Face Up Book a demo

Practical guide · 2026

From role-play to AI simulation

The trainer's guide: convert your role-plays, write your personas, debrief, assess. Methods, templates and pitfalls, with no sales pitch.

  • Convert an existing role-play, column by column
  • Write a persona in four layers, with its hidden information
  • Write the debrief, build the quality score, launch in 30 days
Cover of the Face Up guide: From role-play to AI simulation

22 pages. For trainers and instructional designers who already run role-plays.

Do the maths

The limit of classroom role-play is not about teaching. It is arithmetic.

1 to 2role-plays per participant per year
10-15 minof speaking time in a real situation, per person
0chances to replay the same situation

Typical calculation for a one-day session with twelve participants, detailed in the guide.

Inside

What you will find in this guide

01

What the classroom does best

And what it cannot do: the constraint is logistical, not educational.

02

The new division of roles

What stays in the classroom, what moves to self-paced practice, what disappears.

03

The conversion grid

Turning an existing role-play around, column by column, with a worked example.

04

The persona and hidden information

The four layers, the template to fill in, the DISC profile, and the mechanism that makes it real training.

05

The 10 pitfalls and the debrief

What goes wrong in a first scenario, and what you write, or never write, in a debrief.

06

Quality score, framework and launch

A grid per job, Qualiopi, GDPR and the AI Act, then a thirty-day checklist.

Excerpt · The conversion grid

You already have the material, just arranged differently

1

The scene

The context of the situation.

  • Keep only what is physical and constraining
  • The place, the moment, what can interrupt
2

The persona

The brief handed to the peer playing the role.

  • Behaviour under pressure, register of language
  • And above all, what the character hides
3

The scenario

The history of the relationship.

  • What happened before, what is at stake
  • The levers that move the conversation forward
4

The debrief

Your observation grid.

  • It becomes the structure of the feedback
  • The step where most conversions fail
Too little hidden

The character explains everything in the first thirty seconds. The learner succeeds without learning anything, and gets bored.

Too much hidden

Nothing opens up whatever you do. The learner concludes that "the AI does not understand", and is right to think so.

The Face Up AI Studio: a conversation with SAM and the scenario draft
Face Up sessionA conversation with SAM in the AI Studio, and the scenario draft builds itself.

The AI Studio

SAM: describe the situation, it writes the scenario

  • 1
    Describe a real situationIn everyday language, as you would to a colleague. A PDF brief attached is enough.
  • 2
    You remain the authorSAM produces a draft structured in four layers. Every field stays editable.
  • 3
    Test before publishingPlay the scenario once well and once badly before giving it to a group.
  • 4
    Built into your LMSThe published scenario launches from your existing programme (SCORM, LTI, SSO).
“AI role-plays and the quality of the feedback save training centres an enormous amount of time.”
Thierry Bonetto, former HR Director for Training, Danone
“AI simulations let learners practise far more, without adding to the trainers' workload.”
Pierre Bacrot, training centre director

Bring a case. We will convert it with you.

The guide gives you the method, column by column.

Get the guide