Concrete workplace scenario
A marketing assistant has two announcement drafts and asks AI to make the weaker one better.
It sounds like a taste problem. The trap is that iteration without criteria becomes random rewriting.
Worked answer
Bad example first: “make it better” hides the criteria
Weak example
“Make this announcement better and more engaging.”
What goes wrong
The prompt says better and more engaging with no criteria, no comparison, and no keep/repair decision.
Work-ready version: compare before revising
Compare
Judge A/B against audience, accuracy, and format.
Keep
Name what improved and should stay.
Repair
Name what got worse or still fails.
Quick decision check
The compare before revise rule
Iteration starts with comparison criteria, not taste words.
Compare
Judge A/B against audience, accuracy, and format.
Keep
Name what improved and should stay.
Repair
Name what got worse or still fails.
Iteration trace
This is the small visible work change from today. Keep it shallow, real, and reusable tomorrow.
Before / after: what changed in your work?
Use this contrast to check whether today changed the actual work result, not just the prompt wording.
Before learning
After learning
What AI-in-workflow means today
Build today's artifact rows
Complete three short fill-ins. Keep one real work item per row, and change the sample wording so it fits your work.
| No. | Work item | AI role / prompt case | Risk or evidence | Human owner / reader | Next action |
|---|---|---|---|---|---|
| 1 | |||||
| 2 | |||||
| 3 |
Repair one weak row
Use this pass to find the row that could confuse someone, create unsafe AI use, or fail as workplace evidence before you export.
Repair example
Weak row: Prompt: Make this announcement better and more engaging.
Repair move: Add criteria, compare versions, keep strengths, and repair weaknesses.
Better row: Compare A/B for audience fit, accuracy, and format; keep clearer headline from A; repair missing call to action in B.
Choose the row you would be comfortable showing to a manager. It should have a clear task, boundary, owner, and next action.
Choose the row most likely to fail, confuse someone, expose sensitive input, or need review.
Write the exact change. Sentence starter: The weak row is weak because ___. I will repair it by adding ___ before using AI.
Pass / fix / stop standard
Do not iterate by taste alone; compare against the criteria defined before generation.
Bad prompt → improved prompt → changed output
Bad prompt
Bad prompt: “Make it better.”
Improved prompt
Improved prompt: “Create version B that is shorter, keeps the same evidence, names the owner, and removes unsupported claims. Then compare A/B against evidence, reader fit, and action clarity.”
Changed output behavior
Changed output: iteration becomes criteria-based instead of random polishing.
Export artifact
This file is today’s work receipt. It should show what changed in your work habit, not just that you filled a table.
Before you export
- Does the artifact name comparison criteria?
- Does it say what to keep?
- Does it say what to repair next?
What this artifact proves: This file proves you can compare versions before asking AI to revise.
Weak export: vague rows, no owner, no repair note, or no next-use line.
Good export: one clear work case, one boundary/rule, one repair note, and one tomorrow-use or manager-use line.
Click Generate Markdown to create iteration_comparison_note.md.