US424: Grouply: A Smart AI Collaboration Platform For Fair Academic Assessment

Saiful Fikri Bin Shaharum Universiti Pendidikan Sultan Idris

Group assignment is a key part of academic study intended to encourage students to collaborate and work as a team. Nevertheless, managing such group works proves difficult due to the commonly occurring "free-rider" dilemma and uneven distribution of workload. Existing project management software designed for academic settings fails to include essential features that maintain academic integrity. Therefore, the primary goal of this project is to create a safe, intelligent web-based academic group management system called Grouply, followed by a thorough usability evaluation. In creating the system, the RAD methodology was applied. Grouply employs a decoupled architecture leveraging the React 19 frontend framework, Supabase database management system, and Python FastAPI microservice linked to the Groq API. To ensure strict accountability, Grouply imposes strict 48-hour edit lock rules on all milestone evidence submissions. The system earned a perfect score when it passed 31 detailed unit and module test cases. Additionally, the User Acceptance Testing (UAT) produced an excellent System Usability Scale (SUS) score of 84.17 out of 100. Overall, Grouply offers an ideal way to ensure a high level of usability, effectiveness, and security in eliminating fraudulent behavior and unfair grading.