AI Chatbots for Clinic Websites: What's Compliant and What Converts

17 August 2026·7 min read

Quick answer: A chatbot on your clinic website is brilliant for the boring stuff — hours, parking, booking links, after-hours “got it, we’ll call you back” replies. It becomes a real risk the moment it starts answering anything that sounds like a health question. The fix isn’t avoiding chatbots, it’s scripting them tightly: FAQs and booking triage in, anything clinical redirected straight to a human or an appointment. Below is a script you can copy into your chatbot builder today, plus the setup and privacy guardrails that keep it compliant. 📱

Every few weeks I get the same message from a clinic owner: “We installed a chatbot and now I’m nervous about what it’s telling people.” Usually what happened is a well-meaning practice manager turned on an off-the-shelf AI widget, let it run on default settings, and it started answering questions like “is this normal after my knee surgery” with confident, made-up-sounding advice. That’s not a chatbot problem. That’s a scripting problem. Done properly, a chatbot is one of the highest-leverage little tools you can add to a clinic website — it catches after-hours enquiries, filters out the “what are your hours” messages before they hit your front desk, and nudges hesitant people toward booking. Done badly, it’s a liability sitting on your homepage. Here’s how to get it right. 💖

First, what most clinics get wrong with chatbots

Turning on a generic AI chatbot and letting it freely answer open-ended health questions instead of hard-limiting it to logistics and booking
No visible, easy way to reach an actual person — the bot becomes a wall between the patient and your reception team
Capturing symptoms or health details in chat without thinking about where that data goes or who can see it
Running the exact same script for every practitioner type in a multi-disciplinary clinic instead of routing people properly
Expecting the chatbot to compensate for a booking system that’s already slow, confusing, or broken

Steal this: a copy-paste chatbot script

An 8-response decision tree for a clinic chatbot. Copy this straight into your chatbot builder as your core menu, then edit the details for your practice.

1. Opening hours: “We’re open Monday to Friday, 8am–6pm, and Saturday mornings 9am–12pm. Want to book in?”

2. Booking an appointment: “You can book online here [link], or I can take your name and number and our team will call you back within one business day.”

3. Parking and location: “We’re at [address], with free 2-hour parking right out front / in the [X] car park. Need directions?”

4. Do I need a referral: “Great question — it depends on your situation and whether you’re using a health fund, Medicare, or NDIS plan. Our team can check this for you when you book, or you’re welcome to just come in without one and we’ll sort it out together.”

5. Fees and rebates: “Our fees start from $[X] and most private health funds offer a rebate for this service. For your exact out-of-pocket cost, it’s best to chat with our front desk — want me to pass your details along?”

6. Cancellations and rescheduling: “No worries — you can reschedule online here [link] or reply with your name and preferred day and we’ll sort it. We ask for at least 24 hours’ notice where possible.”

7. After-hours message: “Thanks for reaching out! Our team is offline right now, but we’ve got your message and someone will be in touch on the next business day. If this is urgent, please call [emergency line / 000 for medical emergencies].”

8. Any health, symptom, or “is this normal” style question: “That’s something our practitioners are best placed to answer properly, rather than me guessing. Let’s get you booked in so they can take a proper look — would you like me to check our next available appointment?”

How this actually plays out

Physio clinic, after-hours enquiries: A Robina physio was losing enquiries that came in at 7pm when someone had just tweaked their back and started googling. The chatbot now captures name, number, and preferred day, sends an instant “we’ve got you” reply, and the message lands straight in the clinic’s booking inbox for the morning. No advice given, no diagnosis attempted — just a warm holding pattern until a human takes over.

Psychology practice, waitlist and intake: A private practice with a genuine waitlist uses their chatbot purely to set expectations — it explains the current wait time, walks people through what the intake form covers, and flags that a Mental Health Care Plan needs a GP referral first. It never asks what someone is struggling with. That conversation is reserved entirely for the intake call with a clinician.

Multi-practitioner clinic, routing enquiries: A clinic running physio, podiatry, and dietetics under one roof uses the bot as a traffic director. First question is simply “which service are you after?” — the reply then routes to the right booking link and a short, service-specific FAQ set, so a foot pain enquiry doesn’t end up in the dietitian’s calendar by accident.

Getting the setup right

The technical side matters more than the flashy chat bubble. Connect the chatbot to your actual booking software (Cliniko, HealthEngine, Halaxy, or whatever you run) rather than leaving it as a standalone island — every “book now” reply should link somewhere real. Build in a human handoff at every stage, not just as a last resort: a visible “speak to our team” option, a direct phone number, and a promise of a response time you can actually keep. And treat anything typed into that chat window as health information the moment a patient mentions symptoms, conditions, or Medicare details — even if your script tries to redirect them away from it. That means checking where your chatbot platform stores conversation logs, who on your team can access them, how long they’re kept, and whether that setup fits with your practice’s privacy policy and the Australian Privacy Principles. This isn’t a “nice to have” step — it’s the difference between a helpful widget and a data-handling headache.

💡 Heads up: a chatbot cannot fix a booking system that’s genuinely broken. If your online calendar is a mess, your confirmation emails don’t send, or reception is already drowning in calls, the bot just moves the frustration to a different screen. Fix the plumbing first — the chatbot is the finishing touch, not the foundation.

Mistakes that quietly undermine a clinic chatbot

Leaving the AI in “open” mode instead of restricting responses to a defined FAQ and booking script
Hiding the phone number or reception email so the bot feels like the only option
Never testing the bot yourself by asking it a tricky, symptom-shaped question to see how it responds
Forgetting to update the script when hours, fees, or referral requirements change, so it starts giving out-of-date information

Frequently asked questions

Can a chatbot legally give health advice on our website?

No — and it shouldn’t try to. Any AHPRA-registered clinic needs to be careful that nothing on the website, including automated chat, resembles clinical advice, diagnosis, or symptom interpretation. Keep the bot firmly in the lane of logistics, FAQs, and booking, and redirect every health-related question to an appointment or a call with your team.

Should we use patient reviews or testimonials in the chatbot script?

Leave clinical outcome testimonials out of chatbot scripts entirely, along with the rest of your marketing where AHPRA advertising guidelines apply. Stick to factual, practical information the bot can say with total confidence — hours, location, process — rather than anything implying a guaranteed result.

Will patients actually use a chatbot, or do they just want a human?

Both, honestly — it depends who’s asking. Younger and time-poor patients often prefer messaging a bot at 9pm over waiting for a callback. Older patients, and anyone dealing with something they’re anxious about, often actively dislike chatbots and want to hear a real voice. A clear, easy “talk to a person” option isn’t optional — it’s what makes the bot feel like a convenience rather than a barrier.

What’s the biggest privacy risk with clinic chatbots?

Patients typing more than you asked for — symptoms, medication names, Medicare numbers — into a chat window that wasn’t built to handle sensitive health information securely. Choose a platform with proper data handling, limit what you ask for, and make sure your privacy policy actually reflects that a chatbot is collecting enquiries. Please note: general information, not legal advice — check current official guidance before relying on it.

Keep reading 🤍

Written by

Kate, founder of Chronically Online

I help Gold Coast and Brisbane businesses grow with branding, websites and marketing that actually works.

Work with me ✦