The problem
A privately-owned allied-health group — four sites across regional and metropolitan Victoria, around twenty staff, roughly A$5M in annual turnover — ran all of its appointment booking through a shared reception team. Physiotherapy, podiatry, remedial massage and dietetics, all booked by phone during business hours.
Two problems compounded each other. First, the phone. Reception spent a large share of every day on booking calls — taking new appointments, moving existing ones, chasing people who hadn't confirmed. When the phones were busy, calls went to voicemail, and a meaningful number of those callers simply didn't ring back. Anyone who tried to book after 5pm or on the weekend couldn't; the practice estimated it was quietly losing bookings it never even saw.
Second, no-shows. Missed appointments were running at roughly one in seven across the group. Each empty slot was clinician time that couldn't be resold at short notice — a direct hit to revenue, and a frustration for patients on the waitlist who'd have happily taken the spot.
The practice manager didn't want to add reception headcount, and didn't want a booking "bot" making decisions about patients. She wanted the routine mechanics handled automatically so her team could spend their time on the people in front of them.
What we built
A six-week pilot, run first at the two busiest sites before extending to all four. The system did three things, all of them plumbed into the practice's existing practice-management software rather than replacing it:
- Booking capture, including after-hours. Patients could book online at any hour through the practice's own website, choosing service, practitioner and site. Availability came straight from the live calendar, so double-bookings weren't possible.
- Smart reminders. Confirmations and staged reminders went out by SMS and email, timed to the appointment and the patient's history, with a one-tap confirm or reschedule.
- Self-service rescheduling. Patients could move their own appointment inside the practice's rules, which freed the released slot back to the waitlist automatically.
Ahead of each first appointment, the system also collected intake details — reason for visit, relevant history, consent — so the clinician had what they needed before the patient walked in, and reception no longer transcribed it over the phone.
How it works
The design principle was simple: the system assists reception, it does not replace clinical judgement. It handles the mechanical parts of scheduling; anything that needs a person still reaches a person.
Health information is sensitive, and the pilot was built to Privacy Act 2026 expectations from day one. Patient data is held in Australian-resident storage, encrypted in transit and at rest, and the system only ever writes to the practice-management system the group already trusted — it doesn't create a second, shadow copy of the patient record. Intake questions are limited to what the appointment genuinely needs, with clear consent language.
Where the system is unsure — an ambiguous request, a clinical question dressed up as a booking query, anything outside its rules — it doesn't guess. It routes the request to reception with the context attached, and a person takes over. Reception can see and override every automated action. No clinical decision is ever made by the software.
The results
Measured across the four sites at the end of the pilot, against the practice's own prior three-month baseline (figures are estimates the practice was comfortable standing behind):
| Before | After pilot | |
|---|---|---|
| Reception hours spent on booking admin | baseline | ~17 hrs/week less |
| No-show rate | ~14% | ~10–11% |
| After-hours bookings captured | 0 (phones shut) | a steady stream previously lost |
| First-visit intake completed before arrival | ad hoc, by phone | ~80% self-completed |
| Clinical decisions made by the software | n/a | none |
Around a quarter of no-shows disappeared once reminders and easy rescheduling were in place — people who'd have quietly missed an appointment now moved it instead, and the freed slot went to someone on the waitlist. Reclaiming roughly seventeen reception hours a week let the team spend more time with patients on site and less on hold. Combined with the after-hours bookings that used to vanish, the practice estimated the pilot protected in the order of A$30,000 a year in revenue that had been leaking through empty slots and missed calls — a modest but real return for a business this size.
How the team owns it
We built this to be run by the practice, not by us. Reception configures the reminder timing, the booking rules and the intake questions through a plain settings screen — no code, no support ticket. A short written runbook covers the handful of things that can go wrong and how to fix them, and the practice manager holds the switch that pauses automation and reverts everything to phone-only if she ever needs it.
Two reception staff were trained as the in-house owners during the pilot, so the knowledge lived with the team the day we stepped back. We check in periodically, but the system is theirs to run.
If you're weighing up something similar, our AI automation work starts exactly here, and a short readiness audit is often the sensible first step. Talk to us if that's useful.