. Teachers highlighted the value of contextualized explanations and reported clarity and ease in interpreting student LA. However, the study also identified challenges related to data quality, transparency of GenAI outputs and the need for human oversight. Originality/value This work offers an innovative application of GenAI in educational technology by moving beyond data visualization to provide context-rich, automatically generated narratives. It advances the field by proposing and empirically evaluating specific prompt engineering strategies, including the segmentation of prompts into explanation, interpretation and recommendation segments, as well as mitigation of Large Language Model hallucinations through data injection. It addresses a critical gap in making LA dashboards more accessible and actionable for educators and contributes empirical insights into both the practical benefits and limitations of GenAI integration providing actionable guidelines for narrative generation in real educational settings.