A Western Kentucky University undergraduate student presented research at a national engineering education conference that shows how AI-powered conversation tools can boost engagement and personalize support in large technical classrooms. Catherine Grace "Gracie" Fenno, an Architectural Science major, co-authored the study with fellow student Gabriel Jerdon and Dr. Osama E. Mansour, an associate professor in the School of Engineering & Applied Sciences. They shared their findings at the 2026 American Society for Engineering Education (ASEE) Annual Conference in Charlotte, North Carolina, from June 21-24.
Inside the Building Systems course trial
During the Spring 2025 semester, Dr. Mansour integrated Blackboard's AI conversation tool into a Building Systems course enrolling 41 students across Construction Management, Architectural Science, and Civil Engineering. The AI was configured with personas tailored to technical subjects, allowing students to ask questions and receive immediate, customized responses outside of class. The goal was to see whether conversational AI could address knowledge gaps in a multidisciplinary cohort without overwhelming the instructor.
What students reported
A post-course survey revealed that students found clear value in the AI assistant. The team reported three main outcomes:
- Immediate feedback: 88% of students appreciated the 24/7 availability to clarify complex building science concepts on demand.
- Enhanced engagement: 82% said the conversational AI tools noticeably increased their active participation in the course.
- Personalized support: 81% agreed the adaptive dialogue effectively targeted their individual knowledge gaps.
Instructor oversight proved essential
The researchers also documented technical challenges, including AI hallucinations and biased prompts. Dr. Mansour used a human-in-the-loop approach to keep the instructor as the primary authority. He reviewed transcripts of AI-student interactions weekly and addressed inconsistencies during live lectures. This method turned the AI into a supplement, not a replacement, for expert instruction.
These findings add to the ongoing conversation about AI for Education, especially models where instructors remain in control of the learning loop.
"Students arrived at discussions with more refined technical questions, suggesting that real-time interaction with the AI persona helped clarify foundational concepts as they were being introduced," Dr. Mansour said. "This allowed in-class dialogue to advance more quickly toward higher-level considerations of structural loading and system integration."
Why this matters for education professionals
The trial demonstrates that even in technically demanding courses, AI tools can handle routine clarification work, freeing instructors to lead higher-order discussions. For educators managing large or diverse classes, a human-in-the-loop AI system offers a way to provide personalized support without adding to the grading burden. The critical factor is maintaining instructor oversight to catch errors and ensure the AI stays aligned with curriculum goals.
Your membership also unlocks:
