Linguaskill Speaking Parts 1 and 2: Who You Are Talking To

Most Linguaskill Speaking guides tell you the parts and their timings. Very few tell you the thing that changes how you should actually sound, and it is written in plain sight on Cambridge's own task description.

Your audience changes between Part 1 and Part 2, and the test never announces it.

SpeakShark is free if you want to practise talking to a person rather than at a page, three AI conversation sessions a day with no card.

In this guide: the audience switch · Part 1 in detail · the three bullet points · Part 2.1 · Part 2.2 · the warm-up trap · how to practise · method · FAQ

Key takeaways

Part 1 is a presentation with no listener. Part 2 is a report to a named person who missed a meeting. Same test, two registers.
The three bullet points in Part 1 are not decoration. They are the structure you are expected to use for the full minute.
Part 2.1 gives you forty seconds for four key points, which is ten seconds each including your framing words.
Part 2.2 is answered by the same colleague or classmate, who asks three follow up questions about the same topic.
The warm up is three questions and unmarked, which makes it the only free chance to catch a bad microphone.
The whole module is sixteen minutes across five sections, delivered at a computer and marked by humans and machine together.

The audience switch nobody mentions

Read Cambridge's own wording for the two parts side by side.

	Cambridge's description	Who you are speaking to
Part 1	"The candidate is given a topic, with three bullet points, to talk about for 1 minute"	Nobody. No listener is named
Part 2.1	"summarise the four key points for a colleague or classmate who was unable to attend"	A specific person, who was absent
Part 2.2	"three follow-up questions from the same colleague or classmate"	The same person, now asking

Part 1 has no addressee at all. Part 2 has a named one, with a stated situation: they were not there, so they do not know what happened.

That difference is not cosmetic. A summary delivered to somebody who missed the meeting has a natural shape that a presentation does not. You orient them first, then compress, then flag what matters to them. A candidate who delivers Part 2.1 in the same neutral presenting voice as Part 1 is answering a different question from the one asked.

This is the kind of instruction that hides in a subordinate clause. "For a colleague or classmate who was unable to attend" reads like scene setting. It is the task.

Part 1: a presentation to nobody

Cambridge calls the section Presentation. You get a topic and three bullet points, and you speak for one minute.

There is no examiner in the room. The module is taken at a computer with a microphone and headphones, the prompts arrive through the screen and the headphones, and your answers are recorded. Nobody will nod, prompt you, or rescue a stalled sentence.

One minute is longer than it sounds when nothing comes back at you. It is roughly 130 to 160 words at a natural pace. Candidates who have only practised conversational turns tend to run dry at about forty seconds, because a conversation normally hands the floor back by then and here it never does.

If you have prepared for exams where a person sits opposite you, this is a different physical task. Trinity GESE and IELTS speaking put a human in front of you. Versant and Linguaskill do not.

The three bullet points are the mark scheme

The most common wasted opportunity in Part 1 is treating the bullet points as suggestions.

They are the only structure the task gives you, and they map cleanly onto a minute: roughly twenty seconds each with a sentence of opening and a sentence of closing. A candidate who covers all three has produced a complete answer by definition. A candidate who talks fluently about one of them for a minute has produced an incomplete one, however good the language.

Practically:

Say something about each bullet, in order, even if one of them is thin.
Do not spend the first fifteen seconds restating the topic. The examiner already has it.
Leave yourself a closing sentence. Trailing off mid clause when the timer stops is avoidable.

This is the same discipline that OPIc and the TOEIC speaking test reward, and it is learnable in a week of deliberate practice.

Part 2.1: forty seconds and four points

You hear an audio extract of a meeting or a tutorial. Four key points are given in it. You then have forty seconds to summarise those four points for the person who was not there.

Do the arithmetic, because it changes the strategy. Forty seconds for four points is ten seconds each, and that ten seconds has to include whatever words you spend introducing the summary and moving between points. There is no version of this task where you retell the extract.

What survives at that compression is a list, delivered in the order that helps the listener, with almost no connective padding. What dies is narration: "so in the meeting they were discussing, and then someone mentioned that maybe" has already spent a quarter of your time.

Two practical consequences while you listen:

Count to four. The task tells you how many points there are, so you know when you have them all and when you have missed one.
Note the point, not the sentence. You are being asked what was decided or stated, not how it was phrased.

Cambridge publishes a sample Speaking test and a separate Business Speaking sample test, and this is the part most worth doing under a real timer rather than reading about. The same compression problem shows up in CELPIP speaking tasks and in the read-aloud timings we measured for China's CET oral test. Talking daily is what makes forty seconds feel workable, and our free sessions exist for exactly that kind of repetition.

Part 2.2: the same person asks you three things

Part 2.2 keeps the same colleague or classmate and the same topic. They ask three follow up questions and you give your opinions.

The continuity is the point. You have just told this person what they missed; now they are reacting to it. Opinions here can and should refer back to what you summarised, because that is what a real person would do. Answering as though the three questions arrived from nowhere throws away the only piece of context the task has spent two sections building.

