Medical practice management software has quietly absorbed a lot of AI functionality over the past two years, which means the real question for an Australian practice in 2026 isn't whether to adopt AI, most already have some, but which layer of the stack to trust for which task, and where a general-purpose assistant genuinely adds something the built-in tools don't.
What's already built into your practice management software
Best Practice, Medical Director, and Genie have all shipped AI-adjacent features over the past two years: predictive appointment scheduling, basic clinical coding assistance, and some recall automation. These are useful and worth using where they exist, since they're already integrated with your patient records and billing. Their limitation is scope: they handle what the vendor decided to build, and anything outside that (a nuanced patient communication, a practice-specific admin report, a non-standard recall sequence) falls back to manual work regardless of how good the built-in AI is.
Clinical scribe and note-taking tools
A separate category of tools, ambient clinical scribes like Heidi Health and Lyrebird Health, has grown fast in the Australian market specifically for capturing consultation notes without a doctor typing during the appointment. These are genuinely strong at what they do and worth adopting on their own merits for practices wanting to reduce after-hours note-writing. They're purpose-built for the clinical documentation moment specifically, not for the broader admin load around the practice.
Where a general-purpose assistant fits: the admin layer around the clinical work
Recall and reminder sequences tailored to specific patient cohorts, not just a generic due-date trigger.
Referral letter drafting from consultation notes, formatted to a specific specialist's preferred style.
Practice-level reporting: a monthly summary of patient volume, no-show rates, and billing trends that a practice manager would otherwise assemble by hand.
Staff rostering and multi-doctor scheduling support for practices juggling several practitioners' availability.
Patient-facing communication drafting (appointment confirmations, results-ready notifications) that stays within AHPRA and Privacy Act boundaries by design.
The clinical boundary every Australian practice should set explicitly
Any AI tool touching a medical practice needs an unambiguous line: administrative drafting is fair game, clinical advice or diagnosis is not, and anything that references a specific patient's health information stays inside Privacy Act and AHPRA-compliant handling at every step. This isn't a reason to avoid AI tools, it's the specific configuration work that makes them safe to use, and it's the first thing any properly scoped setup for a medical practice should establish before a single workflow goes live.
Cost and a realistic starting point
A practice-level Cowork setup covering recall, referral drafting, and reporting typically runs $3,000 to $4,800 depending on which practice management system is in use and how many workflows launch in the first phase. Most practices get the clearest early win from recall automation alone, since a missed recall is both a patient-care gap and a lost appointment, worth fixing before expanding further.
A worked example: recall automation done properly
A three-doctor general practice in Sydney's inner west was relying on its practice management system's basic recall list, a due-date trigger that generated a generic reminder regardless of what the patient actually needed a follow-up for. Diabetic patients due for a HbA1c check got the same templated message as someone overdue for a routine skin check, and staff had no easy way to prioritise which overdue recalls actually mattered clinically. After setting up a Cowork workflow that read the specific reason for each recall from the clinical notes and tailored both the message and the urgency flagging accordingly, the practice saw a meaningful jump in recall completion for its higher-priority chronic-disease cohort within the first quarter, exactly the patients where a missed recall carries the most clinical risk.
That's the pattern worth taking from this: the built-in recall function wasn't wrong, it just treated every overdue patient the same way, and the fix wasn't a new platform, it was a smarter layer on top of the system already in place.
The same logic extends to referral letters. A specialist referral drafted from raw consultation notes, formatted the way each individual specialist actually prefers to receive it, saves a GP genuine time across a busy clinic day, and getting that formatting right for your specific referral network is exactly the kind of practice-specific configuration a generic, unconfigured AI tool won't get right on the first try.
A short conversation with your practice manager about where the current recall and referral process actually breaks down is the right starting point, well before comparing specific vendors or products.
If your practice is weighing up what to add on top of your existing practice management software, get in touch through our contact page for a straight read on what fits.
