A real OET Speaking role-play, start to finish
Almost every OET Speaking sample you can find is a written transcript. A transcript hides the two things candidates most need: the pace, and how much of the five minutes an unhurried opening actually costs. So here is one you can listen to — a GP and a 72-year-old who does not want the test she is recommending, recorded end to end.
The doctor’s turns were written and synthesised. The patient was not written at all — he is the same AI patient the app runs, improvising from a card in the role-play bank. Nobody scripted him, which is why he raises his friend Arthur unprompted and only softens once the doctor names his concern back to him.
Transcript provenance, since this is a page about accuracy: the doctor’s lines are the exact text that was synthesised, so they are authoritative. Mr Wilson’s are machine transcription of what he actually said, reproduced verbatim — they have not been tidied, because a tidied line would no longer match the audio.
What the doctor was holding
In the exam you get three minutes with a card like this and never see the patient’s. Read the tasks, then listen again and watch where each one gets achieved — none of them are announced.
Setting: Local medical clinic. Your role: Doctor.
Mr George Wilson, 72, a retired mechanic, has had recurrent non-exertional chest discomfort for two months. Initial tests — ECG, bloods including troponin, chest X-ray — were all normal. You recommended a stress echocardiogram and he seems hesitant. This consultation is to address that and plan next steps.
- Explore his specific reasons for hesitating, given the normal results.
- Validate his concerns about more testing without dismissing his symptoms.
- Share why further investigation matters in his case — his age, and the nature of his symptoms.
- Offer an adjusted option that respects his values or practical constraints.
- Agree one concrete step he is willing to take this week.
Task five is the one candidates most often lose. It is easy to explain well for four and a half minutes and finish without anything actually being agreed — and “we’ll see how you get on” is not a concrete step. In the recording it arrives at 4:04, and the patient repeats it back.
What the assessors are listening for
Two assessors score every role-play against nine criteria. Four are about your English. Five are about how you conduct a consultation — and those five are where clinically excellent candidates most often lose marks, because they are not habits anyone is taught to perform out loud.
Linguistic criteria (band 0–6)
Clinical communication criteria (band 0–3)
Your overall result for Speaking is a Grade A–E, not a band — bands are the per-criterion sub-scores. How the two relate.
Where the time actually goes
The role-play is short enough that structure is most of the score. Timings below are from the recording, so they are one real example rather than a rule — but the shape generalises.
- 0:00 – 0:50Open, and hand the floor over
Greeting, name, and one open question. Fifty seconds feels expensive when you have five minutes; it is the cheapest fifty seconds in the exam, because everything after it is a response to something real.
- 0:50 – 2:20Find the actual objection, then name it back
This is Understanding the Patient’s Perspective being earned rather than claimed. Mr Wilson’s stated reason is cost; his real one is his friend Arthur. Naming Arthur is what changes the conversation.
- 2:20 – 3:40Explain — including what the test cannot do
Information Giving rewards chunking and honesty over completeness. Conceding the limits of the investigation buys more credibility than another reason to have it.
- 3:40 – 4:36Offer a choice, check, agree, safety-net
Two options rather than one, an explicit check that the explanation landed, one concrete step, and what to do if things change before then. The safety-net is frequently the last thing said and frequently the thing that is missing.
How long is the OET Speaking sub-test?
About 20 minutes in total, of which the two role-plays are five minutes each. You get three minutes to read each card beforehand. The recording on this page runs 4 minutes 36 seconds, which is what a full five-minute role-play sounds like once the greeting and the closing plan are included.
What is an OET Speaking role-play card?
A one-page brief. Yours sets the scene and lists the tasks you must cover; the interlocutor holds a different card telling them who they are and what they will resist. You never see theirs, which is why the tasks have to be achieved in conversation rather than recited.
How many criteria is OET Speaking scored on?
Nine. Four linguistic criteria — Intelligibility, Fluency, Appropriateness of Language, and Resources of Grammar and Expression — and five clinical communication criteria: Relationship Building, Understanding and Incorporating the Patient’s Perspective, Providing Structure, Information Gathering and Information Giving.
Is the patient in this recording a real person?
No, and nor is the doctor. The doctor’s turns were written and synthesised. The patient was not written at all: it is the same AI patient the app runs, improvising from a case card in the role-play bank, which is why he brings up his friend Arthur unprompted and only softens once his concern has been named back to him.
More on the sub-test itself: the OET Speaking sub-test explained, how the three preparation minutes are best spent, the mistakes that cost the most marks, and what sitting Speaking at home requires. A second recorded role-play, with a nurse in an emergency department, is annotated move by move here.
You cannot practise this by reading. Say it out loud.
The patient in that recording is the one in the app. Run a five-minute role-play against him, and get all nine criteria scored about a minute after you stop talking.
On Android, Windows or a Mac? The app is iPhone-only for now — but you can hear a full role-play, scored turn by turn, right here.
Listen to a scored role-play
iPhone · iOS 16 or later. Android is not out yet.
Not installing today?Want the 60 minutes waiting when you do get an iPhone? Leave your email — 60 free practice minutes (≈ 12 role-plays) go on the account you create.