What Personal Data Should You Minimise Before Using an Automation Platform?

What personal data should a small business exclude or minimise before sending work through an automation platform?

Exclude any personal data that the automation does not need for its specific intended purpose. For data it does need, send the least detail that will still let the task work, explain the processing clearly, and apply minimisation before collection where possible. If the processing is likely to create high risk for individuals, European GDPR guidance says a data protection impact assessment is mandatory. There is no evidence-supported universal exclusion list, so assess fields against the purpose rather than relying on broad category labels.

Experian UK says data minimisation requires organisations to collect, process and store only the minimum personal data required for the specific intended purpose.
Data Minimisation: Guide to Collect Less, Protect More - Experian UK · As of 2026-11-14
The UK Information Commissioner’s Office says minimisation techniques should, where possible, be applied before collecting personal data as part of mitigating risks that large datasets can pose.
How should we assess security and data minimisation in AI? | ICO · As of 2026-11-14
The European Data Protection Board says personal-data processing activities should be explained in clear language and adapted to each category of handler in documents they can refer to.
Secure personal data | Data protection guide for small business | European Data Protection Board · As of 2026-11-14
The European Data Protection Board says GDPR requires a data protection impact assessment for processing of personal data likely to result in high risk for individuals.
Secure personal data | Data protection guide for small business | European Data Protection Board · As of 2026-11-14
The European Data Protection Board says secure personal data is in the interests of both individuals and the organisations processing it.
Secure personal data | Data protection guide for small business | European Data Protection Board · As of 2026-11-14

Start with the automation’s exact purpose

Experian UK says data minimisation requires organisations to collect, process and store only the minimum personal data required for the specific intended purpose. For an automation, write that purpose as one short statement of the task and expected outcome before deciding what any record may contain.

Sources: Data Minimisation: Guide to Collect Less, Protect More - Experian UK.

Use the pilot controls to make a bounded workflow decision before trying to define the minimum data required. Treat each available field as excluded by default until you can state what the task needs it for.

  • Write the task before reviewing fields.
  • Keep only information required for that task.
  • Treat available data as optional until its purpose is clear.

Minimise before personal data enters the workflow

The UK Information Commissioner’s Office says minimisation techniques should, where possible, be applied before collecting personal data as part of mitigating risks that large datasets can pose. The UK Information Commissioner’s Office says a number of techniques can help organisations minimise the amount of personal data they need to process.

Sources: How should we assess security and data minimisation in AI? | ICO.

Use the process readiness guide to check whether the underlying work is stable enough to define what information the task actually needs. For every retained field, ask whether omitting it breaks the task and whether a less detailed value still allows the task to work; omit, reduce or retain based on those answers.

  • Review inputs before a live record enters the workflow.
  • Remove fields with no task-specific purpose.
  • Reduce the detail of retained values where that still supports the task.

Record what is sent, why it is needed and who handles it

The European Data Protection Board says personal-data processing activities should be explained in clear language and adapted to each category of handler in documents they can refer to. Create a short processing record that names the task, the approved minimum inputs, the reason each retained input is needed, the handlers involved and the permitted use.

Sources: Secure personal data | Data protection guide for small business | European Data Protection Board.

A bounded run record can help explain what happened in an automation without making the approved data scope unclear. Use plain wording that a person configuring, submitting to or reviewing the workflow can act on, and revise the record when the boundary changes.

  • Describe the activity and approved purpose.
  • List the minimum inputs and their reason.
  • Identify handlers and give them usable instructions.

Pause for an impact assessment when processing may be high risk

The European Data Protection Board says GDPR requires a data protection impact assessment for processing of personal data likely to result in high risk for individuals. If the proposed activity may meet that threshold, defer the live-data decision and seek the appropriate assessment and controls before proceeding.

Sources: Secure personal data | Data protection guide for small business | European Data Protection Board.

