Design of an AI-Based Chatbot for UNIZIK Student Affairs Information Services

Student affairs offices at Nigerian universities including UNIZIK experience high volumes of repetitive inquiries from students concerning registration deadlines, fee payments, hostel allocation, and academic calendars, overwhelming staff and causing service delays. This study designs an AI-based chatbot for the UNIZIK Student Affairs office, capable of answering frequently asked questions, providing procedural guidance, and escalating complex queries to human staff. The chatbot was developed using the Rasa open-source conversational AI framework, with a custom intent classification model trained on 1,800 student query samples collected through structured interviews and historical inquiry logs. The backend was integrated with UNIZIK's student portal database via a RESTful API to enable personalized responses such as fee balance checks and registration status. A web-widget interface was embedded into the UNIZIK student portal. System evaluation involved 100 student testers interacting with the chatbot over a three-week period using real-world inquiry scenarios. Results indicate that the chatbot correctly handled 86 percent of queries within the trained intent scope, with an average response time of 1.4 seconds. Escalation to staff was triggered appropriately in 94 percent of out-of-scope queries. Student satisfaction ratings averaged 4.2 out of 5. The study concludes that AI chatbots substantially reduce the administrative burden on UNIZIK student services staff while improving student experience. Recommendations include Igbo language intent training and chatbot integration with the UNIZIK mobile app.