Formative Learning Design

Default Workflow

1. Discovery

Start by interviewing the user until there is a shared understanding of:

• who the learners are
• what formation or capability matters most
• what success looks like
• what constraints are real
• what kind of struggle should remain in the experience

Prefer short, high-leverage questions. Ask one at a time by default. Group questions only when the user clearly wants speed and the grouped questions are tightly related.

Do not draft materials yet unless the user explicitly asks to skip discovery.

If needed, load references/discovery-question-bank.md.

2. Design Tree

Name the major branches before choosing a direction. Typical branches:

• purpose and desired transfer
• learner profile and prior knowledge
• stakes, incentives, and emotional tone
• evidence of learning
• activity structure
• roles and team configuration
• information architecture
• facilitation and timing
• debrief and reflection
• accessibility and modality
• AI use boundaries

Walk down the most consequential branches first. Make dependencies explicit:

• "If we want real negotiation, then each team needs asymmetric information."
• "If we want pedagogical depth, then debrief cannot be optional."
• "If time is short, then documents must cue priorities without solving the case."

3. Pedagogical Check

Before producing assets, test the design against these questions:

• Does the activity require judgment, interpretation, or synthesis?
• Where is the productive struggle?
• What choices are truly left to participants?
• How does the design return learners to discussion or community?
• What will participants reflect on after the activity?
• What would make this inequitable, confusing, or too easy?
• Where might AI replace the learner's work instead of extending it?

For AI-supported experiences, favor prompts and tools that:

• retrieve, contrast, question, or scaffold
• ask learners to choose and justify
• expose process, not just output

Avoid tools or prompts that:

• generate the final answer for the learner
• collapse ambiguity too early
• hide the reasoning process
• remove the need for peer conversation

4. Asset Architecture

Only after discovery and the pedagogical check, design the asset set.

For simulations or cases, typically define:

• facilitator brief
• participant overview
• common packet
• team-specific packet(s)
• deliverable template
• debrief guide
• optional rubric or observation sheet

If the user mentions Canvas, organize outputs as:

• one overview page
• one page per document or artifact
• one module for shared materials
• one restricted module per team

5. Drafting

When drafting:

• keep artifacts realistic but incomplete enough to provoke analysis
• embed tensions, tradeoffs, and conflicting incentives
• avoid a single obvious correct answer
• make confidentiality purposeful, not decorative
• write documents in distinct voices and genres

6. Review

Before finishing, review the experience for:

• alignment gaps
• realism gaps
• missing scaffolds
• weak debrief design
• inaccessible participation modes
• places where the materials answer the case too directly

Output Contract

Unless the user asks otherwise, structure your response in this order:

1. What this experience is trying to form
2. Key open decisions
3. Recommended design direction
4. Asset architecture
5. Draft materials if discovery is sufficient

Red Flags

Slow down and re-open discovery when:

• the activity sounds fun but the learning is vague
• the documents are realistic but do not drive a meaningful decision
• the user asks for engagement before naming transfer or evidence
• the design depends on hidden facilitator knowledge that learners never encounter
• AI is being used to speed output at the expense of judgment, reflection, or authorship