AI Generation Workflow
This interactive flowchart illustrates the complete end-to-end workflow for creating MicroSims, from defining a learning objective to deploying a working interactive simulation. The diagram highlights the collaborative interplay between human expertise and AI assistance throughout the process.
About This MicroSim
The workflow is organized into three types of tasks, color-coded for easy identification:
- Human Tasks (Blue) - Tasks requiring instructional design expertise: defining objectives, analyzing prerequisites, selecting visualization types, writing specifications, testing, and adding metadata
- AI Tasks (Green) - Tasks where AI assists: generating code from specifications, iterating based on feedback, and deploying to production
- Quality Gates (Yellow) - Decision points for quality assurance: verifying code works and checking if the MicroSim meets educational standards
Interactions: - Hover over any step to see a detailed description, time estimate, and tips - Click a step to lock the information panel for detailed reading - Click elsewhere to deselect
Workflow Steps Explained
| Step | Type | Description | Estimated Time |
|---|---|---|---|
| 1. Learning Objective Defined | Human | Define what students should understand or be able to do | 5-10 min |
| 2. Analyze Prerequisites | Human | Identify required prior knowledge for appropriate scaffolding | 10-15 min |
| 3. Select Visualization Type | Human | Choose the paradigm (simulation, diagram, chart, network, etc.) | 5-10 min |
| 4. Choose JavaScript Library | Human | Select appropriate library (p5.js, vis-network, Chart.js, etc.) | 5 min |
| 5. Write Specification | Human | Create detailed spec with visuals, interactions, and scaffolding | 15-30 min |
| 6. AI Generates Code | AI | Claude interprets specification and generates complete code | 1-3 min |
| 7. Code Works? | Quality | Test if code runs without errors and behaves correctly | 5-10 min |
| 8. Iterate with AI | AI | Provide specific feedback to refine the code | 2-5 min |
| 9. Test & Validate | Human | Thoroughly test all interactions and educational effectiveness | 10-15 min |
| 10. Meets Standards? | Quality | Evaluate against quality checklist and learning objectives | 5-10 min |
| 11. Refine Specification | Human | Update spec to address quality issues | 10-20 min |
| 12. Add Metadata | Human | Add learning objectives, Bloom's level, tags, accessibility info | 5-10 min |
| 13. Deploy | AI | Commit to repository and deploy via mkdocs gh-deploy | 2-5 min |
| 14. MicroSim Live | Complete | Available to learners on the published site | Ongoing |
Understanding the Iteration Loops
The workflow includes two important feedback loops:
Code Iteration Loop (Steps 6-8)
When testing reveals that the generated code doesn't work correctly: 1. Identify the specific issue (error message, visual bug, interaction problem) 2. Provide clear feedback to the AI 3. AI regenerates code with fixes 4. Re-test until the code works
Typical iterations: 1-3 cycles for most MicroSims
Quality Refinement Loop (Steps 5-11)
When the MicroSim works but doesn't meet educational standards: 1. Identify what's lacking (accessibility, clarity, cognitive load) 2. Update the specification with improvements 3. Regenerate with the refined spec 4. Re-evaluate against standards
Typical iterations: 0-2 cycles depending on specification quality
Key Insights
- Clear specifications reduce iterations - Well-written specs lead to better first-pass code
- Human expertise guides AI output - Domain knowledge and educational goals shape effective MicroSims
- Quality gates ensure effectiveness - Testing against learning objectives, not just functionality
- Iteration is expected - Plan for refinement cycles in your timeline
Total Time Estimates
| Complexity | First-Pass Time | With Iterations |
|---|---|---|
| Simple (single concept) | 45-60 min | 1-2 hours |
| Medium (multiple interactions) | 1-2 hours | 2-3 hours |
| Complex (advanced simulation) | 2-3 hours | 3-5 hours |
Lesson Plan
Grade Level: Professional Development / Graduate
Duration: 45 minutes
Learning Objectives: - Trace the complete workflow from learning objective to deployed MicroSim - Identify human tasks, AI tasks, and quality gates in the process - Recognize feedback loops and understand when they occur - Estimate time requirements for MicroSim development
Activities:
- Explore the Workflow (10 min) - Click through each step to read descriptions and tips
- Follow the Happy Path (5 min) - Trace the ideal flow from start to "MicroSim Live"
- Trace Iteration Loops (10 min) - Follow what happens when code fails or standards aren't met
- Time Estimation (10 min) - Calculate total time for a simple MicroSim with 2 iterations
- Discussion (10 min) - What makes a specification effective? When do you know a MicroSim is ready?
References
- MicroSim Generation Workflow - Swimlane version of this workflow
- Specification Quality Checklist
- Visualization Paradigm Selection
- Human-AI Collaboration Loop