Concrete workplace scenario
A teammate says: “AI says the customer qualifies for the discount. Can I send this?”
It looks like AI already answered. The trap is that fluent wording is not proof, and the worker still owns the evidence check.
Worked answer
Bad example first: the answer sounds certain, but proof is missing
Weak example
“The policy allows this discount. Send the approval.”
What goes wrong
The weak answer treats policy allows as fact with no approved policy text, then jumps to send the approval because AI sounded confident.
Work-ready version: separate the claim from responsibility
Claim
Capture exactly what AI said, without treating it as true yet.
Evidence
Attach the approved source or mark the proof as missing.
Owner
Name who can confirm the policy and own the final answer.
Quick decision check
The claim / evidence / owner rule
A confident answer is only usable when the claim, evidence, and owner are all visible.
Claim
What did AI state?
Evidence
Which source proves it?
Owner
Who owns the truth?
Responsibility 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 says the discount is allowed, so use it.
Repair move: Separate the AI claim from evidence and approval owner.
Better row: Claim: discount may apply. Evidence needed: approved discount policy. Owner: sales/support manager confirms before sending.
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 treat confidence, fluent language, or detailed formatting as proof.
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 separate claim from proof?
- Does it name the evidence source or mark it missing?
- Does it name who owns the final truth?
What this artifact proves: This file proves you do not treat confident AI output as proof.
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 responsibility_boundary_check.md.