2026 ASEE Annual Conference & Exposition

Chemical engineering courses in process control and process design challenge students to integrate mathematical modeling, systems thinking, and open-ended problem solving. However, students often struggle to make the cognitive connections between theory and application, particularly in reasoning through complex, multivariable problems. In this paper, we present the design and implementation of an AI-based Socratic tutor created to promote deeper learning in undergraduate chemical engineering courses through structured questioning and reflection. The AI tutor interacts with students solely through questions, prompting them to articulate reasoning, justify decisions, and confront misconceptions—mirroring the inquiry-based dialogue characteristic of effective one-on-one tutoring.
Grounded in principles from cognitive science that support student acquisition of knowledge and skills using active learning, the Socratic tutor was deployed in two senior-level courses: Process Dynamics and Control and Process Design. The AI agent was trained with a curated primary knowledge base assembled by the course instructors, consisting of publicly available textbooks, technical references, and web resources verified to be accurate and appropriate for the course content. This knowledge base supported alignment between the tutor’s questioning, course objectives, and chemical engineering best practices.. Students engaged with the tutor periodically throughout the semester on topics related to the current course delivery. Following their interactions, students completed surveys reflecting on their experience and evaluating their perceived learning. The results of the perceived learning survey were compared with data from prior course offerings that did not use the Socratic tutor to determine changes in self-reported understanding, confidence, and engagement.
Preliminary findings indicate that students perceived the AI tutor as an effective tool for reinforcing conceptual understanding and promoting self-directed learning. Students reported that the question-based format encouraged them to think more critically before seeking answers and to explain their reasoning more clearly. This paper describes the development of the curated knowledge base and AI Socratic tutor, presents comparative survey results, and discusses implications for integrating AI-driven metacognitive tools into chemical engineering curricula to strengthen conceptual mastery and student independence in learning.