2026 ASEE Annual Conference & Exposition

Expanding Implementation of an AI Conversational Agent to Support Troubleshooting in an Engineering Lab Environment

Presented at Computers in Education (CoED): AI in Education (2 of 9) -- M408A

This work-in-progress paper describes the expanded implementation of an ongoing research and development project implementing an Artificial Intelligence (AI) agent in a mechanical engineering experimentation course. These interactive agents are becoming increasingly integrated into educational environments to enhance learning outcomes and aid instructors in assisting students. This study examines the implementation of a custom AI conversational agent developed to support undergraduate engineering students learning through experimental coursework while alleviating teaching assistant workload. The conversational agent provides course-specific explanations, troubleshooting assistance, and methodological guidance.

The conversational agent is strategically embedded within structured assignments to promote divergent thinking, characterized by exploration of multiple solution pathways and deep conceptual understanding. Chat log analysis and student feedback are used to assess the conversational agent's effectiveness in fostering divergent thinking and to identify areas for refinement in both the human-computer interaction and assignment design. Student adoption rates have indicated aspects of successful implementation during the 2024-2025 academic year

During Fall 2025, we are scaling implementation to additional lab sections while introducing two design revisions: (1) transitioning from broad, course-wide access to case-specific use cases, and (2) shifting from open-ended questions to Socratic dialogue. Socratic dialogue is integrated to progressively reveal case study clues and use rhetorical prompting to guide the student to connect system symptoms to error roots. This redesign allowed students to learn about problems in real time as they interact with the conversational agent.

The paper will contribute a transferable design pattern for course-grounded LLM assistants, i.e., retrieval-first with Socratic scaffolds, guardrails, and user analytics, early usage and learning indicators from a multi-section rollout, and an evaluation blueprint that other instructors can adopt when deploying AI support in laboratory settings. This ongoing work contributes to understanding how structured AI integration in engineering education can support higher-order thinking skills while maintaining pedagogical objectives.

Authors

Dr. D. Matthew Boyer http://orcid.org/0000-0002-4191-260X Clemson University [biography]

Dr. Boyer is a Research Associate Professor in the Department of Engineering and Science Education and an Educational Proposal Writer in the College of Engineering, Computing and Applied Sciences.
Mitra Varun Anand http://orcid.org/0009-0000-3081-6367 Worcester Polytechnic Institute [biography]

Mitra Anand serves as the Associate Director of Makerspace, and Innovation and Entrepreneurship, in addition to being an Adjunct Faculty of Mechanical Engineering at Worcester Polytechnic Institute.

Anand's research interests lie in combining hands-on Maker skills with an entrepreneurial mindset and value creation, aiming to develop practical solutions for real-world problems. He is enthusiastic about innovation in engineering education, design thinking, prototyping, program development, crafting interactive curricula, and bringing ideas to fruition.

With over 8 years of experience in Mechanical and Mechatronics Engineering, Anand possesses a solid background in Innovation and Entrepreneurship education, Additive Manufacturing, and Digital Fabrication technologies. He has taught lectures and workshops on advanced subjects to more than 1000 undergraduate engineering students and 150 graduate students, while advising on over 500 student and faculty research projects.

His academic credentials include an M.S. in Mechatronics and Robotics Engineering from New York University and a B.E in Mechanical Engineering from Anna University.
Sahil Mirani Worcester Polytechnic Institute
Prof. Ahmet Can Sabuncu http://orcid.org/https://0000-0001-7905-421X Worcester Polytechnic Institute [biography]

I hold a Ph.D. in Aerospace Engineering and am currently a faculty member in the Department of Mechanical Engineering at Worcester Polytechnic Institute. My research focuses on engineering education, particularly on problem solving in ill-structured domains, problem typologies, and the role of complexity in learning. I investigate how digital and AI-supported learning environments can be designed to foster adaptive expertise in engineering practice. My work also explores innovation in engineering product design and the integration of entrepreneurial mindset into engineering education. Through both research and teaching, I aim to bridge foundational engineering knowledge with real-world problem-solving practices.
Mr. Tim Ransom http://orcid.org/0000-0003-0357-5427 Clemson University [biography]

Computer Science Lecturer, PhD candidate with the Engineering and Science Education Department at Clemson University.
Jonah Cohen Worcester Polytechnic Institute [biography]

Jonah Cohen is a third-year engineering student at Worcester Polytechnic Institute pursuing a BS/MS in Mechanical Engineering.

Note

The full paper will be available to logged in and registered conference attendees once the conference starts on June 21, 2026, and to all visitors after the conference ends on June 24, 2026

« View session