# Protect client data and reuse rights

Canonical: https://manyspace.aporia.systems/topics/protect-client-data-and-reuse-rights-earn-ai-008
Updated: 2026-10-24

Earning Money with AI · Deliver trustworthy work · EARN-AI-008

- Difficulty: 2 / 5
- Importance: 4 / 5
- Study time: 0.5–1 h

Minimise data, agree permissions and choose a safe review and deletion process.

## Why it matters

Client trust depends on what you collect, where it goes and whether you have permission to use it.

## Where it is used

Write a data-handling checklist covering purpose, minimum fields, approved tools, access, retention, deletion and incident response.

## How you know you have learned it

Complete the scenario checks, explain the decision in your own words, and produce this artifact: Write a data-handling checklist covering purpose, minimum fields, approved tools, access, retention, deletion and incident response.

## Trace each data handoff

Permission and minimisation must be checked before the first upload.

Get permission for the specific purpose and destination. Remove unnecessary personal or confidential data, including hidden metadata and free-text details. Check the current tool’s retention, training use, access controls and deletion options; a subscription label is not proof of confidentiality. Agree retention and incident handling with the client. Never use client work as a portfolio sample without explicit permission. Escalate sensitive or regulated work to appropriate qualified support.

You need to demonstrate a support-ticket classifier before the client has approved a tool.
1. Create synthetic tickets that do not copy real identities or confidential events.
2. Test the classification format and review process with those samples.
3. Document the proposed data fields and destination for client approval before using real records.

A synthetic demonstration tests the workflow while the real-data transfer remains unapproved.

Removing names from tickets makes them safe to publish, correct?

Not necessarily. Dates, addresses, unusual events and combinations of details can identify people. Check context and permission; use synthetic examples for public demonstrations.

## What you need first

- [Create a reliable AI workflow](https://manyspace.aporia.systems/topics/create-a-reliable-ai-workflow-earn-ai-007) (required)

## What it opens

- [Package proof without fake results](https://manyspace.aporia.systems/topics/package-proof-without-fake-results-earn-ai-009)