A professional role-play for your language lesson, prepared with AI
A role-play is only useful if it looks like your learner's job: the person they talk to, what is at stake, their level. With AI you can prepare the scenario brief, the role cards and the debrief grid quickly, and made to measure. Here is the method, in six steps.
In brief
A professional role-play that works starts from a situation the learner really faces at work, gives each player a role, a goal and a constraint, and is pitched at the target CEFR level. AI prepares three documents: the scenario brief, one card per role and a debrief grid. The trainer rereads everything, runs the role-play without interrupting, then gives feedback in two stages: what worked in the exchange, then two or three language points to work on.
1. Start from a real work situation
The starting point is the needs analysis: who does the learner speak to, through which channel, to obtain what? The situations LangueXpert certified trainers have worked on with their learners give the idea: for a corporate lawyer, negotiating contracts in English with Americans, Brits and Indians; for a sales director, a whole negotiation, from greeting the supplier to signing; for a director of sport, facing journalists at a press conference. Their accounts are on the certification page. Choose a situation that is precise and close in time: Thursday's meeting is better than “meetings” in general.
2. Set the roles and the stakes
A role-play with nothing at stake turns into recitation. Each role gets three things: who they are, what they want to obtain, and a constraint the other side does not know about (a capped budget, an impossible deadline, a piece of information not to give away too early). That information gap is what forces the players to listen, rephrase and negotiate. The learner plays their own role, the one they hold at work; the trainer, or another participant, plays the other person.
3. Pitch it at the right CEFR level
The same situation is not played the same way at A2 and at B2. At A2: short, predictable exchanges, with a few phrases supplied in advance. At B1: the learner has to explain, give an opinion and handle something unexpected. At B2: they argue a case, qualify what they say and rephrase what the other person says. Tell the AI the current level and the target level: the scenario, the length of the cards and the phrases suggested all depend on it.
4. The AI's part: scenario brief, role cards, debrief grid
Ask for three documents:
- The scenario brief: the context in a few lines, shared by both roles.
- One card per role: goal, constraint, information to use, and five or six useful phrases at the target level; each player reads only their own.
- The debrief grid: three or four observable criteria, to tick during the role-play.
Then reread everything: an approximate technical term, an unrealistic constraint or a card too long for the level is fixed before the lesson. Describe the job and the situation, without naming any person or company.
5. Run the role-play
State the objective, allow two or three minutes of preparation with the card, then start: five to ten minutes of play is enough. During that time, do not interrupt: note on the grid what works, where the learner gets stuck and the errors that keep coming back. Online, send each card as a private message, so that the information gap holds. If the learner agrees, record it: you can listen to a passage again together.
6. Debrief, then replay
Feedback comes in two stages. First the exchange itself: was the goal reached, what worked? Then the language: two or three points at most, chosen from the grid, each with the wording you would expect. Then have the scene replayed with one parameter changed (a more hurried counterpart, a new constraint): the second run lets the learner reuse the corrections straight away.
Example
A purchasing manager, from B1 towards B2: she meets a supplier who announces a price increase. Her card: keep the current price or obtain something in return, without damaging the relationship; her constraint: she cannot change supplier for six months. The supplier's card, played by the trainer: announce the increase; he can give way on payment terms, not on the price. Debrief grid: rephrases the other side's position, proposes something in return, closes with a clear agreement. Language to keep: disagreeing politely, setting a condition, summing up.
A starter prompt
You are an English trainer for adults. My learner works as [job], current level [B1], target [B2]. Prepare a 10-minute role-play on this work situation: [situation]. Write: 1) the scenario brief shared by both roles, in 5 lines; 2) a card for each role, with its goal, one constraint the other side does not know about and 6 useful phrases at the target level; 3) a debrief grid with 4 observable criteria. Do not use any real name of a person or a company.
The full version — the needs analysis, the lesson plan, the personalised resources and assessment during training — is built in the preparatory course of the certification, on one of your own learners. See also: the speaking assessment rubric, the placement test and the listening lesson with AI-generated audio.
Frequently asked questions
How long should a role-play last in a language lesson?
Five to ten minutes of play is enough, after two or three minutes of preparation and followed by a debrief at least as long as the role-play itself. A short scene played twice works better than one long simulation played once.
Can I run a role-play in a one-to-one lesson?
Yes: the learner plays their own role and the trainer plays the other person. The debrief grid is then filled in straight after the role-play, or from the recording if the learner has agreed to one.
From what level can learners do a role-play?
From A2, provided the situation is short and predictable and a few phrases are supplied in advance. From B1, add something unexpected; from B2, a disagreement to handle and a case to argue.
Which AI tool should I use to prepare a role-play?
A general-purpose AI assistant is enough to write the scenario brief, the role cards and the debrief grid. In the LangueXpert preparatory course, the tools worked on are Claude, ElevenLabs and Google AI Studio. Either way, the trainer rereads and corrects everything before the lesson.