Fictional client
By then she had booked with a clinic two streets away. Braemore took 206 aesthetics enquiries last year, 141 waited until the next working day or longer and 43 were never answered at all — but speed on its own would have made things worse, not better. Of the 64 enquiries that were answered within the hour, eleven carried a claim that should never have gone out, and every one of those was typed by a person, out of hours, trying to be helpful. Advertising a cosmetic treatment in Australia is a regulated act. A clinic needs answers that are immediate and defensible, and most manage one at a time.
Braemore Dental & Skin does not exist. We cannot show live client systems, so this study is built around a fictional practice — invented clinic, invented clinicians, invented patients and invented figures. What is not invented is the work: the way enquiries arrive out of hours, the way an automated reply is itself advertising, the way a phrase spreads across surfaces nobody has a list of, and the failure modes are all real. The screens are genuine working pages rather than pictures of pages, and every mockup opens and responds to clicks.
On scope: this is written for private dental and cosmetic practice in Australia, with a Perth clinic as the example. Regulatory requirements are described here in general terms and nothing on this page is advice. Advertising rules for regulated health services change, and differ by procedure — a real build would be configured against the current guidelines with the practice's own advisers, and the claim register is a tool for applying a decision somebody qualified has made, never a substitute for making it. All figures are Australian dollars.
Braemore runs three dentists, an oral health therapist, two cosmetic nurses and a dermal clinician across four chairs and two treatment rooms in Claremont. Four thousand two hundred active patients, $4.6m of fees — $3.3m dentistry, $1.3m skin. The dental half runs on recall: it is predictable, it books itself, and the software for it has existed for thirty years. The skin half runs on enquiries from people who are not yet patients, arrive at night, are comparing three clinics, and are asking about treatments the practice is not free to describe. Nothing in the practice was built for that half.
Every one of these is a version of the same problem: the clinic is judged on what it said, and it has no record of what it said, to whom, or on whose authority.
Two ideas, and the second is what makes the first safe. First, the assistant cannot write a sentence. It assembles replies from phrases that have been reviewed and approved, and where it has no approved phrase it says so and hands the question to a clinician rather than improvising. Speed stops being a compliance risk, because the thing answering at ten at night has a smaller vocabulary than the person who used to.
Second, the register is the only source. The website, the automated replies, the email templates and the phone script all draw from the same approved phrases, so withdrawing one removes it from every surface at once — and the surfaces it cannot reach are listed by name, because pretending otherwise is worse than admitting it.
Seven screens follow. Each one opens as a real page — click through them.
One list of everything that is stopping something from happening — an enquiry from Sunday nobody was assigned, a consent that has to exist before Thursday's treatment, a history form sent twice, and a sentence on the website that has never been reviewed.
An enquiry about lip filler at 22:10, answered in fourteen seconds and holding a Saturday consultation by 22:14. Every outbound message carries the approved phrases it was built from and, more usefully, what it refused to say — a price for the treatment, an estimate of recovery time.
The refusals are the point. A blocked answer is a question the patient wanted answered, so it goes on the record and in front of the clinician before the consultation instead of being fudged at midnight.
Every sentence the practice publishes or sends about a treatment, with each surface it appears on, who approved it and when. Eighty-four phrases across eleven surfaces; six need action today, including a withdrawn phrase still live in an email template and a patient review republished as a testimonial.
The three surfaces with no owner — an unlisted blog post that is still indexed, a directory listing, printed material in circulation — are listed separately, because those are the ones the system cannot reach.
Every enquiry, its source, how long it waited and what became of it. The social lead form produces the second highest volume and the worst outcome by a distance — 62 enquiries, 31 of them never answered — because it was nobody's job.
At the bottom, the reasons enquiries did not proceed, including the five patients who were assessed and told the treatment was not right for them. That bar is the one the practice should want to keep.
Seven years of dental history and three days of skin enquiry, on one record, so the nurse taking a cosmetic consultation can see a sensitivity a dentist wrote down in 2021.
Consent is held as separate items with separate scopes and separate dates — treatment, radiographs, photography for the record, photography for promotion, marketing contact. Five questions, five answers, because they are not the same question.
A cosmetic treatment as the sequence it actually is — enquiry, history, assessment by the practitioner who will treat, photography, written consent specific to what was assessed, a recorded reflection period, treatment with product and batch against the patient, and a review booked at the time rather than left to the patient to arrange.
Across all live pathways: six patients assessed with no consent recorded, and seven treated with no review booked. Nobody chose either of those.
Dental and skin appointments in one list, because the practice should not ask patients to know which half of the building they are dealing with. The full message transcript is there, including the two questions the clinic declined to answer by text.
Consent is not offered here to sign in advance, and the portal says why: it comes after the assessment, because until then there is nothing specific to consent to.
Four phases of two weeks each — eight weeks end to end. The claim register goes in first and the assistant second, in that order, because an assistant that can answer in seconds is only an improvement if it is drawing on language somebody has approved. Each phase ends with something the practice actually uses, not a demo.
Weeks 1–2.
Weeks 3–4.
Weeks 5–6.
Weeks 7–8.
Eight weeks is possible because most of it already exists. The patient and contact records, document handling, consent and expiry tracking, escalation rules, the messaging engine and the portal are running in production across six sectors already, and the conversational SMS and email layer behind the first-response screen is a product we already sell. What actually gets built for a dental and skin practice is the sector core: the claim register, the guardrail that sits between it and the assistant, and the cosmetic pathway. That is the part this study is about, and it is the part we would spend the eight weeks getting right.
This took a fictional practice and made its process visible. The same exercise works better with a real one. Tell us how your clinic actually runs and we will build a working mockup of it — your treatments, your language, your brand — before anybody commits to a build.
Start a Conversation See other sector blueprints