What is happening?
Check the task and the information the tool could see.
A practical guide for Australian support teams
Useful tools.
Clearer choices.
Less guesswork.
Before you use AI or automation, check what it does, what information it sees and what needs a human review.
Choose from simple categories. Get a clear next step, the reasons behind it and safer workflows to consider.
Seven short sections. No sign-in or record upload.
The whole workflow matters.A tool name alone does not tell you where information goes or who makes the decision.
Describe the work. Keep the person out of it.
Use categories only, not names, case notes, diagnoses, recordings or real records. Answers stay in this open page and clear on refresh. Downloads and printouts are separate copies you control.
One task. Three useful questions.
Where does it go?
Look at processing, supplier terms and connected systems.
What needs to change?
Get a practical direction and the first actions to review.
This is a screening guide, not legal advice, a compliance certificate, a final privacy policy or NDIS approval. The underlying pilot sources were checked 29 September 2026. The updated OAIC APP 1 guidance was rechecked 2 October 2026; the screening rules are unchanged. Source links open external sites only when selected.
See how the details change the result
Explore 14 fictional workflows
Selecting an example loads a completed check; use Edit answers to explore a variation.
Your workflow
Start with one task
Your practical direction
Ordinary AI or automation
Privacy and supplier handling
December 2026 disclosure
First actions for this workflow
Brief tool-use inventory
Draft staff-use rule to adapt
Limits: This direction depends on the categories supplied. It does not decide legal coverage, certify a tool, replace a privacy impact assessment or produce a final privacy policy. Current privacy duties and NDIS responsibilities can apply even when the December requirement does not. Read the detailed reasons before use.
The December test, element by element
The law concerns an APP entity’s use of a computer program, personal information about the affected individual, a substantial and direct decision contribution, and a potentially significant effect. It includes beneficial outcomes, refusals and failure to decide.
If the test applies, what would the policy need to describe?
APP 1.8 requires the kinds of personal information used, the kinds of decisions made solely by computer programs, and the kinds of decisions that programs substantially and directly help to make. Give a specialist your workflow inventory to assess these categories. This check does not write or approve the final privacy-policy wording.
Why this result and what to change
These are the findings triggered by your answers. Each identifies its basis; a recommended safeguard is not presented as a new legal rule.
A narrower workflow to consider
Alternatives matched to this task
These options may reduce particular risks. A tool swap does not settle permission to use information or remove an automated-decision obligation. No option has been tested against your records or systems.
Questions for the supplier or technical owner
All selected categories
Options to investigate
Change the workflow, the information or the tool
Start with the specific problem. A local model can reduce cloud disclosure; a managed account can improve organisational control; a simpler rule-based process may avoid unnecessary AI. None resolves every risk.
Desktop does not automatically mean offline. Separate the app interface, model processing, connected services and permission to act. A robot that controls a local app may still send screenshots to a cloud model. A fully local robot can still make unwanted changes.
Understand the basis
Examples, rules and evidence
The worked app scenarios are Betty’s fictional applications, not official legal determinations. The examples below identify what appears in the law or regulator material.
Open the source register
Pilot sources checked 29 September 2026. On 30 September, the OAIC published updated APP 1 guidance and new automated-decision transparency resources. That update was reviewed on 2 October 2026. The May issues paper remains historical consultation material, not final guidance. Check current sources before relying on a result.