AI for Physios Polishes Notes. Patients Still Quit at Session Three

A bloke named Craig did his calf at touch footy in May.

His physio prescribed six sessions over eight weeks. Craig came to two. He felt better, work got busy, and the reminder about session three arrived the same week his ute needed a new clutch. Nobody at the clinic followed up, because nobody is paid to follow up. The calf went again in September, and Craig booked in with a different clinic closer to work.

That same month, the AI scribe your practice trialled was writing beautiful SOAP notes before each patient reached reception. Your senior physio stopped taking notes home for the first time in years. A real win, on consults that were already booked and paid for.

That is the shape of AI for physios as it is sold today. Almost all of it points at the treatment room: scribes that turn the consult into a note, tools that draft the exercise program. I build these systems for Australian businesses, so let me say the unpopular half out loud. Your clinic does not bleed inside the treatment room. It bleeds between appointments, in the plans that quietly stop.

What AI for physios gets right

Give the scribes their due first, because the case for them is real.

Clinical notes have eaten physio evenings forever, and the documentation load on EPC, DVA, NDIS and workers compensation cases makes it worse. A scribe that drafts the note while you treat gives a clinician back real hours, and in a market where every practice struggles to hire, if it keeps one good physio from burning out, it has paid for itself.

Two things are true about it anyway. The scribe only works on appointments that happen, and the note stays your responsibility, so you read it before you sign it. You have bought speed on work you were already going to be paid for.

The half of AI for physios nobody demos

Now look at the other side of the same clinic.

The six-session plan that stopped at two. The 7am cancellation that left a hole in the diary nobody rang the waitlist to fill. The chronic back patient discharged in March who is due for a review nobody scheduled. The GP who sent you four referrals last year and none since.

That work was always mechanical. It never had anyone spare, so it slipped, and it kept slipping, because none of it feels urgent until the diary has more white space than it did last winter.

A scribe saves minutes on the patients in front of you. The patients who stop at session two take the rest of their plan, and their next injury, somewhere else.

Your practice bleeds between appointments

Open your practice software and run one report: patients with incomplete treatment plans in the last six months. Every clinic owner I have sat with winces at that list, because each name is care you recommended, in writing, that never happened.

The uncomfortable part is that the dropout is rarely a clinical decision. People stop coming when the pain drops below the level that overrides a busy week. They do not ring to cancel the plan. They just let the next booking lapse, and your front desk, flat out with the phone and the counter, lets it go.

Nobody decided to quit your care. They just stopped being asked to continue it.

Start with the plans that stall

Pick one workflow, not the whole practice. For most clinics the highest-value place to start is the one that gets dropped first every busy week: following up incomplete plans.

A system reads the treatment plans and attendance already sitting in your practice software, flags the patients who have fallen behind the plan their physio wrote, and messages each one in your clinic's voice, not a vendor's template. It answers the easy replies, books the yeses into real gaps in the diary, nudges the quiet ones once more, and puts the rest on a short list for your front desk to ring on a slow afternoon. Nothing in that loop makes a clinical judgement. All of it takes persistence no reception team has spare.

A follow-up costs you a text message. A patient who drifts costs you the rest of their plan and every referral they would have sent you.

The second workflow, once the first is proven, is the diary itself: a cancellation triggers an offer to the right waitlisted patients within minutes, and review-due lists actually get worked. That gives you numbers to defend at the end of the quarter: plans completed, holes backfilled, against what you spent. Arithmetic, not a vibe, and I set out how to run that calculation honestly in the AI ROI framework.

Automate the chasing, not the treating

Here is the line that keeps this safe, and Ahpra makes it sharper for you than for most industries.

Diagnosis, treatment and discharge decisions sit with a registered physiotherapist. A tool that drafts an exercise program is offering a starting point for a professional to edit. The moment a system starts telling patients what their injury probably is, or promising outcomes in its messages, you have a professional standards problem wearing a chatbot, and the advertising rules on outcome claims apply to an automated SMS just as they apply to your website.

There is a cleaner place to point the same technology. Drafting the reminder, reading the reply, moving the booking, writing the contact onto the file: that is admin. It was always admin. It is the work your team hates and the part a machine does without sighing.

Point the software at the follow-up nobody was doing, not at the judgement you are registered to make.

What AI for physios costs

Work down this list in order and stop the moment something works.

• The tools already inside your practice software. Recall lists, waitlists, online bookings, SMS reminders. Often half-configured. Turn them on before you spend a dollar, because a build that duplicates them is money set on fire.
• A scribe subscription, priced per clinician each month. Cheap enough to trial with one physio and judge on your own numbers.
• A custom build that connects plan follow-ups, replies and bookings to your actual practice software. Usually a few thousand up to the mid teens of thousands of dollars to go live, and what moves you up that range is how many systems it writes back into.

What catches owners out is the running cost, not the build: the subscriptions underneath, the model usage, and the front desk time spent on whatever the system escalates. I broke those layers down in what AI actually costs a small business. A vendor whose every answer is their priciest option is solving for their invoice.

The Australian layer: your practice software and patient data

Two things separate a system that works in an Australian clinic from an overseas template.

The first is integration. You run on Cliniko, Nookal, Zanda or Halaxy, and that database is the business. A tool that cannot read your real treatment plans and write bookings back has not removed the double handling, it has moved it to your admin team. Ask what it writes back, and where, before the demo wins you over.

The second is data. Patient files are health information, the most protected class under the Privacy Act, before you add Medicare numbers and DVA details. Send them offshore and the cross-border rules keep you accountable for what the overseas recipient does with them. Ask where the data is stored, whether it trains someone else's model, and whether you can delete a record on request.

Common questions about AI for physios

What should a physio clinic automate first?

Incomplete treatment plan follow-ups, then cancellation backfill. Not diagnosis, not triage. Prove one workflow before you widen, because switching on five things at once means you cannot tell which one paid.

Will AI replace physiotherapists?

No. The assessment, the hands and the registration are the business. The software clears the chasing and the paperwork so the clinicians you can hire spend their hours treating.

Is a custom build worth it over what my practice software gives me?

Only after you have switched on what you already pay for and hit its limits, and only if you own the system and the data at the end.

Go back to Craig. The clinic that has him now did not treat calves better than you. It was simply in front of him the week his calf went again, because you never asked him to finish the plan you wrote.

If you want a straight read on how much of next quarter's revenue is sitting in your own incomplete plans list, that is what a first call is for. Book a strategy call and bring that report and last month's unfilled cancellations. We will find the patients you are quietly losing before we talk about building anything.