Use this when you have a live product exercise coming up (a PM interview whiteboard, a client working session, a kickoff) and you want a board plan that lets people watch you think. It takes a raw prompt and produces a frame-by-frame board layout, color-coded sticky clusters, a spoken talk-track, and a time-box so your reasoning is legible in real time, not hidden in your head. Works for product-sense, prioritization, strategy, modernization, and accessibility prompts.
Related skills: Uses
/2x2-prioritizeand/rice-scorefor the prioritize step,/opportunity-solution-treefor the opportunity step, and/now-next-later-roadmapfor sequencing. Pairs with/interview-prepand/interview-scriptfor the rest of the loop, and/persona-createwhen the prompt needs user segments.
Inputs
- {{PROMPT}}: The exercise prompt as given (or your best paraphrase). Paste it verbatim if you have it.
- {{TIME_LIMIT}}: How long you have (for example 30 minutes live, or a take-home). Drives the time-box.
- {{CONTEXT}}: (Optional, high value) Company, role, interviewer, product domain, and anything you know about what they are testing for.
Process
The board has six frames, one per move in the method spine. Build them left to right so the board reads like a narrative.
Step 1: Translate the prompt (Frame 1)
Pull the load-bearing phrases out of {{PROMPT}} and pin them up verbatim. Decompose into Audience / Technology / Problem-to-solve. Restate the goal in the asker's own words as a one-line product or feature vision. Add the measurement question: "how might we measure this?"
Output a labeled cluster:
AUDIENCE: "<quoted phrase>" → <who this really is>
TECHNOLOGY: "<quoted phrase>" → <capability in play>
PROBLEM TO SOLVE: "<quoted phrase>" → <what is actually broken>
─────────────────────────────────────────
PRODUCT / FEATURE VISION: <one line that ties them together>
^ How might we measure this?
Step 2: Define the opportunity (Frame 2)
Build an opportunity tree: one framing question, three to five child questions, then sharper sub-questions. Convert the sharpest into How-Might-We reframes. Go wide here before narrowing.
DEFINE THE OPPORTUNITY: <framing question>
├─ <child question 1> → HMW <reframe>
├─ <child question 2> → HMW <reframe>
└─ <child question 3> → HMW <reframe>
Step 3: Ground in reality (Frame 3)
Anchor the abstraction. Pick the right grounding for the prompt:
- User-segment spectrum when the prompt is about a population (draw the spectrum, mark what breaks at each point).
- Research snapshot when you need to know who the user is (demographics, behaviors, three takeaways).
- Live data read when you are handed metrics (turn the table into a heat map; see Step 4).
Reason with concrete numbers even if illustrative. Note assumptions you are making.
Step 4: Prioritize visibly (Frame 4)
Make the tradeoff legible. Choose one:
- 2x2 with axes named in plain, memorable language (call
/2x2-prioritize). Spell out what each axis end means ("we can begin solving this tomorrow"). - Heat map for a metric-by-team or metric-by-segment grid (color = at/above benchmark vs. opportunity to improve).
- Action tiers when numeric scoring would be slower than the decision needs: DO NOW / PREPARE-NEXT / GET READY.
Pair each prioritized item with a How-Might-We so a problem reads as an opportunity. Optionally call /rice-score if the panel wants quantitative scoring.
Step 5: Sequence (Frame 5)
Turn the prioritized set into a plan with a shape. Use the spine Outcome → Focus Areas → Features → Next Steps, and/or a Now / Next / Later roadmap (call /now-next-later-roadmap). Write a falsifiable hypothesis. Keep one item visibly traveling from the 2x2 into the roadmap so the thread is unbroken.
OUTCOME: <the change you are trying to create>
FOCUS AREAS: <2-4 themes>
FEATURES: <the bets under each focus area>
NEXT STEPS: <what you would do Monday>
HYPOTHESIS: We believe <bet> leads to <outcome>.
Step 6: Close the loop (Frame 6)
Never leave a solution un-instrumented. End with success metrics, a validation plan, and observability: how you would know it worked, what you would test before committing, what you would watch after shipping. Mark assumptions validated or invalidated. Close human: "Celebrate. Keep going."
Step 7: Apply the craft layer (across all frames)
Before finalizing, pass the whole board through the craft checklist in reference/craft-checklist.md: color legend, visible-cognition stickies, at least one memorable rename, spatial left-to-right narrative, triage of anything inherited, and the big decision named out loud.
Step 8: Write the talk-track and time-box
Produce a spoken talk-track (what you say as you build each frame) and a time-box that fits {{TIME_LIMIT}}, allocating minutes per frame and leaving room to narrate.
Output format
# Whiteboard Plan: <prompt in a phrase>
**Time limit:** {{TIME_LIMIT}} **Context:** <company / role / domain>
## Board layout (left to right)
Frame 1 Translate · Frame 2 Opportunity · Frame 3 Ground · Frame 4 Prioritize · Frame 5 Sequence · Frame 6 Close
## Color legend
<context> · <users> · <solutions> · <metrics> · <open questions>
## Frame-by-frame
### Frame 1: Translate the prompt
<the cluster>
### Frame 2: Define the opportunity
<the tree>
... (frames 3-6)
## Talk-track
A short line per frame: what you say as you build it.
## Time-box (total = {{TIME_LIMIT}})
| Frame | Minutes | Goal |
|---|---|---|
## Craft check
A line confirming each item in the craft checklist is satisfied.
Related resources
reference/board-template.md: a blank six-frame board you can copy into Miro.reference/craft-checklist.md: the legibility checklist to pass the board through.
Output location
Present the whiteboard plan as formatted text in the conversation. The user recreates the frames in Miro (or any whiteboard) and uses the talk-track and time-box live.