A practice manager at the front desk of a bright Australian general practice, with a small Mac mini beside her laptop.

Bring your practice into the AI era. Keep every record in Australia.

We assess the practice, design a compliant architecture, build the automation and run it on a machine in your rooms.

Referrals triaged before reception has opened the email.

MedAI reads every document in a referral, extracts the details, and recommends a priority and pathway against your own criteria. A clinician confirms. Reception gets a ready-to-paste note.

The MedAI referral queue: three incoming referrals, one marked HIGH priority, one ROUTINE, one awaiting triage. — The referral queue, shown with synthetic patients.

Implementation, not a workshop.

Courses teach clinicians to use AI themselves. We come in, build the automation, hand it over, and stay on to keep it running. Three engagements, each a natural step to the next.

Most practices start here

Practice AI audit

A one-day engagement that maps every workflow touching patient data, scores each automation opportunity for effort, value and compliance risk, and gives you a prioritised roadmap you can take to your principals.

Workflow mapping with your practice manager
Compliance snapshot, including the AI tools staff already use
Vendor shortlist with a compliance verdict on each
Board-ready report and a debrief call

Fixed price, 2 to 8 weeks

Implementation project

We build the agreed workflow, install it on a dedicated Mac mini in your rooms, document it, and train your staff. You own the result.

Monthly

Managed service

We monitor the automation, ship updates, handle vendor changes and stay your point of contact for anything AI. Remote support is restricted and logged.

How a referral gets triaged.

The same pattern applies to results, intake forms and any other document-to-system job.

1
Mail arrives
A rule in the practice mailbox hands referral attachments to MedAI on the local machine. No email body is copied, nothing is sent anywhere.
2
One AI read, in Australia
The documents go to a frontier model hosted in AWS Sydney under a data processing agreement. Fields are extracted and a priority and pathway are recommended against the practice's own criteria.
3
A clinician confirms
The recommendation, its reasoning and a draft reception note appear in the queue. The practitioner reviews and decides. AI never makes the clinical call.
4
Reception acts
Patient details map to the practice software's fields. Write-back is enabled only once the software vendor has given written permission.

A clinician reviewing a referral on a tablet in a modern consulting room.

A MedAI run view: the four pipeline steps on the left and, on the right, the AI's summary, reasoning, recommended pathway, receptionist note and extracted patient fields. — Every step is inspectable. The run view, shown with synthetic patients.

Compliance is the architecture, not a policy document.

Australian privacy law and AHPRA guidance shape how we build, before a line of code is written.

Data stays in Australia

Model inference runs in AWS Sydney on an Australia-only profile, with a signed data processing agreement. APP 8 cross-border disclosure is never triggered.

No consumer AI, ever

Patient-identifiable information never reaches ChatGPT, Claude.ai, Gemini or any consumer tier. We will tell you, in writing, if your staff already use them.

Runs in your rooms

The automation lives on a dedicated machine at the practice. Documents, results and the audit trail stay on that machine. The only network call is the one model read.

Clinicians decide

Every output is a recommendation a practitioner reviews and confirms. Our systems are locked until the practice's privacy impact assessment and consent basis are in place.

Metadata-only audit

The audit trail records what ran, when, and which fields were touched. It is built so that patient values physically cannot be written to it.

Vendor permission first

We integrate with your practice software through official channels or with the vendor's written consent. Never by working around their terms.

Built for practice managers and principals.

You own the operating budget and carry the compliance liability, so the work is designed around your decisions, not an individual doctor's enthusiasm for a new tool.

General practiceSpecialist roomsCardiologyAllied healthReferral-heavy clinics

About MedAI

MedAI is run by Matt Platts, a Melbourne software engineer and engineering manager with fifteen years building and operating production systems. The practice exists because the gap between "AI could help here" and a compliant, running system in a clinic is wide, and most of that gap is engineering and privacy work, not prompting.

Clinical judgement stays with clinicians. We bring the architecture, the build, and the discipline to keep it inside Australian law.

Start with an audit.

One day, one report, and a clear answer to "are we compliant, and what should we automate first?" Tell us about the practice and we will reply within two business days.

Book a practice audit

contact@med-ai.com.au · Melbourne, serving practices across Australia