AI features, the AI register and automated decisions
See which AI and automated features your organisation uses and where they run, switch AI drafting on or off, and build the privacy policy wording due from 10 December 2026.
Where: Settings in the menu, then AI register. The privacy policy wording is in Safety and quality, Privacy, Automated decisions and services.
The AI register
The tabs are Features, Services and locations, Usage log and What workers and participants are told.
- Features lists each AI and automated feature, whether it is on, What leaves Inclusi, the Label on the output and who decides. Safety checks (for example the roster's hard stops) cannot be switched off: they only stop or warn, and a person decides.
- Services and locations shows which outside service each feature uses and where it processes data.
- Usage log shows who used which feature and how it went. Inclusi does not keep what was sent or written.
- What workers and participants are told shows the wording in the apps and the portal.
AI drafting is off until your organisation turns it on and Inclusi has an AI service set up. Everything AI writes is a labelled draft that a person checks and saves.
Your privacy policy: automated decisions
From 10 December 2026, under the Privacy Act (APP 1.7 to 1.9), your privacy policy must describe decisions made by computer programs that could significantly affect people. On Automated decisions and services, under Copy into your privacy policy, choose Build the text. Inclusi writes a Privacy policy: automated decisions paragraph from the features you have switched on, and a Collection notice: use of AI paragraph. Copy them into your own documents.
Automated decisions: before 10 December 2026 is a checklist of what to do, ticked off with the date and who ticked it: list the automated features you use, decide which decisions could significantly affect someone, build your privacy policy paragraph, have your privacy officer or lawyer check it, publish the updated policy before 10 December 2026, check people are really deciding, be ready to explain a decision, and build the text again when you change features.
Overrides lists each time someone chose differently from a suggestion, accepted warnings or disagreed with a reportable incident suggestion, which shows that people make these decisions.
Inclusi's classification of its own features is a starting point. Have someone qualified check your privacy policy.
Related articles
- Handle a request for information or a correction
- For workers: voice notes, AI tidy-up and sign-off at clock-out
- Reports and Ask Inclusi
- A worker's phone is lost or stolen
- API keys and webhooks
Need more help?
Contact support with the page you were on and what you expected to happen. Do not send participants' health details. Is Inclusi down? See the status page.