Quick Guide · Data handling
How to Plan What Happens to Data After AI Use
Decide what to keep, move, restrict, review or delete after an AI task instead of leaving data behind by accident.
The idea to keep
The task is not finished until the data has a clear destination and review trigger.
Best for
- Customer-support transcripts
- AI-generated drafts and extracts
- Personal or confidential data
What you need
- Outputs and source data
- Retention and access rules
- An owner and delete trigger
Give every output a destination
1
Identify what was created
List the source data, prompt, output, export and copied version.
2
Decide retention value
Keep the approved outcome; do not retain scratch material just because it exists.
3
Assign destination and access
Name the system of record, permitted users and restricted version.
4
Set a review or delete trigger
Choose a date, event or owner that checks whether the data is still needed.
Example
Customer-support transcript
Keep
Approved outcome and final response in the case record.
Move
Relevant source evidence to the restricted case folder.
Restrict
Raw personal data to authorised support staff.
Delete
Scratch copies after the case review.
Check before you use it
- Source and output identified.
- Access limited appropriately.
- Destination is the right system.
- Review or delete trigger visible.
Common mistake
Keeping every copy “just in case” without deciding which version is authoritative.
Nova9 view
Data handling is part of the task
A clear after-use decision prevents accidental retention and makes accountability visible.
