2026 ASEE Annual Conference & Exposition

Artificial Intelligence (AI) technologies are rapidly transforming higher education, yet their integration into core engineering courses remains limited. This study presents the development, implementation, and evaluation of a Virtual Teaching Assistant (VTA), an AI-driven pedagogical companion, deployed in undergraduate courses across Aerospace, Civil, and Mechanical Engineering. These courses traditionally emphasize mathematical modeling and engineering reasoning but, like many large engineering classes, face enduring challenges such as limited instructor availability, diverse student preparation, and reduced opportunities for individualized feedback.

The VTAs were designed to complement the instructor by offering immediate, context-aware explanations, interactive problem-solving guidance, and on-demand clarifications of course concepts, with seamless access for students anytime and anywhere. Each VTA was trained using course-specific instructional materials, example problems, and feedback strategies aligned with the instructor’s pedagogical approach. Students engaged with the VTA through a chat-based interface integrated within the learning management system, enabling continuous, personalized support throughout the course.

Student surveys were administered at the end of the semester to evaluate the perceived usefulness, accessibility, and learning impact of the VTA in two of the three course offerings. The results indicated that most students found the VTA helpful for understanding core concepts, particularly in areas requiring iterative reasoning such as design-oriented problems. Students emphasized the immediacy of responses, the freedom to explore “what-if” scenarios without judgment, and the availability of personalized guidance beyond regular office hours. Some concerns were raised regarding potential overreliance and occasional numerical inaccuracies, prompting targeted design improvements in subsequent offerings. The third course is currently in development, and its implementation will provide further insights once deployed.

The paper presents the VTA’s architecture, integration strategy, and methods used to align its responses with course learning outcomes. The observed pedagogical benefits include increased student engagement, reduced hesitation among reserved learners, and deeper conceptual reinforcement through self-paced inquiry. From the faculty perspective, the tool effectively offloaded routine clarification requests, allowing instructors to focus on higher-order mentoring and assessment.

Beyond the specific case study, this work discusses the broader implications of AI-assisted instruction in engineering education, emphasizing scalability, ethical considerations, and human–AI collaboration. The results demonstrate that Virtual Teaching Assistants can meaningfully enhance student learning when thoughtfully embedded within a well-structured pedagogical framework. Overall, the findings provide a replicable model for AI integration across engineering disciplines and contribute to the ongoing transformation of education into a data-informed, student-centered ecosystem.

The final paper will elaborate on the implementation across the three course offerings, detailing contextual differences in deployment and usage. It will further analyze common trends in student feedback, highlighting recurring themes of engagement and support, as well as notable dissimilarities reflecting the distinct disciplinary and instructional contexts of Aerospace, Civil, and Mechanical Engineering.