Data Privacy & Security
How we handle
your residents' data.
Indicator Pro analyses aged-care progress notes to surface clinical risk patterns. Here is exactly what happens to your data — from the moment you export it to the moment it appears on your dashboard.
- 01
Export from your care system
You export progress notes as a CSV or PDF file directly from your care management software (e.g. eCareOne, Telstra Health). No automated connection — you control exactly what leaves your system.
Your facilityYour data - 02
Encrypted upload to Australia
The file travels over an encrypted HTTPS connection (TLS 1.2 or higher) to secure storage in Sydney (Supabase Storage, ap-southeast-2). It never touches any server outside Australia.
TLS 1.2+ encryptedSydney · ap-southeast-2 - 03
De-identification — before the AI sees anything
Before a single note is sent for analysis, our software automatically strips all personal identifiers and replaces them with neutral placeholders. The AI only ever sees the clinical content.
What gets replaced
Resident names(full name, surname, nicknames, honorifics)[CLIENT]Phone numbers(mobile, landline, +61 forms)[PHONE]Email addresses[EMAIL]Medicare numbers[MEDICARE]Formatted dates of birth(DD/MM/YYYY format)[DOB]Structured street addresses(with AU postcode)[ADDRESS]Before
"Mrs Klose was assisted with her morning shower. She reported pain in her left hip. Phone family on 0412 345 678."
After de-identification
"[CLIENT] was assisted with her morning shower. She reported pain in her left hip. Phone family on [PHONE]."
- 04
AI analysis — Sydney only
De-identified note text is sent to the AI model running inside AWS Sydney (ap-southeast-2). The AI reads the clinical content and assigns it to indicator categories. AWS Bedrock's standard service terms prohibit using customer data for model training, and model invocation logging (which would store inputs or outputs) is off by default.
AWS Bedrock · SydneyNo invocation loggingNo model training - 05
Categories returned — no note content stored offshore
The AI returns only structured classification metadata — category names, confidence scores, and clinical sub-types. No note text is returned. These results are matched back to the original note record in the Sydney database. The note text never leaves Australia at any point.
No note text returnedNote text stays in AU - 06
Clinician review, then your dashboard
A clinical staff member reviews the AI-assigned categories before a batch's data appears in reports. Corrections are recorded as new entries — the original AI output is never deleted. Your provider team then sees the final indicators on the dashboard.
Staff review requiredFull audit trail
In summary
All data stored in Australia
Sydney AWS ap-southeast-2
Encrypted in transit and at rest
TLS 1.2+ · AES-256
AI never trains on your data
AWS Bedrock — no invocation logging by default
Note text never leaves Australia
De-identified text only sent to AI
Staff review on every batch
Human approval before data appears in reports
Privacy Act 1988 compliant
APPs + Notifiable Data Breaches scheme
Questions?
If your IT team or compliance officer has questions about our data handling, we are happy to provide further detail.
Boucher Advisory Group Pty Ltd · ABN 36 685 315 477
Operated under Privacy Act 1988 (Cth) · Australian Privacy Principles