Skip to main content

No account?

Start free trial

Just a moment

Before you can get stuck in, we need just a little information from you.

Your Name
Receive Marketing (UK)

Your ideas could help us define the State of Allied Health. Take our survey and read the final industry insights report before anyone else. Take the survey.

Start free trial
Latest features
15 Sep, 2026

Understanding how AI Transcribe works in practice

splose AI transcription
Jane Jongen
8 mins to read

Most Healthcare practitioners have a complicated relationship with practice notes. They are an invaluable part of client care but they also take between 30 to 60 minutes a day to write. Add those minutes up and you’re looking at at least 3 hours a week. Go a step further and the time snowballs into nearly four weeks a year. 

Worse still, all this is the work that happens after you’re done working. And it’s non-billable. 

This is where AI Transcribe & Dictate from splose can be a game changer. Not only can it automate work that’s considered immutably manual, but it creates a new workflow that feels intuitive and instantly useful. 

Getting started with splose AI Transcribe

With splose AI, you don’t have to spend hours getting set up, or have to spend days learning how to prompt the system. To get started, all you have to do is record your session. splose AI then transcribes what was said and produces a structured progress note from it. There are two ways to record:

In-session recording: Open the progress note before or during the session and start recording. The microphone picks up the audio during the session. When you finish, AI processes the transcript and drafts the note.

Post-session dictation: After the session, dictate your clinical observations into splose in the same way you'd leave a voice note for yourself. AI transcribes it and builds the note from your audio dictation. 

The choice of using in-session recording vs post-session dictation comes down to a practitioner's preference and the type of session AI Transcribe is being used for. Manual therapy and hands-on assessments often work better with post-session dictation. This is because it’s easier to narrate what you did without having to pause repeatedly during the session. Talk-based sessions such as counselling, speech, psych tend to work well with live recording because the session itself contains most of what needs to be documented.

What does a splose AI transcribed note look like?

Let's use a real example in which a physiotherapist runs a 30-minute follow-up with a patient managing chronic lower back pain. They've seen this patient five times. Today's session covers exercise progression, a new functional assessment, and updated home program.

The session is recorded. Afterward, AI produces a draft note. Here's roughly what it includes:

splose AI transcription example

The output isn't a word-for-word transcript. It's a structured note that includes observations, assessment, and a plan built from what was said and referenced against the patient's existing history.

In our example, previous notes are referenced automatically (the last three, by default), so AI already knows the patient's baseline, their previous functional scores, and what was agreed at the last session. The practitioner doesn't need to recap any prior information. It simply gets pulled into the note by splose AI.

After reviewing the draft, the physio makes a couple of edits - adjusts one exercise name, adds a specific load note, and changes the home program wording. The total time from the end of the session to seeing a finalised note is about 4 minutes. Before splose AI, this entire process would have taken anywhere between 18–25 minutes.

The questions practitioners ask about AI transcription

"Will patients mind being recorded?"

Most don't but consent always matters. The standard approach is to let clients know at the start of the session that you'll be recording their responses for documentation purposes. Be sure to  ask if they're comfortable with that. You can also reassure clients that splose doesn't store audio recordings after they've been transcribed. The transcript is generated and the audio is discarded. If you’re transparent about the process upfront, the vast majority of clients will agree to be recorded. For clients who'd prefer not to be recorded, you still have the option of post-session dictation that is also augmented by AI. 

"What if the AI gets something wrong?"

AI will occasionally make minor mistakes, but it shouldn't be a cause for concern. Unusual medication names, discipline-specific terminology, and accents can all cause transcription errors. That's why the system is designed to have you review the transcription before its finalised. Nothing goes into the clinical record without your sign-off. Most practitioners find they're fixing a few words per note rather than rewriting anything substantial.

"Does it produce the right kind of note structure?"

If you want a Subjective, Objective, Assessment, and Plan (SOAP) note, a Data, Assessment, and Plan (DAP) note, or a specific format your clinic uses, you need to tell it before it starts transcribing. Adding one line to the prompt at start of each session ("Please structure this as a SOAP note") is usually enough. You can also save that instruction as a default prompt to save time across multiple notes. However, left entirely on its own, AI will still produce a reasonable narrative note. 

"What if I forget to start recording?"

If you forget to start recording before the session, you can always use post-session dictation. All you have to do is open the note, record a 2-minute verbal summary of the session, and AI builds from that. Live recording is more seamless, but post-session dictation is still faster than writing from scratch.

A sample prompt that works for most practitioners

By default, AI Transcribe will produce a clean summary of the session. If you want more structure, you can add the following prompt after the recording:

"I'm a physiotherapist. This is a follow-up session — refer to the last 3 progress notes. Structure the note as SOAP. Include functional goal progress and updated home program."

If you find the output from this prompt useful, you can save it in your AI settings and reuse it for subsequent sessions.

splose AI transcription - prompting

Ready to get started? Here’s a quick setup guide

splose AI is available as a paid add-on to any splose plan. To get started, just follow these simple steps:  

  1. Open a new progress note
  2. Click the AI panel on the right → Click the microphone button
  3. Hit record — either before the session starts, or after for dictation
  4. Click to generate note or write your own prompt
  5. Review the draft note when processing completes
  6. Edit, finalise, done

Your first session with it will present a small learning curve. But by the third or fourth session, the process will feel more natural. 

Need a little more help?

If you still need some additional help getting started with splose AI, we’ve created handy guides to walk you through the process in a bit more detail. 

Add splose AI to your practice

Start your free trial today to see how seamlessly smart automation fits into your current workflow.

Add splose AI

Get rewarded for sharing your best splose templates

If you're in Australia, the splose template search is now open. We're giving away up to $10,000 in cash prizes for the best progress note templates built in splose. Bonus points for templates that use AI prompts.

Other insights of interest

See more insights