Case study
Reflections GPTs
Interactive reflection systems that replace static journals with guided Socratic dialogue, challenge, and quick feedback, and produce measurably deeper student engagement.
Purpose
Interactive reflection systems that replace static journals with guided Socratic dialogue, challenge, and quick feedback, and produce measurably deeper student engagement.
Overview
Reflections GPTs replaces static remote-learning journal entries with interactive reflection systems.
Instead of writing one response and submitting it, students work through a guided dialogue that questions their assumptions, asks for clarification, and pushes for deeper analysis. The system never hands over answers. Its job is to make students think harder, reason better, and stay with the material.
What happened when it ran
The original journal format asked for at least 250 words on a prompt. Most submissions came in under 500, just enough to clear the bar.
Students using the Socratic system routinely wrote thousands of words across a back-and-forth. After several real rounds of it, one student sent a message that caught the dynamic exactly, asking how they were supposed to finish.
They were not confused. They were so far into the conversation that there was no natural stopping point, which is the opposite of the problem the original format was trying to solve.
That prompted a small design change, making it clearer that students should use their own judgment about when they had learned what they needed. The underlying point held. A well-designed challenge takes the ceiling off engagement.
What I built
It is a set of Socratic reflection flows tied to specific lesson objectives across roughly 20 lessons, with guardrails that block one-word or barely-engaged responses, and a submission step where students paste their full chat transcript as the journal entry for completion credit.
Why the transcript submission model matters
Transcript submission gives instructors something no traditional journal can, which is visibility into the quality of a student's thinking as it happens rather than the one polished sentence they chose to turn in.
They can see where a student engaged for real, where they pushed back, where they gave a surface answer that the system challenged, and whether they rose to it. It also keeps a clear record of how students are using AI, which matters anywhere AI policy is still taking shape.
Current state
This is active and still developing as a reflection-assignment framework for remote and hybrid courses.
Related