Generative AI for Data-Informed Educational Decision-Making: A Systematic Review and Framework

Generative artificial intelligence (GenAI) is creating new possibilities for data-informed decision-making in higher education, yet how it translates educational data into actionable decisions for different stakeholders remains undertheorised. This study reports a PRISMA-guided systematic review of 38 peer-reviewed empirical studies addressing three questions: what educational decisions GenAI supports in higher education (RQ1); how educational data and GenAI capabilities enable those decisions (RQ2); and what framework explains GenAI-driven educational decision-making (RQ3). Across the 38 studies, eight decision categories are identified inductively and organised into micro (learner-facing), meso (instructor and program), and macro (institutional) levels. Five enabling GenAI capabilities (natural language generation, behavioural and semantic analysis, multimodal processing, synthetic data generation, and conversational analytics) are mapped to each category. Synthesising these findings, this study proposes the GenAI-Driven Educational Decision-Making (GDEDM) framework: a four-layer model connecting educational data sources, GenAI capabilities, educational decision levels, and governance. The evidence reveals important cross-level patterns, divergences, and research gaps, exposing a critical governance gap between rapidly advancing GenAI capabilities and lagging institutional safeguards. The GDEDM framework offers researchers, practitioners, and policymakers a coherent, evidence-grounded basis for responsible GenAI deployment in higher education.

@article{Razami2026,
  title     = {Generative AI for Data-Informed Educational Decision-Making: A Systematic Review and Framework},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {17},
  number    = {9},
  year      = {2026},
  publisher = {The Science and Information Organization},
  author    = {Husna Hafiza Razami and Abtar Darshan Singh and Roslina Ibrahim and Ali Najeeb and Helio Augusto da Costa Xavier Mauquei},
  doi       = {10.14569/IJACSA.2026.0170948},
  url       = {https://doi.org/10.14569/IJACSA.2026.0170948}
}