Tool bench
The first tool lab is guided, not assumed.
- AI critique setup: open one clean chat, paste the structured prompt, and copy only useful evidence.
- FigJam/Figma setup: create the first board and type the exact observation zones.
- Evidence save: keep the prompt, AI output, board link or screenshot, critique notes, and final decision.
Weak versus strong
Click
Weak
Start with "open ChatGPT and Figma" as if the first step is already obvious.
Strong
Follow the path into the tool before producing anything.
Type
Weak
Ask for a board or prompt without showing the expected text structure.
Strong
Give exact text for the first prompt, board labels, and decision sentence.
Check
Weak
Trust an AI answer or visual board because it looks finished.
Strong
Inspect specificity, evidence, uncertainty, and what to accept, adapt, or reject.