Concrete workplace scenario
A team asks whether AI can help with a complaint, a contract note, and an employee issue because all three are writing-heavy.
The trap is that the risk is not writing. The risk is authority, consequence, and trust.
Worked answer
Bad example first: text-heavy does not mean automation-ready
Weak example
“Write the final complaint decision, contract position, and employee warning.”
What goes wrong
The prompt asks AI for a final complaint decision, contract position, and employee warning — direct automation of consequence.
Work-ready version: prepare, escalate, do not decide
Prepare
Summarize facts, list questions, draft neutral options.
Escalate
Route to legal, HR, manager, or support owner.
Do not decide
No final judgment, promise, approval, or discipline by AI.
Quick decision check
The prepare, don’t decide rule
For sensitive tasks, AI can prepare the work around the decision boundary; it must not cross the boundary.
Prepare
What can AI safely organize?
Escalate
Who owns the consequence?
Do not decide
What must stay human-owned?
No-direct-automation 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 small contrast to see what changed today. The goal is not more paperwork; it is a clearer AI-in-workflow move.
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: AI writes final employee warning.
Repair move: Move AI before the decision boundary.
Better row: AI may summarize facts and draft neutral questions; manager/HR owns any warning or disciplinary wording.
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 directly automate legal, financial, medical, hiring, disciplinary, refund, contract, or customer-trust decisions.
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 it say what AI may prepare?
- Does it say what AI must not decide?
- Does it name the human owner of the consequence?
What this artifact proves: This file proves you can keep AI in the preparation zone and outside human decision authority.
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 do_not_automate_list.md.