STEP 1
← Back to case studies
FEATURED CASE STUDY / AI PRODUCT DEVELOPMENT
Give teachers their time back. Keep their judgment in the driver's seat.
Designing an AI-supported substitute planning workflow that starts with teacher expertise—not a blank prompt.
Project Substitute Plan Builder / Teacher Workload AssistantMy role Product concept, learning design & workflow developmentStatus Working prototype / ongoing development
THE DESIGN PRINCIPLE
AI can help with the draft. The teacher owns the decisions.
PROTOTYPE • NOT A MEASURED DEPLOYMENTPRODUCT VISUALS / PROTOTYPE CONCEPT
What the workflow looks like.
These are interface-style concept visuals based on the prototype workflow. They illustrate the experience and structure of the tool while the product continues to evolve.
STEP 2
Validation Check
✓Lesson objective identified
!Dismissal location missing
!Behavior plan requires teacher review
✓Materials list extracted
System action
Flag gaps. Do not guess.STEP 3
Teacher Review
8:00–8:20
Morning Work
Students complete journal prompt and silent reading.
10:15–10:45
Small Group Rotation
Use teacher-provided reading groups and station directions.
AI
Optional transition script
AI-generated suggestion pending teacher approval.
INPUTS
Teacher expertise
- Classroom profile
- Lesson plans
- School procedures
PROCESS
Validation + drafting
- Extract key details
- Flag missing information
- Generate structured draft
OUTPUT
Teacher-approved plan
- Readable for substitutes
- Safer and more complete
- Still owned by the teacher
01 / THE PROBLEM
A substitute plan should not take hours to prepare.
Teachers know their students, their classrooms, and what needs to happen during the day. Turning that knowledge into a plan someone else can follow is another task entirely.
A usable substitute plan needs more than lesson instructions. It must account for classroom routines, materials, transitions, student supports, school procedures, and details that a substitute may not know.
The opportunity I saw was straightforward: could AI reduce the repetitive drafting work while leaving the professional decisions with the person who knows the classroom best?
02 / THE PRODUCT IDEA
Start with what teachers already know.
The tool combines a reusable classroom profile with the teacher's lesson information to create a structured draft substitute plan.
Instead of asking a teacher to invent a detailed AI prompt each time, the workflow captures essential context up front: how the classroom runs, what the substitute needs to do, and what requires particular care.
That context is used to prepare a draft for the teacher to review—not a finished document that silently makes decisions on their behalf.
03 / HOW THE WORKFLOW OPERATES
A practical workflow with human checkpoints.
Capture context
The teacher supplies a classroom profile, lesson information, and relevant procedures.
Extract & validate
The workflow organizes the source material and flags missing or conflicting details rather than quietly guessing.
Generate a draft
AI helps turn approved information into a structured, substitute-friendly plan.
Evaluate & review
Checks assess the draft's completeness and usability; the teacher reviews and approves before use.
This describes the prototype's intended workflow. Features are being refined; it is not presented here as a district-deployed product.
04 / THE DESIGN DECISIONS
Automation with boundaries.
The question wasn't just “Can AI generate a plan?” It was “How do we make the output reliable enough to be useful?”
Don't invent classroom facts.
Unknown or inconsistent information should be flagged for the teacher, not filled in as if it were true.
Make AI additions visible.
When a teacher authorizes AI to suggest a missing example, the suggestion is identified as AI-generated and held for review.
Build in evaluation.
Use an extractor/validator step and an evaluator step to catch omissions and improve plan clarity before teacher approval.
Keep safety and privacy in view.
School procedures and sensitive student information require deliberate handling, appropriate permissions, and alignment with district policies.
05 / WHAT I BUILT & LEARNED
From learning concept to working prototype.
I wanted to test the workflow, not just describe what a useful AI tool might do.
I developed a web-app prototype with a dashboard, classroom profile, and plans workflow, and iterated on how information moves from teacher input through validation, drafting, and evaluation.
Testing surfaced the practical details that matter: missing dismissal information, unclear behavior expectations, safety procedures, and assumptions that might look reasonable to an AI system but be wrong for a particular classroom.
It has also pushed me back into hands-on development, learning new tools and revising the product as problems emerge. My focus is on a workflow teachers could realistically use—not technology for its own sake.
06 / HOW SUCCESS WILL BE MEASURED
Save time, but measure what that time makes possible.
The product is still in development. I haven't claimed measured reductions in teacher workload or burnout.
The next step is to test the workflow with teachers and evaluate draft quality, time to complete a substitute plan, number of corrections, and whether the process genuinely feels easier.
Longer term, I'd want to understand whether time saved is redirected toward instruction, feedback, interventions, and relationships—the work that depends on human expertise.
Design target, not a result: One early project objective was to reduce time spent on administrative tasks by 20%. This remains a hypothesis to test, not an achieved outcome.
07 / WHAT THIS DEMONSTRATES
Product thinking grounded in real human needs.
This project brings together my experience as a classroom teacher, instructional designer, product leader, and curious builder.
It shows how I approach emerging technology: diagnose a real problem, design around the user, test the edge cases, and create guardrails so the tool supports—not replaces—professional judgment.
A note about the prototype
This is a personal product development project, not a commercial product launch or a validated field study. The visuals on this page are concept mockups based on the prototype workflow. Actual screenshots and an interactive demo can be added once a shareable version is ready and privacy/security checks are complete.
LET'S CONNECT
Have a challenge worth solving?
I bring learning strategy, product thinking, and thoughtful execution together to create meaningful change.