New Patient Enquiry Replies and Intake Forms

What does a new patient enquiry reply need to do?

Answer the question asked, point to the next step, and sound like the clinic. Typical enquiries are about opening hours, whether the clinic is taking new patients, how to book, what to bring and where to park.

Every one of those answers comes from the clinic. An AI tool can only help with how it's put.

How can AI help without seeing patient details?

Strip the enquiry down before you paste anything. Replace the name, phone number and email with placeholders, and drop any health detail the person included.

Then ask for a friendly reply to the general question, such as whether a clinic takes new patients and how to book. Add the clinic facts from your own page. Staff paste the result into the real email and add the person's name there.

LLMs can typically write, shorten and translate text.¹ Shortening and rewording are the useful parts here. Making up opening hours isn't.

Why strip the details first?

Because you may not know where the text goes. Ahpra mentions public AI software only as an example in its notes on recording consultations.² We suggest leaving patient details out of public tools as a precaution. Ahpra also tells practitioners to find out whether patient data is used to train the AI model for future patients.²

That guidance is Australian and written for registered practitioners. It still shows what to ask of any tool. The questions are laid out in the patient privacy guide.

Can AI help with intake form wording?

Yes, for the plain-language part. A form's instructions often read like a legal notice. A tool can swap "furnish your particulars" for "fill in your details". It can shorten a long paragraph about how the clinic handles information and make the headings clearer.

Give it the clinic's text with no real patient data in it. Ask for short sentences and everyday words. Ask for a version for someone who reads English as a second language.

What must stay with the clinic?

The questions themselves, if they're clinical. A medical or dental history section asks about health, and the dentist decides what's asked and how. Don't let a tool invent or reorder those questions. The page on tasks to keep with staff explains why.

The privacy notice and consent wording also belong to the clinic and its adviser. A tool can make them easier to read, but it can't make them correct. After any rewrite, a person must confirm the meaning hasn't shifted.

In the United States, the Privacy Rule requires covered entities to notify individuals of the use of their PHI.³ PHI is protected health information. That's a US rule for covered entities. Elsewhere, ask the OAIC (Australia), the ICO (UK) or your own regulator what a privacy notice must say.

What should staff check in a drafted reply?

A short list does it:

• Opening hours, address and phone number match the clinic's website.
• The reply says whether the clinic is taking new patients only if that's currently true.
• Any fees quoted come straight from the clinic's current fee list.
• No promise of a treatment outcome or a particular dentist.
• No health advice, even a friendly one. If an enquiry describes pain or symptoms, a staff member passes it to the dentist or the clinic's urgent-care process.

Where should the data go instead?

Into the clinic's own booking or records system, set up under the clinic's privacy policy. The AI tool should only ever see text you'd be comfortable seeing on a public noticeboard. If you can't do the task without pasting in real details, that's the signal to stop and check the tool's terms first.