What is the AI-Driven Feedback Activity?

Didask’s AI-driven feedback activity lets you personalize each learner's pathway. This activity allows learners to respond individually to a shared prompt, with the AI virtual coach tailoring feedback based on each response, fostering a more individualized, immersive learning experience.

As the course designer, you’ll need to:

👉 For instructional design guidelines and step-by-step advice, see the "Design" section later in this article.

1️⃣ We recommend selecting between 4 and 16 objectives. The virtual coach will review learners' submissions against these objectives and identify the top 3 areas where learners could benefit most from feedback.

👉 This way, each learner receives personalized feedback based on their needs.

2️⃣ Preview the Activity: Test the activity’s flow by previewing it as often as needed.

👉 Unlike learners, who have only one attempt, you can reload the page to simulate multiple trials. Learners will see only their response and any ongoing feedback dialogue with the coach after submitting their answers.

Once the feedback-based activity is added to a publication in the learner catalog, learners can complete it as part of their learning path.

1️⃣ Completing the Feedback-Based Activity

Learners will:

2️⃣ Evaluating Mastery Levels

Each objective has three mastery levels within the AI’s feedback structure:

⚠️ Learners are granted one additional attempt to address objectives marked as "Partially Achieved" or "Not Achieved."

3️⃣ Completing the Activity

The activity is considered complete when:

After publishing, you can track performance metrics for the AI-driven feedback activity on the publication statistics page.

1️⃣ Global Statistics

Access an overview with essential information, including:

2️⃣Individual Statistics

In the individual statistics export, each feedback-based activity will appear as a module, with columns for:

👉 You’ll also find progression details on learners' profiles.

Before creating the activity, define your instructional goals. This will help you shape the instructions, objectives, and vary activities if desired.

Here’s a goal example for each type of feedback-based activity:

Recall Activity: I want my learner to recall what they’ve just learned (e.g., list best practices, define terms).

Application Activity: I want my learner to complete a task to apply what they’ve learned (e.g., write a cover letter, design a process).

Application Activity: I want my learner to complete a task to apply what they’ve learned (e.g., write a cover letter, design a process).

Reflective Activity: I want my learner to take a stance on a topic based on the training (e.g., “Why is your role as a school leader crucial for sustainability?”).

1️⃣ Identify the Activity Type

Ask yourself:

2️⃣Create the Activity

Creating an AI-driven feedback activity includes two main steps: drafting the instructions and defining objectives. Ensure these elements are cohesive:

1/ Writing the Instructions

The instructions for an application-based exercise should include all the elements learners need to respond effectively. Learners engage more easily in an activity when the context is clear, allowing them to immerse themselves in the scenario, even if it doesn’t perfectly match their exact use case.

👉 Approach it as a scenario-based exercise.

2/ Selecting Objectives

Evaluation objectives should be based on “learning granules” (information or practice). You can pull from existing course objectives or create custom ones. ⚠️ Note: Only the objective itself is used by the AI, not the full granule content.

💡 Tips for Designing Objectives:

🎨 Creating an Objectives Bank

Custom objectives can be formatted as specific actions or learning outcomes (e.g., “Instead of praising only effort, learners should recognize improvement”). If needed, create separate information granules for easy reuse across multiple activities.

This approach helps you avoid rephrasing errors and enables easy reuse of these objectives across multiple activities.

🧭 Guiding AI in Prioritizing Key Objectives

During the activity, the AI coach will first evaluate the learner’s response across all objectives and then provide feedback on the top 3 objectives it identifies as:

This means that not all learners will receive the same feedback. If you believe certain objectives are especially crucial, you can guide the AI by marking these as “high-priority” with notes like “this is absolutely ESSENTIAL.”

👉 The AI coach’s feedback is based on these selected objectives, so choosing objectives intentionally is a crucial step.

For example: Untrained learners tend to praise someone for their qualities or talent rather than the effort made. Instead, they should focus on the learner's progress by acknowledging the effort—this is absolutely ESSENTIAL.

⚠️ Note: Avoid selecting all topic-related granules as objectives, as the AI coach's feedback may be inaccurate if the objectives do not align with the prompt provided to learners.

🏁 Verifying Your Objective Choices

To ensure your objectives are well-chosen, consider asking yourself:

Testing an activity allows you to verify that your objectives and instructions are well-aligned and to adjust their wording if needed.

👉 Keep in mind that the AI coach’s feedback can vary slightly with each attempt.

👉 This is a practice exercise; the AI coach may sometimes challenge learners to expand on a response that seems adequate, but the learner is not graded.

Sur le même thème, découvrez aussi :

Keywords: AI, educational AI assistant, personalized learning, intelligent feedback, auto-generated feedback

Need more help? Feel free to contact us at [email protected]. Our team is here to support you with your projects! 💬