Concrete workplace scenario
A customer thread has names, account IDs, order numbers, and internal notes. You want AI to help draft a reply.
It looks like you should paste the thread so AI has enough context. The trap is that the input itself may be unsafe.
Worked answer
Bad example first: more context can create a privacy problem
Weak example
“Here is the full customer thread with account details and internal notes. Draft a reply.”
What goes wrong
The bad prompt includes the full customer thread, account details, and internal notes with no redaction.
Work-ready version: make the input AI-ready first
Remove
Delete names, IDs, credentials, and confidential notes.
Replace
Use neutral placeholders like [customer], [order], or [policy].
Ask
Escalate if restricted detail is required for the task.
Quick decision check
The remove / replace / ask rule
Safe AI use starts before the prompt: remove what should not be shared, replace what AI can still reason about, and ask when the boundary is unclear.
Remove
Take out sensitive fields.
Replace
Use safe placeholders.
Ask
Check the approved process if needed.
Privacy 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: Paste the full customer email thread for AI to rewrite.
Repair move: Remove identifiers and replace private details with placeholders.
Better row: Use [customer], [order], and [issue summary]; omit account ID and internal notes before asking AI for reply structure.
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 paste names, identifiers, credentials, employee records, or confidential terms into AI without an approved process.
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
- Did you remove sensitive identifiers?
- Did you replace needed context with placeholders?
- Did you avoid pasting restricted notes or credentials?
What this artifact proves: This file proves you can turn raw work input into safe AI-ready input.
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 privacy_input_filter.md.