⏱️ The Essentials in 3 Minutes |
🧠 Understanding the Pedagogical Value of the Corrected Assignment
The corrected assignment moves the learner from passive learning to active practice: they don't choose from options, they produce. This production is evaluated by the AI, which adapts its feedback to what the learner has actually written.
This format is particularly effective for:
anchoring complex skills that cannot be reduced to a single correct answer;
developing the ability to transfer knowledge to concrete situations;
personalizing learning: two learners who respond differently receive different feedback.
📌 Example: in a management training course, a "reflection" assignment asks the learner to explain how they would respond to a demotivated team member. The AI assesses their response against the defined objectives (active listening, constructive feedback, managerial posture) and indicates their strengths and areas for improvement.
🔀 Choosing the Assignment Type
Type | Objective | Example prompt |
Recall | The learner restates what they have learned. | "List the 3 best practices covered in this module." |
Application | The learner applies a skill. | "Write a response email to this customer complaint." |
Reflection | The learner takes a position on a situation. | "Why is your role key when it comes to sustainable development?" |
🛠️ Configuring the Assignment on the Platform
1. Create a corrected assignment module From the project, add the module title and click Create. Then select the Corrected assignment module type.
2. Enter the module Click on the module you created, then click Edit corrected assignment.
3. Write the prompt and context The prompt is what the learner sees. It must contain all the elements needed to respond.
Frame it as an exercise situation or a concrete scenario.
The more precise the context, the better the learner can project themselves.
Add a context if needed to give the AI more information about your learners, their level, or their environment.
4. Define the pedagogical objectives Objectives guide the AI in its corrections: they are what it relies on to assess the response and formulate feedback. This is the most important step.
Recommendation: select between 3 and 8 objectives. The AI evaluates the response against all objectives, then formulates 3 pieces of feedback on those it considers most important given the learner's answer. Each learner therefore receives different feedback.
⚠️ Only the granule's objective is used by the AI, not its content. Make sure the selected objectives are directly linked to the prompt.
Three ways to add objectives:
Select existing granules from your project to retrieve their associated pedagogical objectives.
Create new objectives in the form: "Untrained learners [common mistake]. Instead, they should [expected behavior]."
Create information granules in a separate module to build a bank of objectives reusable across multiple assignments.
💡 If an objective seems particularly critical, flag it to the AI by adding to its wording: "this is truly ESSENTIAL." The AI will take it into account in its feedback.
5. Save and exit
🤖 Testing the Assignment Before Distribution
Testing the assignment lets you verify that the prompt, objectives, and feedback are consistent.
Simulate an average response: a response that meets some objectives but not all.
Simulate a good response such as a well-trained learner might write.
Ask a colleague who has not worked on the assignment to test it cold.
💡 The AI's feedback varies from one attempt to the next, this is normal. The corrected assignment is a practice exercise, not a graded evaluation. The AI can sometimes be demanding even with a good response.
Keywords : corrected assignment, pedagogical AI, personalized feedback, pedagogical objectives, free production, recall, application, reflection.