TOWARD AI‑ENABLED INSTITUTIONAL REPORTING: A CONCEPTUAL PROPOSAL WITH A CROATIAN HEI CASE STUDY
Abstract:The aim of this paper is to assess the possibilities and limitations of applying AI—huge language models (LLMs) and retrieval-augmented generation (RAG)— to support the preparation of institutional quality and performance reports. The study explores the key challenges of current reporting practices and assesses the potential of AI to enhance the quality, efficiency, and usefulness of reporting in higher education. The methodology is based on a scoping literature review covering quality assurance in higher education, stakeholder information needs, institutional quality and performance reporting, and the use of LLMs and related AI technologies in reporting processes. A case study of a Croatian higher education institution was used to analyse stakeholder information requirements, types of existing reports, data sources and databases, reporting frequency, and the main limitations of current reporting practices. The findings show that stakeholders’ needs differ significantly, requiring a range of report formats (e.g., KPIs, plans, self-assessment reports, survey results). Additionally, data sources are fragmented and dispersed across multiple systems, making data collection and analysis difficult and increasing the subjectivity of result interpretation. Based on these findings, the paper proposes a conceptual model that links specific business and reporting challenges with potential AI-based solutions. The results suggest that AI can streamline and accelerate report preparation, while tailoring outputs to diverse stakeholder groups. However, due to institutional diversity, a universal “one-size-fits-all” solution is unlikely. Instead, institutions should conduct pilot projects using real documentation enriched with metadata and AI models adapted to the Croatian language and institutional context. Following implementation, it will be necessary to critically evaluate the accuracy of AI-generated outputs, assess the ability to link conclusions with supporting evidence, and identify any unintended consequences.
Keywords: Higher education, Quality assurance, Institutional reporting, Large Language, Models, RAG.
