AI solutions to classroom challenges: transforming learning at PolyU
What if every student could access a personal tutor at any time of day, in their preferred language, and tailored specifically to their course content? At PolyU, Professor Haitian Lu from the School of Accounting and Finance has been developing an innovative solution through course-specific AI-Tutor Agents designed to address longstanding educational challenges and enhance learning experiences.
After two decades of teaching, Professor Lu identified key pain points in conventional education: failure to accommodate individual learning speeds in large classes, excessive administrative tasks for teachers, and delayed feedback affecting student motivation.
To tackle these local challenges, Professor Lu’s team gathered feedback from students across various disciplines. The survey revealed common issues, including incomplete comprehension of core concepts, limited questioning opportunities during class, and language barriers reported by many Chinese Mainland students.
The AI-Tutor Agent directly addresses these issues with three core features:
Course-Specific Learning: Trained exclusively on course materials, lecture slides, past exam papers, and reading lists to ensure relevant, aligned support.
Traceable Answers: Provides precise answers to questions while citing specific lecture slides and page numbers, enabling independent verification and building trust.
Analytics for Educators: Identifies common weak points and learning patterns, empowering teachers to adjust their strategies effectively.
Pilot testing demonstrated immediate success, with over 80% of participating students reporting satisfaction with the tool’s accuracy in clarifying complex concepts. The 24/7 bilingual support proved exceptionally beneficial for students studying in a multilingual learning environment. Beyond providing immediate academic support, the AI-Tutor Agent also encourages students to take greater ownership of their studies. By offering timely explanations, transparent references and opportunities for self-paced exploration, it helps learners build confidence while fostering deeper engagement with course content and independent learning habits that extend beyond the classroom. Recognised as Hong Kong’s first university AI tutor, the initiative earned funding from the Teaching Development Grant for a two-year cross-faculty pilot across ten subjects.
Employing a “human-in-the-loop” approach, educators serve as co-creators who review and validate AI outputs. Some professors noted high confidence after evaluating 200 responses, stating the AI tutor performed at a B+ grade level. "True innovation lies in creating subject-specific knowledge databases scalable across university-wide implementation,” notes Professor Lu. “This approach establishes a paradigm where AI amplifies human teaching rather than replacing it.”
As universities worldwide explore the role of artificial intelligence in education, PolyU’s AI-Tutor Agent demonstrates how the University integrates smart technologies into a student-centred approach to prepare learners for the future.

Through the development of course-specific AI-Tutor Agents, Prof. Haitian Lu is transforming how students learn, addressing long-standing educational challenges while providing more personalised support.






