How to Turn a Worked Example Photo Into a One-Page Visual Explainer With Gemini

Problem this solves

This workflow is for teachers, tutors, lab instructors, and coaches who have a good worked example on a whiteboard, document camera, handout, or notebook page and want to turn it into a cleaner one-page explainer fast. It is especially useful right after class, while the explanation is still fresh.

Prerequisites

• A Gemini app or web access with image upload support
• A clear photo of the worked example, board explanation, lab setup, or annotated model
• Good lighting and a readable image
• Optional second pass in NotebookLM, ChatGPT, or Claude if you want a more polished handout later

How to capture or gather the source material

• Use your phone camera straight on, not at an angle.
• Crop out extra clutter so the main example fills the frame.
• If the board has glare, retake the photo before you move on.
• For paper notes, use your phone's document scan mode if available.
• If there are multiple steps on different boards or pages, take one photo per step instead of one far-away photo of everything.

Step-by-step workflow

1. Take a clean photo of the worked example or board explanation.
2. Upload the photo to Gemini with a prompt that asks for extraction, cleanup, and a student-friendly explainer.
3. Have Gemini return the content in a structured format first, such as title, step list, common mistakes, and a quick check question.
4. Review the extracted steps against the original photo. Fix any math symbol, label, or sequence error immediately.
5. Ask Gemini for a one-page visual explainer draft with short labels and plain-language captions.
6. If you want a more polished final product, move the cleaned text into ChatGPT, Claude, or a design tool for final layout.
7. Share the final explainer as a PDF, classroom post, or printed sheet.

Tool-specific instructions

• Primary path: Gemini. Best when the job starts with a phone photo and you need fast extraction plus drafting on the same device.
• Fallback: ChatGPT. Useful for image upload plus second-pass cleanup, especially if you want a more structured final layout.
• Fallback: Claude. Strong for turning the cleaned extraction into a calmer, clearer explainer.
• Optional second stage: NotebookLM. Use it only after you have cleaned text or a scanned PDF and want additional source-grounded study assets.

Copy and paste prompt blocks

{
  "role": "worked-example explainer builder",
  "goal": "Turn the uploaded image into a clear one-page student explainer.",
  "instructions": [
    "Read the image carefully before drafting.",
    "Extract the exact steps in the correct order.",
    "Rewrite the explanation in plain language for the target learner level.",
    "Add a short list of common mistakes or confusion points.",
    "Add one quick self-check question.",
    "Keep the final draft short enough for one page.",
    "If any part of the image is unclear, say what is unclear instead of guessing."
  ],
  "target_user": "replace with grade level, course, or team level",
  "output_format": {
    "section_1": "What this example shows",
    "section_2": "Step-by-step explanation",
    "section_3": "Common mistakes",
    "section_4": "Quick self-check"
  }
}

Quality checks

• Every extracted step matches the photo.
• Symbols, labels, and sequence are accurate.
• The explanation is simpler than the source, not more confusing.
• The quick check actually tests the same skill.

Common failure modes and fixes

• The photo is hard to read. Retake it before asking for the draft.
• Gemini guessed a symbol or label. Correct the extraction first, then regenerate the explainer.
• The result is too long. Ask for a one-page handout with only the essential steps.
• The explanation is still too advanced. Ask for a rewrite at the exact grade or learner level.