2026 ASEE Annual Conference & Exposition

Shear-moment diagrams remain a persistent challenge in engineering courses, from lower-division statics and mechanics of materials to upper-level structural analysis. Student success depends on correctly linking free-body diagrams to equilibrium, maintaining consistent sign conventions, and translating applied loads into shear and moment behavior along a structure. This Work-in-Progress paper describes the design and preliminary evaluation of an AI-guided study workspace that supports learning through a graphical, process-focused approach. The workspace emphasizes a qualitative load–shear-moment scaffold that helps students connect external actions to key diagram features, such as identifying shear jumps at point loads and reactions, assigning slopes under distributed loads, drawing moment behavior from shear, and locating moment extrema where shear is zero. Each problem follows a small set of explicit steps: drawing the free-body diagram, computing reactions, constructing the shear and moment diagrams, and performing consistency checks on signs, continuity, end values, extrema, and units. The AI tutor provides short, targeted prompts when common conceptual errors are detected, encouraging students to reflect on their reasoning rather than supplying final answers.

Implementation occurred in two undergraduate course sections with the same content and pacing. Students completed a traditional problem-solving round without AI support, followed by a second round using an AI-guided Study Mode workspace during the same class session. The AI-guided activity used the same graphical methods emphasized in class and was accessed through a free account configured to support step-by-step reasoning without providing final answers. During the activity, the instructor monitored progress, addressed conceptual questions, and encouraged reflection when errors were identified. Paired, isomorphic problem statements were used across rounds to ensure comparable difficulty and isolate the effect of the practice mode. Students submitted their written work and AI interaction excerpts through the learning-management system, and credit was awarded for completion of required steps rather than solution accuracy.

Evaluation included a short concept check administered before and after the AI-guided workspace activity, related questions on a later course exam to examine carryover of learning, basic process data such as time-on-task and coded error types (e.g., shear jumps, slopes, moment extrema, and boundary checks), and a brief survey of student confidence and perceived mental effort. Traditional and AI-guided practice were compared on these measures, with attention to pre-to-post gains, changes in error patterns, and performance on the later exam. This Work-in-Progress paper describes the study guides and example prompts used in the pilot, which leverage ChatGPT’s built-in Study and Learn modes, to illustrate how the activity may be adopted or adapted in similar courses.