Agentic Orchestration for Pedagogy: Knowledge-aligned Multi-agent Learning Assistant

Topic 1
Enhancing Teaching and Learning with RAG-powered Knowledge Base Chatbot (by Dr. Albert LEE)
Topic 2
How GenAI-based Chatbot Promote Students’ Knowledge Assimilation and Design Thinking in a Project-based Course (by Dr. Vincent Tam)
Topic 3
Multi-agent Educational Chatbot Framework for Teaching and Learning Excellence (by Dr. Zhenglong LI)
Topic 4
Methodologies to Enhance Alignment of Chatbot Repsonses to Instructors’ Tacit Knowledge and Students’ Learning Status for Adaptive Teaching and Learning (by Mr. Alex Kiang)

Our team presents a multi-agent educational AI chatbot that supports teaching and learning at scale across courses in Electrical and Electronic Engineering, Philosophy, Biomedical Sciences, and other areas. We combine a proven knowledge-base chatbot, deployed over three consecutive semesters, with a newly developed multi-agent framework and a knowledge-graph approach. The knowledge-base chatbot answers student questions using approved materials and is designed to keep responses aligned with instructors’ teaching insights and experience. To broaden support, the multi-agent framework routes requests to focused helpers, for example, an agent for course administration, an agent for explaining core concepts and linking lecture and lab ideas to projects, and an agent for generating practice problems based on each student’s current learning status. Student learning status is evaluated using a knowledge graph that links learning goals, concepts, and common misunderstandings to form a learning map that supports personalised learning. The system is modular, so agents and data sources can be added or adjusted over time, and results can be reported as they become available. Early use of the knowledge-base chatbot across three semesters shows strong student feedback, fewer routine queries, and clearer guidance.