Copywriting

It’s not just a matter of reordering the words, or getting rid of the well-known AI hallmarks like em dashes and colons.
Even changing a sentence here and there won’t make much difference as Anthropic’s version subtly biases the model’s word choices rather than using hidden characters.
The watermark survives copy and paste as well as light editing.

Fixing a typo or rewording a sentence will not remove it.

• The only real way around this is to rewrite the majority of the copy, so it’s recognised as human text, rather than AI.
• Watermark detection isn’t perfect either way. A mark being present doesn’t prove whether it was edited afterwards – and a mark not present doesn’t prove a human wrote it.
• For a medical practice or a hospital, however, the takeaway is simple: assume that unedited AI drafts are identifiable or will be very soon.

Think about it. If it was your mother or sister having a kidney cancer operation or a knee arthroscopy, would you prefer a doctor-reviewed website or “health churnalism” pushed out by AI? Whilst AI works well for many industries, for your patients and your peers, a real voice matters. Why? Because patients are not buying a toaster, they are seeking vital information about a health condition. This means that generic, non-specific copy will impact the trust and credibility of your website. Regulatory and ethical scrutiny is higher in healthcare as well. Copy that turns out to be unedited AI output can quite often be a bigger liability than copy that is carefully copywriter and doctor-reviewed.

No. Merging man and machine usually gives the best result. AI is a genuinely useful drafting and research tool and is excellent at summarising complex clinical material into layperson language. The issue is not abandoning AI altogether, but using it judiciously, interspersing it with journalistic acumen, a doctor’s voice, and automated compliance checks.

Clinical accuracy and appropriate disclaimers.
A copywriter with extensive experience working in Australian healthcare is likely to red-flag claims that need qualifiers.
Health writers also often know which phrases cross a line that regulators such as AHPRA and TGA care about.
Physician review provides a second set of eyes, and a final AI compliance review will even further clarify things that need scrutiny.
Keep in mind that all legal responsibility rests with the provider, and whilst we are not lawyers, we can confidently state things we believe need further scrutiny.

A distinct voice.
Generic AI frequently reads as wordy, generic and voiceless and lacks the kind of specific information that can make or break a patient’s choice to see a particular provider.
We make your practice sound exactly like YOUR practice, not everyone else’s practice.

Patient-centred framing
Good copywriting anticipates every question a patient might ask.
This means understanding your patients through years of experience in interviewing them as journalists, rather than just discussing a topic in general and generic terms.
Our focus is symptom first, not service first. Decision support, not sales talk.

Substantive edits
Light edits don’t cut it for watermarks; real improvement means restructuring and adding real expertise.

Google rankings reward content that is high in E-E-A-T (experience, expertise, authoritativeness and trust) – and this is what big institutions like the Mayo Clinic and Cleveland Clinic do by default. A watermark-free piece typically has real, specific details baked in that show genuine clinical nuance and reflection rather than a raw AI draft. Not using watermarked copy also shows specific expertise – e.g. a faster referral gateway through your practice. This is what both humans and machines read as “expertise.” Journalist-written copy will also frequently cite credible sources, such as journals or a meta-analysis that may be relevant to a particular condition. A heavily-edited, expert-informed piece is far more persuasive than a generic AI-edited draft. When it comes to trust, this is an especially big factor for healthcare. Trust is built through warmth, accuracy, specificity and voice consistency.

Yes, but our goal isn’t to launder AI text past a detector; it’s to wind up with the content that is genuinely accurate, on-brand and something you’d be proud to publish under your name. In other words, it has your voice and is still actually YOUR copy, not something that is there specifically to game an AI detector.

Our writers specialise in healthcare and medical marketing. We have worked in newsrooms, hospitals and universities so can view healthcare from every angle. We know how to translate clinical accuracy into patient-friendly language. We also know that while AI helps the world move faster, it will never replace the judgement and expertise your patients are trusting YOU to have.