It is also a register shift in the other direction: Part 2.1 is compressed and factual, Part 2.2 is your view. Candidates who stay in clipped summary mode sound evasive, and candidates who stay in opinion mode during 2.1 run out of seconds.

The warm-up is unmarked and still matters

The test opens with three simple questions that Cambridge states are not marked.

Treat them as free, because they are, but not as worthless. They are the only moment before scoring starts when you hear your own voice through the setup. If you are too far from the microphone, if the headphones are leaking, if you are speaking so quietly that the recording will be thin, this is where you find out.

A candidate who discovers a bad audio setup in Part 1 has already lost the marked minute they discovered it in.

How to practise Parts 1 and 2

Three drills, in the order they pay off.

The unanswered minute. Pick any topic, write three bullet points, and speak for sixty seconds with nobody responding. Record it. The first three attempts will end early.
Four points in forty seconds. Take any short talk, extract four points, and summarise them to a timer. Cut until it fits. This is the single most transferable skill in the test.
Switch audience deliberately. Say the same content twice, once as a presentation and once as a report to a person who missed it. Hearing the difference in your own voice is what makes the switch automatic in the exam.

The third one needs a listener to be worth anything, which is the awkward part of self study. Our free plan gives three full conversation sessions a day with no card, and the pricing page sets out what the paid tier adds. For the wider picture of what these tools can and cannot do, we keep an honest comparison of speaking apps and a plain how it works page.

What we could not verify

Cambridge publishes timings for Part 1 and Part 2.1 and does not publish per part timings for Part 2.2, Part 3 or Part 4 on the task types page. We have not inferred them, and neither should you: the sixteen minute total covers listening, reading and preparation as well as speech.

We also make no claim about how the hybrid marking splits between the human examiner and the auto marking system, or which criteria each contributes to. Cambridge states that both are involved and does not publish the division.

How we researched this guide

Every description of a part is quoted or closely paraphrased from Cambridge English's own exam formats and task types page for Linguaskill, opened directly rather than summarised from a preparation site. Where Cambridge gives a number we use their number, and where they do not we say so above rather than filling the gap.

We are a speaking improvement tool. We are not an exam preparation provider, and we are not affiliated with Cambridge English or any exam board. For test format, booking and official practice material, go to the exam body directly. Use SpeakShark to make your spoken English stronger, and use official material to learn the test.

Site: speakshark.com
Support: speaksharksupport@gmail.com
Free tier: 3 AI conversation sessions per day, no card

Sources

FAQ

What happens in Linguaskill Speaking Part 1?

You are given a topic with three bullet points and you talk about it for one minute. Cambridge calls this section Presentation. There is no listener in the task, no examiner asking follow up questions and no partner, because the whole module is taken at a computer with a microphone and headphones while your answers are recorded. The three bullet points are the structure you are expected to use, so a candidate who ignores them and free associates for sixty seconds has thrown away the only scaffolding the task gives.

How long do you get in Linguaskill Speaking Part 2.1?

Forty seconds. You listen to an audio extract of a meeting or a tutorial in which four key points are given, and you then have forty seconds to summarise those four points for a colleague or classmate who could not attend. That works out at ten seconds per point including the words you spend introducing them, which is why this task rewards compression over completeness and punishes anyone who tries to retell the extract in order.

Is there a person in Linguaskill Speaking or is it all computer based?

The delivery is entirely computer based. Cambridge states that the module is taken using a computer with a microphone and headphones, that questions arrive through the screen and headphones, and that responses are recorded. The marking is a different matter: Cambridge describes it as hybrid marking, carried out by human examiners together with auto marking technology. So no person is present while you speak, but a person is involved in the score.

Are the Linguaskill Speaking warm-up questions marked?

No. Cambridge states that the test begins with three simple questions intended to help candidates warm up and that these are not marked. They are still worth taking seriously for a practical reason: they are the moment to discover that your microphone level is wrong or that you are speaking too far from it. Candidates who treat the warm up as dead air often carry a bad recording setup into Part 1, which is marked.

How many sections does Linguaskill Speaking have and how long is it?

Five sections in sixteen minutes. They are Part 1 Presentation, Part 2.1 Summary, Part 2.2 Discussion questions, Part 3 Recommendation, and Part 4 Argument presentation and evaluation, preceded by the unmarked warm up. Sixteen minutes covers everything including listening to prompts, reading materials on screen and preparation, so the amount of time you spend actually speaking is a fraction of the sixteen.

Does Linguaskill Business Speaking have the same structure?

The structure is the same and the subject matter is not. Cambridge describes Linguaskill as testing language used in daily life with topics such as studying and working, future plans, travel and technology, while Linguaskill Business tests English used in a business and corporate setting with topics such as buying and selling products or services, the office, business travel and human resources. The tasks you practise transfer between the two; the vocabulary you prepare does not.