Standalone session 9 min Reviewed 2026-07-12

From Static Cases to AI Role-Play

Case discussions usually look backward: the decision is already made and the pressure is gone. This session shows how a bounded chatbot role-play puts students inside the situation while the decision is still unfolding, so they practice judgment under realistic pressure instead of recalling someone else's answer.

Watch

Session video

9:06
Video summary

Case discussions look backward: the decision is made and the pressure is gone. This session flips the timing by letting students step into a situation while the decision is still unfolding, with a chatbot as scenario partner — a nervous investor, skeptical client, or confused patient — never an answer machine. The loop is simple: students enter the scenario, respond, the AI responds in character, the exchange repeats with rising pressure, then reflection and your debrief. The key design principle: assess the reflection, not the chatbot conversation — the student's revised judgment is the evidence. A five-step build (define the objective, build the scaffold, write the prompt, test and iterate, run and debrief) and a five-block prompt anatomy (role, situation, boundaries, behavior, debrief) keep the bot from solving the task or grading without you. A green/yellow/red boundary keeps faculty judgment in the loop: simulate and support reflection, verify any formative feedback, never private student data or final grades. It closes with a modest pilot — one role, one decision, one debrief question, on a 5/10/5/10-minute clock — and three assessment questions: What did you decide? What evidence did you use? What did the role-play make you reconsider?

The one thing to keep

Key takeaway

Leave with one role, one decision, and one debrief question you could pilot in a course you already teach. The chatbot is the practice environment, not the teacher or the assessment. The student's revised judgment is the evidence.

Make it real

Try it next week

  1. Start with one existing case, scenario, or hard concept you already teach. You already know where students struggle.
  2. Name one decision students should practice. Not five. One.
  3. Build one AI role that creates realistic pressure: a skeptical client, a confused patient, an angry customer, a worried parent, a regulator.
  4. Keep the timing modest: five minutes setup, ten minutes role-play, five minutes reflection, ten minutes debrief.
  5. Assess three questions: What did you decide? What evidence did you use? What did the role-play make you reconsider?

Take it with you

Session resources

LINK

Role-Play in Finance Example

An example of an AI role-play application in the finance classroom.

Go deeper

Connect this session to the course

This session pairs naturally with Module 4: Assignment Redesign. The module adds the frameworks, reading, and portfolio artifact behind this design move.

Keep the conversation going

One useful update. Once a week.

Follow new resources, course improvements, and faculty-support tools without chasing every headline.