A patient gets your appointment reminder, thumbs out "yep see you then" or "can't make it Tuesday", and puts the phone down. What happens next decides whether your reminder was a one-way broadcast or a working part of your front desk. If every reply lands in a staff inbox to be read and actioned by hand, the reminder has only moved the work around.
This guide walks through exactly how DataFlows handles patient replies for Cliniko clinics. You will see how confirmations are recorded in Cliniko without staff involvement, how cancellations and reschedule requests reach the right person, how free-text replies are classified, and what the audit log keeps. One point up front: this feature classifies replies. It does not write replies and it is not a chatbot.
What Is Cliniko SMS Reply Handling?
Cliniko SMS reply handling is the process of reading a patient's text reply to an appointment reminder, working out what they mean, and taking the right action in response. In the DataFlows and Cliniko integration, a reply is sorted into one of a small set of intents: confirm, cancel or reschedule, with anything unclear treated as a question or low-confidence reply. Simple one-word replies are matched by fast keywords. Free-text replies such as "can't make it Tuesday" are classified by AI, which returns an intent and a confidence score. Clean confirmations are marked as confirmed in Cliniko automatically. Anything that needs a human decision, such as a cancellation, a reschedule request, a genuine question or a low-confidence reply, is passed to staff with the patient's name attached. Every reply is logged with its intent, confidence score and the action taken, so the clinic can always see what happened and why.
Why It Matters for Australian Clinics
Most Australian clinics run lean front desks. Reception answers phones, checks patients in, handles Medicare and private health claims, and manages the diary at the same time. Reading and sorting every SMS reply by hand competes with all of that.
Replies also arrive at awkward times. A patient may answer at 9pm the night before, or at 7am on the day. If a cancellation sits unread until mid-morning, the slot is harder to refill. Fast, consistent handling of replies protects the diary.
There is also a trust issue. Patients expect a clear response when they text a clinic. A reply that disappears into silence, or a confirmation that never shows in the diary, creates doubt. Reply handling closes that loop.
Finally, health information is sensitive. A clear record of what was sent, what came back and what was done about it is good practice for any clinic. If you are reviewing your consent approach, see our guide to Cliniko SMS consent workflows.
Key Benefits
- Confirmations need no staff time. A clean confirmation is recorded in Cliniko automatically, so reception only sees the replies that need a person.
- Free-text replies are understood. Patients do not have to remember a keyword. "Can't make it Tuesday" is read for what it means, not for an exact match.
- Staff alerts carry the patient's name. Every escalation by email and/or SMS names the patient, so staff can go straight to the right appointment.
- Uncertain replies go to a human. Low-confidence replies and genuine questions are escalated rather than guessed at.
- Edge cases are handled calmly. A reply to a booking confirmation, or a second reply to something already resolved, gets a clear response instead of a false alarm.
- Everything is logged. Each reply is stored with intent, confidence score and action taken, which makes the process easy to review.
How a Patient Reply Is Handled, Step by Step
Here is the path a reply takes, from the moment the patient hits send.
Step 1: The patient replies to the reminder
Reminders are sent from your Cliniko appointment calendar on the timing you choose. The patient simply replies to that message from their phone. Reply handling needs a two-way channel, so your clinic should send from a dedicated virtual number. A registered alpha Sender ID shows your clinic name to patients, but it is one-way only and cannot receive replies.
Step 2: Keyword matching reads the simple replies
The first pass is fast keyword matching. Short, obvious replies such as a confirmation word, a cancel word or a reschedule word are recognised straight away.
Step 3: AI classification reads the free text
Not every patient follows instructions. Some write full sentences, add context, or reply in their own words. For these free-text replies, AI classification works out which intent the patient most likely means and attaches a confidence score. A message like "can't make it Tuesday" is read as a cancellation signal, even though it contains no keyword.
This is the only job the AI does here. It sorts the reply into an intent. It does not write a response, hold a conversation or decide what the clinic should do about a booking.
Step 4: Clean confirmations are confirmed in Cliniko
When a reply is a clear confirmation, DataFlows confirms the appointment in Cliniko automatically. No staff member needs to open the message, find the patient or click anything. If you have turned on the option to log confirmations onto the appointment, a timestamped note is added to the Cliniko appointment as well.