A process that should not be automated until its risks and operating conditions are clearer should remain outside the automation boundary. This is GDPR guidance, not a determination that GDPR applies to every reader; confirm the rules and advice relevant to your business and location.

  • Do not assume a reduced data set removes all risk.
  • Pause when processing may be high risk for individuals.
  • Confirm applicable legal and professional requirements for the business.

Make minimisation part of the approval decision

The European Data Protection Board says secure personal data is in the interests of both individuals and the organisations processing it. Experian UK says data minimisation can help meet personal-data obligations, reduce security risks and support key GDPR rules.

Sources: Secure personal data | Data protection guide for small business | European Data Protection Board; Data Minimisation: Guide to Collect Less, Protect More - Experian UK.

Approve the boundary only after adding human approval, limits and review appropriate to the proposed workflow. Choose go when the purpose and every retained field are clear, revise when a field can be reduced or better explained, and stop or defer when the business cannot define a controlled minimum set.

  • Go when every retained field has a stated purpose and clear handling instructions.
  • Revise when a needed field is too detailed or poorly explained.
  • Stop or defer when the purpose, risk or controls remain unclear.

Purpose-to-data approval decision guide

Use this decision guide for one proposed automation. It does not create a universal exclusion list; it helps you decide whether the workflow has a clear, limited and documented data boundary.

DecisionPurpose testWhat to recordAction
GoEach retained field is required for the stated task.Task, minimum inputs, reasons, handlers and approved use are clear.Approve the defined boundary and keep it under review.
ReviseA field may help, but its necessity or level of detail is unclear.Record the uncertainty and the reduced or omitted alternative to test.Remove or reduce the field, then test the revised input set.
Stop or deferThe business cannot define the purpose, minimum input set or handling instructions.Record why the boundary cannot yet be approved.Do not send live personal data until the gap is resolved.
EscalateThe processing may be high risk for individuals.Record the proposed activity and assessment question.Pause for the appropriate assessment and controls before proceeding.

This is an operational decision aid, not a legal opinion or proof of compliance. Confirm applicable obligations and seek appropriate advice where the processing may be high risk for individuals.

When this is not the right approach: Do not use this guide as a substitute for an assessment required by applicable law or for professional advice on a potentially high-risk activity. Do not proceed with live records when the workflow’s purpose or minimum inputs cannot be stated clearly.

Another perspective

A universal list of personal-data fields to ban can create false confidence. Test every field against the automation's specific purpose and minimise it before collection where possible.

What follow-up questions matter most?

Is there a universal list of data fields to exclude?
Use the automation's specific intended purpose as the test. Omit a field when the task does not need it, reduce its detail when a less detailed value will work, and retain it only when you can explain its role in the task.
When should data minimisation happen?
Minimise before collection where possible. Review the workflow and its proposed inputs before a live record is sent, rather than treating the source system's full record as the default input.
Does every automation need a DPIA?
The European Data Protection Board says a DPIA is mandatory under GDPR for processing likely to result in high risk for individuals. Whether that threshold or GDPR applies to a particular business depends on its circumstances and applicable law.
What should we document before sending data?
Record the task, the minimum required inputs, why each retained input is needed, the handlers involved and the approved use. Keep the record practical enough for the people handling the workflow to use.

What steps does this workflow follow?

Set a personal-data boundary before automation

  1. Name the task: Write one plain-language sentence describing what the automation must do and the outcome it is meant to produce.
  2. Inspect proposed inputs: List the information the workflow could receive, then remove fields that do not serve the stated task.
  3. Reduce retained detail: For each remaining field, decide whether a less detailed version would still let the automation complete its task.
  4. Document the boundary: Record the required inputs, why they are needed, the approved use and the people who handle the activity.
  5. Escalate possible high risk: Pause for an appropriate assessment when the processing may be high risk for individuals, rather than treating the reduced input set as automatic approval.
  6. Approve, revise or stop: Proceed only when the purpose, minimum inputs and handling instructions are clear; otherwise revise the workflow or defer it.