Design and implementation of a generative AI-supported entrepreneurship education system based on project-based learning

This study designs and evaluates ABL-Startup, a generative AI-supported entrepreneurship education system grounded in Project-Based Learning (PBL). The study mapped five PBL stages onto entrepreneurship tasks: idea formulation, market and persona analysis, AI-guided business modeling, feasibility and revenue analysis, and Demo Day, and implemented them in a four-layer architecture integrating a prompt engine, PSST-based document generation, presentation generation, and mock investment evaluation. Functional alignment, task-efficiency data, and exploratory pilot satisfaction results for the system were analyzed. ABL-Startup connected stage-based learning activities with cumulative outputs and iterative feedback. In the pilot context, business plan drafting decreased from approximately four weeks and 12 working hours to one week and three hours, while overall satisfaction reached 4.88/5; writing efficiency, process appropriateness, and AI usefulness scored 4.95, 4.90, and 4.89, respectively. Generative AI can operate as process-embedded instructional scaffolding rather than merely as a writing tool, although controlled studies are required to establish learning effects. The system can lower novice learners’ documentation burden, support consistent instructor feedback, and be adapted for university courses, startup camps, and public entrepreneurship programs.