Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for chief digital officers

Data Request Privacy Review Agent

Extracts that contain only what the purpose needs and cannot easily identify a person

Data Request Privacy Review Agent: what goes in, what the agent does and what you get

What it does

Data extracts go out with personal data that the recipient does not need. For each extract, this agent checks the fields against the stated purpose and your policy. It tests whether combinations, such as postcode, birth date and job title, could identify a person. It proposes masking, generalizing or aggregating fields. After each change, it rechecks the extract to see whether the risk is gone and the extract still serves its purpose. The data owner approves release. Edge case: a small group has only three people in one postcode, so the agent groups the area.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Extract is about to be shared 2 USES A TOOL Read the extract fields and the stated purpose 3 DOES Mark each field as needed, not needed or sensitive 4 DOES Remove or mask fields that are not needed 5 USES A TOOL Test combinations of fields for small groups 6 CHECKS THE RESULT Is every group at or above the minimum size? If not: generalize or aggregate the fields until it is.Back to step 4. 7 CHECKS THE RESULT Does the extract still meet the purpose? If not: restore the minimum needed detail and retest.Back to step 2. 8 YOU APPROVE Data owner approves release 9 RESULT Release record with fields and changes
Read the steps as a list
  1. Extract is about to be shared
  2. Read the extract fields and the stated purpose
  3. Mark each field as needed, not needed or sensitive
  4. Remove or mask fields that are not needed
  5. Test combinations of fields for small groups
  6. Is every group at or above the minimum size?If not: generalize or aggregate the fields until it is. Back to step 4.
  7. Does the extract still meet the purpose?If not: restore the minimum needed detail and retest. Back to step 2.
  8. Data owner approves releaseThe agent waits here for your OK.
  9. Release record with fields and changes

How it decides

A field stays only if the purpose needs it. A combination is risky when a group has fewer than a set number of people.

  • Keep only fields the purpose needs
  • Require a minimum group size of 10
  • Generalize dates and locations before removing them
  • Block release when the purpose is unclear

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Minimum group size (default: 10)
  • Sensitive field list
  • Masking methods
  • Who approves

What keeps you in control

It always asks you first

  • Data owner approves release

Hard limits

  • Never releases data itself
  • Never keeps fields the purpose does not need

It stops when

  • Done: the extract is approved and logged
  • Stop: the purpose is missing or the recipient is unknown

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA request for staff survey data asked for age, postcode, job title and scores. The agent removed name and email, then found 12 groups smaller than 10. It replaced age with ten-year bands and postcode with region. The recheck found two groups of 8, which it merged. The purpose was still met. The owner approved.

More agents for chief digital officers