Expect 3 to 8 questions, roughly 30 seconds of think time once each prompt appears, and 2 to 3 minutes of recording per answer. Retakes exist only if the employer switched them on, and plenty do not. Every one of those numbers is a dial the employer turned before sending the link, which is why the practice question matters more than anything else on screen. It shows you your actual configuration before a single scored second is recorded.
The six settings that define your interview
A one-way interview is not really one product. It is a template with half a dozen dials. Nothing below is a rule from the platform, it is the range candidates commonly report, plus the place you personally confirm the value.
| Setting | Common range | Where you confirm it |
|---|---|---|
| Number of questions | 3 to 8, five is a typical screen | Invitation email, then the instructions screen before question one |
| Think time after the prompt appears | About 30 seconds, sometimes 20, occasionally none | The countdown on the question screen, and the practice run |
| Recording limit per answer | 2 to 3 minutes, some questions capped at 60 or 90 seconds | Displayed beside the record button, per question, not once at the start |
| Retakes | Zero, one or two. Frequently zero | The recording screen shows attempts remaining. The practice question shows you the policy first |
| Whether recording auto-starts | Usually automatic when think time expires, sometimes a manual click | The practice question. This is the most common surprise of the whole flow |
| Deadline to complete | Days, not hours, in most invitations | The invitation email. It covers the whole interview, there is no per-question deadline |
Two things follow. Advice built on someone else's HireVue is close to worthless, because the same employer can configure a graduate role and a senior role differently. And the think time and the answer limit are the only two numbers you need internalised before you sit down, because everything about how you speak is downstream of them.
One question, second by second
Here is the shape of a single question when the limits are 30 seconds of think time and 2 minutes of recording. Rehearse against this clock, not against a generic answer structure.
- Prompt appears, 0:30 remaining. The question shows as text, usually alongside a short recorded clip of someone reading it aloud. Read the text, not the video. The text is the part you can re-read.
- 0:30 to 0:20. Pick the story. One decision, made fast. The second-best story told well beats the ideal story you are still choosing at second five of the recording.
- 0:20 to 0:05. Say your first sentence in your head. Not the whole answer, the opening line only. That is the one that always comes out mangled.
- 0:05 to 0:00. Sit back, put your hands somewhere they will stay, look at the lens. Recording may begin on its own.
- 0:00 to 0:15 of recording. One sentence of context. The company, your role in it, the problem. No warm-up phrases, no restating the question back.
- 0:15 to 0:40. The situation and the constraint that made it hard. The constraint is the detail reviewers remember.
- 0:40 to 1:25. What you specifically did. Named tools, named decisions, first person singular. This is the bulk of the answer and the part where most people go vague.
- 1:25 to 1:45. The result, with a number if one honestly exists, plus one line on what changed afterwards.
- 1:45 to 2:00. Silence you never use. Stop early and on purpose.
That last row is the one worth changing your behaviour over. There is no grace period when the timer expires. The upload contains whatever was captured, mid-word if that is where you were, and a reviewer cannot tell the difference between an answer that ran long and a candidate who had nothing to land on.
The practice question is a settings readout, not a mic check
Most people treat it as a webcam test and click through in fifteen seconds. It is the only place you learn the rules before they cost you anything.
Read the numbers before you touch anything
Think time, recording limit, attempts remaining. Write all three on paper next to you. If the practice screen shows retakes and the instructions screen said nothing about them, believe the practice screen.
Watch what happens when the countdown hits zero
Does it start recording by itself, or wait for a click? That is the difference between a composed opening and eight seconds of you looking startled. Find out on the question that does not count.
Record the full length, at real volume
Not a ten second hello. Talk for the entire limit, so you learn what two minutes feels like from the inside and so the audio level you play back is the level the reviewer will hear.
Play it back once and fix exactly one thing
Usually the light behind you, the laptop sitting below your chin, or a microphone picking up the room. Fix the worst one. Do not restage the entire room and lose your nerve doing it.
What thirty seconds is actually for
Thirty seconds is not enough to plan an answer. It is enough to make one decision and rehearse one sentence, which is why the real preparation happens days earlier. People who freeze in that window are almost always composing from scratch. People who use it well are choosing between three stories they already had ready.
That is the honest shape of this format. A live interviewer follows up, rephrases, and rescues a slow start. A one-way recording does none of it. No clarifying question, no reading the room, no recovering in the second half because someone nodded at the right moment. You get a prompt, a countdown, and your own voice in a quiet room. Preparation is not one input among several here, it is the entire mechanism.
The prompts are more predictable than the format feels
Because these clips are recorded and compared across candidates, employers keep the question set stable and fairly neutral. Expect a mix of four kinds, roughly in this order of likelihood.
- Motivation. Why this company, why this role, why now. Nearly always present, often question one, and the one most people answer with something they could say to any employer on earth.
- Behavioural. A conflict, a failure, a deadline, a time you influenced without authority. Familiar stems, told inside a fixed two minutes rather than a conversation.
- Role specific. A short technical or domain question answered out loud instead of typed. Judged on whether you can explain, not whether you can implement.
- Situational. A hypothetical about how you would handle something. Answer with a real precedent first, then the hypothetical.
If the employer wants coding, it arrives as a separate link with its own editor and its own rules. It is not part of the recorded video flow, so do not try to prepare for both in one sitting.
Where OfferPilot AI is the wrong tool
Worth saying plainly, because people often arrive here from a different question than the one they typed. A recorded one-way interview is a poor fit for a live copilot, and not for policy reasons.
- There is no interviewer audio. Interview Copilot works by listening to an interviewer's voice arriving through system playback. Here there is a text prompt, a countdown, then you talking alone. Nothing for it to listen to.
- The camera is a channel our overlay does not touch. The overlay is excluded from screen capture at the operating system level, so it stays out of Zoom, Teams, Meet, OBS and screenshots. A webcam pointed at your face is a different channel entirely. That exclusion does nothing about it, and nothing about lockdown browsers, an invigilator in the room, eye tracking, or a phone propped up as a second camera.
- Reading is legible on camera. A human watches these clips. Eyes tracking across text look different from eyes thinking, and in a medium built entirely from your face there is nowhere to put that. Six bullet points taped just under the lens are fine. A script on a second monitor is not.
- Plenty of people record on a phone. We run on Windows 10 and 11 only. No macOS, no Linux, no mobile, no browser extension.
The useful window is earlier, in the days before the link expires. Because these question sets repeat across a company's funnel, running the company question bank generator against the specific firm and role is worth an hour, so you arrive with a two minute version of the dozen prompts most likely to appear. That is preparation, which is the only thing this format rewards.
The short version
Three to eight questions. Around thirty seconds to think, two to three minutes to record, retakes only if they were enabled, and no grace period when the clock ends. The practice question tells you which configuration you are facing, so work it properly instead of clicking past it. Then build every answer to finish fifteen seconds early, and put the result before the flourish.