Care hours, spent on paperwork.
named for every AI system, with the authority to pause it. Health information is sensitive information under the Privacy Act, so we scope around that first.
Australian providers run on documentation: clinical notes, referrals, care plans, claims, rosters and compliance reporting. Every hour spent on it is an hour not spent with a patient, and the workforce shortage makes those hours the scarcest asset in the sector.
Private practice · GP clinics · Aged care · Disability and social care · Life sciences
What we hear from healthcare.
The same four problems, in almost every practice, clinic and provider we talk to.
Documentation load
Clinicians and carers finish their day with hours of notes, care plans and progress reports. Burnout follows the paperwork, not the caring.
Missed bookings
No-shows, unfilled cancellations and phone queues cost revenue and frustrate patients who just want an appointment.
Referral and intake friction
Referrals, intake forms and eligibility checks are keyed by hand between systems that do not talk to each other.
Compliance reporting
Aged care and NDIS providers carry heavy, repeating reporting obligations, assembled by hand from scattered records.
First use cases, with the number we scope to.
Boring and useful first. Each one is scoped in the free audit with a target, a named owner and a human review point before anything is built. The targets below are what we set; the audit puts your own number on them.
Assisted documentation
Draft notes, care plans and letters from the encounter, reviewed and signed by the clinician. The record stays the clinician's; the drafting time goes.
Referral and intake triage
Referrals read, classified and routed within the hour, missing information chased automatically, eligibility checked against the rules you give it.
Booking recovery
Cancellations offered to the waitlist and reminders handled across SMS and phone around the clock, with the front desk taking only the calls that need a person.
The rules that apply.
Every use case we scope in healthcare starts with who may see what, where the data flows, and where a person must review. This is what we check it against. It is not legal advice; it is the list your auditor will ask about.
- 01Privacy Act 1988 and the Australian Privacy Principles. Health information is sensitive information. Collection, use, disclosure and overseas transfer each need a lawful basis. The OAIC's October 2024 guidance covers using commercial AI products.
- 02Ahpra and the National Law. Registered practitioners stay responsible for the clinical record and the clinical decision, whatever drafted it. Ahpra's position on AI in practice applies to every assisted-documentation tool.
- 03Aged Care Act and the Aged Care Quality Standards. Providers carry documentation and reporting obligations to the Aged Care Quality and Safety Commission that an assembled compliance pack must satisfy, with a person signing it off.
- 04NDIS Practice Standards. Participant records, incident reporting and the NDIS Quality and Safeguards Commission's requirements shape what an agent may write and who reviews it.
- 05My Health Records Act 2012 and state health records law. Access to My Health Record, and the Victorian and NSW health records acts, bound what any system may read and retain.
Guardrails that matter most here
- 03Data governed: minimum data per task, permissions that follow the user, every answer traced to a source.
- 05Human in control: nothing clinical leaves without a practitioner's review and sign-off.
- 06Plainly disclosed: patients are told when a booking reminder or a letter came from an agent.
How a first engagement usually runs.
Week one and two: the free audit. We walk the front desk, the clinical documentation flow and the compliance calendar with the people who run them, and score the six pillars. Governance usually scores lowest, because nobody has written down who owns the data or where a person must review.
Weeks three to eight: one boring use case, built with the staff who will live with it, on a free trial. Referral triage or booking recovery most often, because both produce a countable number inside a month.
Then the roadmap. The next use case is chosen by the number the first one produced, and the whole thing runs against the Guardrails Framework with a named owner from day one. See the case studies for how this ran in other sectors.
Questions.
Is patient data safe with AI?
Only if the architecture enforces it. We design so agents see the minimum data the task needs, permissions follow the user, and every answer traces to a source. Health information handling is assessed against the Privacy Act and your professional obligations before anything is deployed.
Will clinicians have to change how they work?
The tools fit the workflow, or they get worked around. We start with documentation and admin that clinicians already resent, keep them in control of anything clinical, and train the team properly. Adoption is a design requirement, not a hope.
Does patient data leave Australia?
Not unless you decide it should. The audit maps where each use case's data would flow, including which models sit underneath and where they run. Overseas disclosure under APP 8 is a decision you make with the facts in front of you, not a default.
Where do we start?
With the free audit: a readiness scorecard, your three highest-return use cases with a number on each, and a 90-day roadmap. Three hours of your time.