Step 5: Anything needing a person goes to staff
Some replies should never be handled automatically. A cancellation changes the diary and may open a slot to fill. A reschedule request needs someone to find a new time. A genuine question needs an answer. A low-confidence classification means the system is not sure, and a person should look.
These are escalated to staff by email and/or SMS. Every alert carries the patient's name, so the person receiving it knows who to look up. Staff can also step into the conversation from the patient inbox, which shows each SMS thread on the clinic number with the patient name resolved automatically.
Step 6: The reply is logged
Whatever the outcome, the reply is written to the log with the intent that was detected, the confidence score and the action that was taken. We cover the log in more detail below.
The Two Edge Cases
Real inboxes are messy. Two situations come up often enough that DataFlows handles them specifically.
A reply to the booking confirmation. When a new appointment is booked, a booking confirmation SMS is sent straight away. Patients sometimes reply to that message with "thanks" or "great". There is nothing for staff to do, and no reminder to confirm yet. Instead of raising a false alert, the patient receives a short response saying the clinic will text them closer to the date.
A duplicate reply. Patients often send a second text, such as "yes" followed by "yes confirmed" a minute later. If the first reply has already been resolved, the second one gets an "already taken care of" acknowledgement. Staff are not alerted twice, and the patient knows it is done.
Both of these are small details, but they are what keeps reply handling from generating noise for your team.
The Reply Audit Log
Every reply is logged with three pieces of information: the intent that was detected, the confidence score, and the action taken. That means you can answer questions such as "why did this patient get an alert?" or "was this confirmation automatic?" without guessing.
The log supports routine checks, such as spotting patterns in low-confidence replies. It also supports a full two-sided CSV export, covering both the messages sent and the replies received, which is useful for internal audits and for reviewing how the process is working across a branch.
How DataFlows Helps
The reply handling described above is part of the native DataFlows and Cliniko integration. DataFlows connects directly to Cliniko using a Cliniko API key, found in Cliniko under My Info and then Manage API keys. It polls Cliniko's API for appointment data, so there is no Zapier route or webhook set-up to build or maintain.
Around reply handling, the same integration covers the rest of the patient communication cycle:
- Smart reminders with rolling or fixed timing, multiple reminder slots and per-appointment-type message overrides.
- Booking confirmation, cancellation and reschedule notices sent when those events happen in Cliniko.
- Waitlist SMS that offers a cancelled slot to waiting patients, with staff finalising the booking manually in Cliniko.
- Patient inbox so staff can step in alongside the automation.
- Consent-aware messaging that respects each patient's Cliniko-recorded reminder consent setting.
- Multi-branch settings so each business location can have its own timing, sender number and templates.
DataFlows is pay-as-you-go with no monthly or annual fees. You buy SMS credits and top them up as needed. For the full feature list, read our complete guide to Cliniko SMS features, or start with how Cliniko SMS reminders work with DataFlows.
Best Practices
- Use a dedicated virtual number. Two-way reply handling needs a number that can receive texts. Keep alpha Sender IDs for one-way messages only.
- Tell patients how to reply. Include a short line in the reminder telling patients how to confirm, so more of them use a simple keyword. The AI covers the rest.
- Set up staff alerts before going live. Decide who receives escalations by email and/or SMS at each location, and make sure those contact details are current.
- Review the log in the first weeks. Check intent, confidence and action taken on a sample of replies. This shows you how your patients actually write.
- Treat low-confidence replies as a prompt. If you see many, look at your reminder wording. Clearer instructions usually produce clearer replies.
- Edit your templates. Every message template is editable with merge tags such as {patient_name}, {date}, {time}, {practitioner} and {business}, so wording can match your clinic's voice.
- Test with your own phone. Book a test appointment, then try a keyword reply, a free-text reply and a duplicate reply to see each path.
Conclusion
Reply handling turns an SMS reminder from a broadcast into a process. Simple confirmations are recorded in Cliniko without anyone touching them. Cancellations, reschedule requests and questions reach staff with the patient's name attached. The edge cases are covered, and every reply is logged with intent, confidence and action taken.
It is a classifier working inside a defined workflow, not a chatbot. That keeps decisions about bookings with your team, where they belong.
Ready to try it? Sign up at dataflows.com.au, connect your Cliniko account with an API key, and start with reminders on a dedicated virtual number. You can then roll out reply handling one branch at a time.