We invest a lot in high-impact practices. Internships, undergraduate research, service learning, study abroad, first-year seminars: the evidence for these experiences is strong, and institutions have responded by building them into curricula at scale. But there's a quieter problem that often goes unaddressed. These experiences work best when students reflect on them deeply, and most of the time, that reflection isn't happening in any meaningful way.
If you've read end-of-semester reflection essays, you know what this looks like. A student describes what happened. They note that the experience was "interesting" or that it "helped them grow." But the thinking behind those claims stays invisible. The experience happened. The meaning didn't land.
This isn't a student motivation problem. It's a scaffolding problem.
The Prompt Isn't Enough
Most students haven't been taught to reflect at the level faculty are hoping for, and traditional prompts ("What did you learn?") don't get them there. Asking a broad question at the end of a semester invites broad answers. Students write to complete a task, not to think.
What changes the outcome isn't a better prompt in isolation. It's a guided conversation that pushes thinking further, asks follow-up questions, connects a student's response to their actual work, and keeps going until something real surfaces.
At Bucknell University, Rebecca Thomas piloted exactly that approach in an engineering course. The difference in student writing was striking. Without guided reflection, one student wrote that the class was "pretty straightforward" and that they never felt "at a loss." With AI-guided reflection, a different student from the same course wrote: "I would ask for help sooner when we were confused on the code. We were so intent on finding the error ourselves that we spent an hour combing through it. It's helped my perspective on mistakes. Mistakes do not equal failure."
Same course. Same students. Same semester. The only difference was how they were asked to reflect.
A Framework That Goes Deeper Than Technology
The approach we presented at AAC&U CLASS 2026 is grounded in TORI, the Taxonomy of Reflective Inquiry. TORI organizes reflection across six core domains: cognitive and analytical thinking, emotional and affective processing, social and interpersonal dynamics, personal growth, cultural and ethical considerations, and life transitions.
This gives faculty a shared language for what meaningful reflection actually looks like. It also gives AI tools a pedagogical foundation, so they're scaffolding genuine thinking rather than generating text that sounds like reflection but isn't.
Built on TORI, the AI reflection tool functions as a conversational partner rather than a form to fill out. Instructors set the learning outcomes and the framing; students experience a dialogue where follow-up questions are tailored to their responses and tied to their actual work. In our informal polling, students said the 10-minute reflection process didn't feel long. Many noted an unexpected benefit: the guided process helped them articulate their experiences in ways they could use on a resume or in a job interview.
One student put it simply: "It felt like a real conversation, not a box to check."
What This Unlocks for Your Institution
Beyond the learning experience itself, AI-guided reflection opens up something administrators have been trying to solve for a long time: evidence.
Every guided reflection is an artifact of learning, authentic, timestamped, and archived. When accreditation season arrives, you don't need a separate evidence-collection process because the evidence has been building all along. And when you zoom out across a semester or a program, a collection of reflections becomes a dataset. AI-assisted analysis can surface recurring themes, common misconceptions, and growth trajectories that no single instructor could identify alone.
The practice that deepens student learning is the same practice that generates program-level insight, without adding reporting burden to faculty or coordinators.
It's also worth saying clearly: this does not replace human mentorship. Faculty conversations, peer discussion, and advisor relationships remain essential. The AI extends reach and access. The people provide the care.
Where to Start
You don't need to overhaul a program to begin. Start with one course, one HIP context. Try asking one open question instead of four. Brief, frequent reflections build the habit over time and produce a richer longitudinal record than a single end-of-semester essay.
If you're curious about what this could look like at your institution, we'd welcome the conversation. You can also find the slides and resources from our AAC&U CLASS 2026 session.
Tagged