Clinicians lose hours to notes, letters, referrals and reporting. We help health organisations take that load off with AI, while keeping patient information inside the rules that govern it: the Privacy Act, TGA medical device regulation and AHPRA's expectations of practitioners.
Most of the AI tools now sold to practices do something useful. An ambient scribe drafts the consult note, a document tool reads incoming referrals, a reporting agent pulls numbers out of the practice management system. The difficulty is everything around the tool. Health information is sensitive information under the Privacy Act 1988, and the OAIC is clear that the Act applies to every private sector health service provider whatever its size or turnover. Recording a consultation and sending it to a vendor is a collection and a disclosure, so consent, notice and where the data goes all need answers before go-live.
Then there is the TGA. Software, including AI, is regulated as a medical device when its intended purpose is diagnosis, prevention, monitoring, prediction, prognosis or treatment. The TGA's guidance gives the example of a digital scribe that starts outside the definition and moves inside it once the developer adds diagnosis or treatment suggestions. Features change with every vendor update, so a tool that was administrative when you bought it may not stay that way. Our TGA software as a medical device guide explains where the line sits.
Finally, AHPRA's guidance confirms that practitioners remain accountable for anything AI contributes to the record or to a decision. The AHPRA AI guidance post sets out what that means day to day.
We qualify by the problem in front of you. A single clinic and a multi-site group can have the same one.
Clinicians finishing notes, letters and reports after hours, or referrals sitting in an inbox waiting to be keyed in.
A vendor demo looked good and you need someone to check data flows, consent and the TGA position before you sign.
Practice management, billing, rostering and reporting in separate systems, with staff copying data between them.
Aged care, NDIS or accreditation reporting assembled by hand from several sources each cycle.
Staff already pasting patient details into public chatbots because nothing approved exists yet.
Clinical or HR documents in SharePoint and Teams, with Copilot on the roadmap.
If patient files live in Microsoft 365, start with our Copilot readiness assessment, because Copilot answers from whatever each user can already open. For the public chatbot risk, see how the wrong AI tools leak business data.
Each candidate use case is checked against its intended purpose, the TGA definition of a medical device and the APPs that apply to the data it touches. Administrative use cases with clear value go first.
For every tool, we document what is captured, where it is processed and stored, whether it leaves Australia, how long it is retained and whether the vendor trains on it.
Referral intake, correspondence, recalls, claims paperwork and reporting, built with a clinician or staff member approving anything that reaches a patient record. See our process automation service.
Connecting practice management, finance and reporting systems so data is entered once. Our system integration work covers the APIs and the access controls around them.
Where contracts or policy require in-country processing, we design self-hosted models inside your environment through our private AI infrastructure service.
Patient notices, staff AI use policy, review steps for AI-drafted notes and an incident path, aligned to the AHPRA guidance and your privacy policy.
Related services: process automation, system integration and private AI infrastructure. Background on hosting choices is in our data sovereignty guide.
See what this would take in your business
A free 30-minute review on Microsoft Teams. Bring the process you want fixed and leave with a clear first step.
This screen is a starting point and does not replace a regulatory determination. Intended purpose is set by the manufacturer's own claims, so we read the vendor's materials alongside the product. Where the answer is unclear, the TGA accepts questions and pre-submission meetings, and we recommend you use them.
Timeframes depend on how quickly vendors answer data handling questions, so we set them after discovery. Engagement options are on our pricing page, and if you have not yet settled on priorities, our AI strategy service is the place to start.
For front-desk call handling, see healthcare practices. For a wider view of the obligations covered here, read Privacy Act compliance and AI, our post on aged care AI under the new Act and the guide to NDIS registration and automation. Once tools are live, our managed AI services keep the review cycle running.
Book a 30-minute scoping call. Bring the process that costs your clinicians the most time and any tool you are already evaluating, and we will tell you what we would check first.