Exercises grounded in workplace practice

How can you use AI to produce a realistic exercise that is suited to learners and can be assessed unambiguously?

1250 € per session, excl. VAT

In short

In a half-day session of 3 hours participants use generative AI to turn a workplace situation into a usable exercise. They write an exercise brief, prepare the necessary resources, develop an answer key and adjust the difficulty. Peer testing checks consistency and feasibility.

Why this training?

An exercise can seem realistic while assessing something other than the intended competence. Unnecessary information, an implicit instruction or an overly restrictive answer key can turn a scenario into a guessing game rather than an opportunity to practise. AI makes it easier to produce scenarios and variants, but may add facts that were not provided or unintentionally alter the difficulty. The training teaches participants to set boundaries for its suggestions, separate the exercise brief from the answer key, and then check their alignment. Running case study: The fictitious organisation « Montalis Rayonnages » welcomes assistants responsible for responding to quotation requests. Using a fictitious catalogue and three incomplete customer messages, participants design an exercise: identify missing information, prioritise questions and draft a response without inventing product characteristics. Transfer to the workplace: Within the following two weeks, rework an existing exercise using the template produced, with anonymised data. Compare the ambiguities identified before and after revision, then record learners’ questions the next time the exercise is used in order to adjust the brief.

Objectives of this training

  • Define an observable competence and the output expected from the learner.
  • Select the contextual elements needed to complete the exercise.
  • Use AI to write an exercise brief and its resources without contradictory information.
  • Develop an exercise at two difficulty levels without changing the competence being assessed.
  • Evaluate the exercise brief and answer key using an explicit assessment rubric.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Trainers and facilitators wishing to refresh their practical exercises.
  • Learning content designers and internal subject contacts preparing short scenario-based activities.

Prior skills

  • Have previously made a simple request to a generative AI tool.
  • Be able to describe the audience and objective of a training activity.

What you will work on

  1. Defining the learning situation

    Objective: Link a credible situation to a specific competence. Distinction between workplace scene-setting and useful context; audience, assumed prior learning, target duration and observable evidence.

  2. Producing an exercise brief that can be completed

    Objective: Obtain a first usable version with AI. Request structure: objective, permitted data, constraints, response format; prohibition on adding facts that have not been provided.

  3. Adjusting difficulty and the answer key

    Objective: Adapt the effort required and make the assessment explicit. Guidance, number of missing pieces of information, distractors; reasoned answer key, common errors and alternative answers.

  4. Testing without verbal explanation

    Objective: Check that the exercise works using only its supporting materials. Timed peer test; distinction between intended difficulty and obstacles caused by wording; review of unanticipated answers.

What you take away

  • Scoping sheet: competence, audience, context and expected output.
  • Ready-to-distribute exercise brief, resources and a difficulty variant.
  • Reasoned answer key and assessment rubric.
  • Test log and reusable AI prompt.

Method and assessment

  • Short input sessions followed by immediate practice using the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group review: what works, what needs correcting and what remains a human decision.
  • Preparation: Bring a learning objective and a fictitious or anonymised exercise; the running case study remains available. Check access to a generative AI tool authorised by the organisation and remove all confidential data from the materials.

Observable criteria: Each task set corresponds to the stated competence. ; All essential information is included in the resources or explicitly identified as information to be requested. ; The variant changes an identified difficulty factor without changing the objective. ; The answer key justifies the answers and acknowledges acceptable alternatives. ; The test meets the target time and any ambiguities identified are corrected..

Frequently asked questions

Is a particular tool required?

No. A generative AI tool authorised by your organisation is sufficient; the prompts remain transferable.

Can you start from a real situation?

Yes, after effective anonymisation and removal of confidential information; otherwise, use the fictitious case study.

Can AI validate the answer key on its own?

No. The participant checks the answers against the resources and decides which alternatives are educationally acceptable.

Does the training cover a complete learning pathway?

No. It covers a contextualised exercise, its variant and its assessment, not the structure of a curriculum.

Last updated: