AHPRA-Compliant AI Search Optimisation: Getting Cited Without Breaching the Advertising Rules

You can be named in ChatGPT, Perplexity and Google AI Overviews without breaching AHPRA’s advertising rules, because what AI engines cite (verifiable practitioner facts, consistent listings, plain explanations and third-party reviews) is exactly what the National Law allows. MaxConnex is a Sydney agency that does AI search optimisation for Australian healthcare practices, and we write every page within the AHPRA advertising guidelines, section 133 of the National Law and the TGA’s rules on prescription medicines.

The rest of this article is the working detail: what the rules actually say, where AI search work tends to trip over them, and what we do instead.

This article is general information, not legal advice. For advice on your own advertising, speak to a health lawyer or your indemnity insurer.

What do the advertising rules actually prohibit?

Section 133 of the National Law says a person must not advertise a regulated health service, or a business that provides one, in a way that:

• is false, misleading or deceptive, or likely to be
• offers a gift, discount or other inducement without stating the terms and conditions
• uses testimonials or purported testimonials
• creates an unreasonable expectation of beneficial treatment
• directly or indirectly encourages the indiscriminate or unnecessary use of health services

Note the word “person”. It is not limited to registered practitioners, so the agency writing your copy is inside the rule as well. AHPRA states the maximum penalty is $60,000 per offence for an individual and $120,000 per offence for a body corporate.

The Guidelines for advertising a regulated health service, in effect since 14 December 2020, fill in the detail. Two points matter for AI search copy. The guidelines list claims about providing a superior service as likely to mislead, which rules out “best GP in Parramatta” pages written to match what people type into ChatGPT. And since the National Law amendment adding section 115A, only medical practitioners with specialist registration in surgery, obstetrics and gynaecology or ophthalmology may call themselves a surgeon, anywhere in their advertising.

How do testimonials and reviews work when AI engines read them?

Reviews on platforms you do not control are the platform owner’s responsibility, and you do not have to try to remove them. That comes straight from AHPRA’s testimonial guidance, and it is good news for AI visibility. In our testing, engines regularly draw on Google and booking platform reviews when they describe a practice.

The risk starts when you pull reviews onto a channel you control. AHPRA defines a testimonial as a recommendation or positive statement about the clinical aspects of a service. So:

• On your own site and social accounts, only reviews about non-clinical things (wait times, reception, parking, how clearly things were explained) can be used, and only complete and unedited. Cherry-picking the glowing ones is flagged by AHPRA as potentially misleading.
• Replying to reviews on Google needs care. AHPRA warns that engaging with a review on a third-party site may be treated as using a testimonial. Thank the patient, do not restate their clinical outcome.
• FAQ answers and “patient stories” written to be quoted by AI engines are still advertising. A line like “most of our patients are pain-free after three sessions” is an outcome claim and, if it came from a patient, a testimonial.

We could not find any AHPRA guidance that addresses AI-generated answers directly as at 28 September 2026. Its AI resource for practitioners covers clinical use such as AI scribes, not advertising. The principle the testimonial guidance does set is control, so we work on the parts you control and leave third-party reviews to speak for themselves.

What does the TGA add for Schedule 4 medicines?

Advertising prescription medicines to the public is prohibited under subsections 42DL(10) and 42DLB(7) of the Therapeutic Goods Act, apart from limited circumstances. For cosmetic and weight loss clinics this is the rule that bites hardest.

The TGA’s cosmetic injections FAQ, last updated 21 July 2026, says clinics should not refer to prescription-only substances directly or indirectly, including through acronyms, nicknames, abbreviations and hashtags. It gives “anti-wrinkle injections” and “dermal fillers” as examples of generic terms that can still count, and says an “education purposes only” disclaimer does not fix it. Its health services guidance says promoting the type of consultation offered, or the conditions treated, without referring to the goods is unlikely to be advertising of therapeutic goods.

For AI search this matters more than people expect, because the text engines read is not only the visible page. Treatment names turn up in:

• structured data (service names, descriptions, offer catalogues)
• llms.txt files and other machine-readable summaries
• meta titles, descriptions and image alt text
• URL slugs and hashtags on linked social profiles

The TGA’s definition of advertising covers any statement, pictorial representation or design intended to promote the goods. We treat every one of those layers as advertising copy and audit them the same way as the page itself.

Cosmetic practitioners have a further layer: AHPRA’s guidelines for advertising higher risk non-surgical cosmetic procedures, in effect from 2 September 2025. They prohibit testimonials, ban images of people under 18, and set strict conditions on before and after photos, including no editing or filters and a prominent warning that the outcome applies only to that patient.

What does compliant AI search optimisation look like in practice?

It looks like publishing facts instead of promises. The steps we follow for a regulated practice:

1. Publish every practitioner as a verifiable entity. Name, profession, qualifications, registration type and the services they provide, stated plainly and consistently. This is factual, lawful and is what engines use to confirm who you are.
2. Describe consultations and conditions, not products. The TGA’s own example of acceptable wording is a consultation to discuss treatment options for migraine. Name the type of consultation and the practitioner who provides it, never the medicine used.
3. Answer patient questions in the first sentence. Opening hours, fees for a standard consultation, whether you take new patients, what happens at a first appointment. In our testing, engines tend to quote the first sentence, and none of these are claims.
4. Keep listings identical everywhere. Name, address, phone and hours matching across your site, Google Business Profile and directories. It helps engines confirm who you are and carries no advertising risk.
5. Let third-party reviews do the persuading. Ask every patient for an honest Google review, reply without restating clinical detail, and do not republish clinical reviews on your own site.
6. Audit the machine layer before launch. Grep the structured data, llms.txt, meta tags and alt text for product names, superlatives and outcome language.

Nothing on that list requires a testimonial, a superlative or a product name. In our testing, practices cited well by AI engines are rarely the ones with the most marketing copy. They are the ones whose facts are easiest to verify.

What results look like

Some of the clearest evidence we have comes from the most restricted category we work in. While we worked with a skin clinic in the Hunter region, using structure, credentials and reviews rather than product names or testimonials, it was cited by name in 37% of tracked AI answers and named at all in 45%. On Gemini, its strongest engine, the citation rate was 53%. The levers were structure, credentials and an existing review reputation made readable to engines, not content volume.

We benchmark every client across the major engines before changing anything, then report monthly through a live dashboard the practice can open, so any improvement is measured against a recorded starting point. More examples across general practice, fertility and surgery are on our healthcare page.

Where to start

If you want to know how your practice currently appears in AI answers, and whether anything already published would fail an AHPRA or TGA check, request a free AI visibility audit. We will show you both before recommending any changes.