Case scenarios and role-plays created with AI

Knowing the right answer on paper is not the same as doing the right thing when a real person is in front of you. A new manager can recite the steps for a difficult conversation and still freeze when a team member gets upset. A property manager can know the rules on repairs and still handle an angry tenant badly. Courses that stop at explanation leave that gap wide open.

Case scenarios and role-plays close it by letting learners practice decisions in a safe setting. Writing realistic ones takes time, and it is hard to invent varied situations after the first few. AI is a strong drafting partner for this work, and it can even play the other person in a practice conversation. You still need to check every scenario for accuracy, realism and fairness.

What makes a scenario worth practicing

A scenario is a short, realistic situation that asks the learner to decide what to do. The good ones share a few features:

• Concrete detail. Names, stakes and a specific moment, not “a customer is unhappy.”
• A real decision. Several reasonable options, with trade-offs between them.
• Consequences. Each choice leads somewhere, so learners see the effect of their decision.
• A link to your objectives. The decision exercises a skill the course teaches.

For the storytelling side of teaching, using stories to make lessons stick covers the non-AI basics. Here the focus is on producing scenarios quickly and making sure they hold up.

Drafting branching scenarios

A branching scenario lets learners choose at each step and follow the consequences. Nadia runs a membership for new residential property managers. She uses scenarios to practice tenant situations, such as a late-night leak complaint from a tenant who has already complained twice.

Write a branching practice scenario for learners who are [describe learners and experience]. The skill being practiced is [paste learning objective]. The situation is [describe the setting and the problem in a sentence or two].

Open with a short, vivid description of the moment. Give the learner three options at each decision point, all of them plausible, with one best choice. Show a realistic consequence for each option, then a second decision point. Keep it to two decision points. End with a debrief that explains what the best path did well and links it to the skill. Use neutral, varied names and avoid stereotypes. Mark anything that depends on local laws or rules so I can check it.

The last line matters in any field with rules that vary by place, such as housing, employment or health and safety. AI will often state a rule with confidence that applies in one place and not another. The marked sections tell you exactly what to verify, with a qualified professional where the rules are complex.

Role-plays with AI playing the other side

Scenarios on paper are useful. Live practice is even better. General assistants such as ChatGPT, Claude and Gemini can play a character in a conversation, reacting to what the learner says. Their voice modes let learners practice out loud, which is closer to the real thing. You can give members a ready-made prompt to start the role-play in their own AI tool.

Ruben runs a coaching program for first-time team leaders. He gives members this prompt to rehearse a conversation about missed deadlines:

Let's do a practice role-play. You will play [character description, for example a capable team member who has missed three deadlines and feels overloaded]. I will play [my role]. Stay in character and react realistically: if I am vague, be confused; if I am blaming, get defensive; if I listen well, open up gradually. Keep your replies short, like real speech. Do not coach me during the role-play. When I type “debrief,” step out of character and give me feedback on [the skills taught, such as asking open questions and agreeing next steps], quoting what I said.

Tell members to keep role-plays fictional and not to type real colleagues' names or confidential details. AI feedback on a role-play is a practice aid, not an assessment. For anything that counts toward completion, you or a trained facilitator should judge the performance.

Check realism and fairness

Every scenario needs a review before it goes into your course. Look for:

• Accuracy. Would an experienced practitioner handle the situation the way the debrief says?
• Realism. Does the dialogue sound like real people, or like a training manual?
• Stereotypes. Check who is cast as difficult, who is cast as the expert, and whether names and backgrounds lean one way. Bias in AI explains why this happens.
• Obvious answers. If one option is clearly silly, replace it with a tempting mistake.

If you can, ask one or two experienced members to try a scenario and tell you where it rings false. Their comments are the best way to make it feel like the real job.

Putting scenarios to work

Once you have a set of reviewed scenarios, use them in several places. Put a short one at the end of a lesson as a check on understanding. Use a longer branching case as the focus of a live session, pausing at each decision for a group vote. Turn the decision points into scenario-based quiz questions, using the approach in creating quizzes and question banks with AI. And share the role-play prompts as optional practice for members who want more.

Next steps

1. Choose one skill in your course that learners find hard to apply.
2. List three realistic situations where they would need it.
3. Draft one branching scenario with the prompt and check every marked rule.
4. Review for accuracy, realism and stereotypes, and ask an experienced member to try it.
5. Write a role-play prompt for the same skill and test it yourself first.
6. Add the scenario and prompt to the lesson, with clear guidance on keeping practice fictional.