What a one-way video interview is
A one-way video interview (also called an asynchronous or on-demand video interview) is a recorded first-round interview. The employer sets the questions. The candidate gets a link, reads or hears each question and records an answer on camera. Nobody is on the other end at that moment. Later, one or more people on the hiring team watch the recordings and decide who moves forward.
It replaces the 20-minute screening call. The questions are the ones a recruiter would have asked anyway, but the call has been turned into a task the candidate does on their own time. The word "one-way" describes the flow of the conversation: it only goes from candidate to employer, with no follow-up questions and no replies.
How it works, step by step
- The employer chooses the questions. Usually five to eight, tied to the job. Good ones ask for specific past situations.
- Candidates are invited. By email or a link in an application flow, usually with a deadline.
- The candidate checks their setup. Camera, microphone and a short practice question, so the first real answer is not the first time they have tried it.
- The candidate records answers. Each question has a time limit. Some tools allow re-recording, and some give one take.
- The team reviews. Reviewers watch, share and rate. Newer tools add transcripts, summaries or scores to cut the watching time.
- People decide. The recordings narrow the list, and live conversations follow for the people who go forward.
Pros and cons
For employers
| Strengths | Weaknesses |
|---|---|
| No scheduling. Candidates interview at any hour, across time zones. | Reviewing video takes time. Fifty candidates times six answers is a lot of watching without transcripts or summaries. |
| Every candidate gets the same questions, so answers can be compared. | No follow-up questions. A vague answer stays vague. |
| The whole team can watch the same recording, not a recruiter's notes. | Reviewers may react to lighting, background and presentation instead of the content, unless the process guards against it. |
| It scales: a hundred first rounds do not need a hundred calls. | Some strong candidates drop out because they dislike recording themselves. |
For candidates
| Strengths | Weaknesses |
|---|---|
| They choose the time, which helps people with jobs, kids or shifts. | Talking to a camera feels unnatural and can be stressful. |
| Everyone gets asked the same thing, so it feels more even-handed than an unplanned chat. | There is no way to ask a clarifying question or read the interviewer's reaction. |
| No travel or video call coordination. | Candidates need a working camera, microphone and connection, and a quiet place. |
How to run a fair one
One-way video is only as fair as the process around it. These habits make the biggest difference:
- Ask the same questions of everyone for the role. This is the core of a structured interview.
- Write down what a good answer looks like before you watch. Rate each answer against that, not against a general feeling.
- Rate content, not presentation. Unless looking polished on camera is part of the job, lighting, backgrounds, accents and style should not move a score.
- Tell candidates what to expect. How long it takes, how many questions, who sees the recording, how long it is kept, and whether AI is involved.
- Offer an alternative. Someone without a camera or with a disability may need another way. Decide in advance who handles requests.
- Keep people in the decision. Use recordings to inform a person, not to reject automatically.
- Check the law where you hire. Some places regulate automated tools or video interviews. Our NYC Local Law 144 guide is one example, and your own counsel can tell you what else applies.
How Infyva's AI interview relates to it
Infyva's standard AI interview is a one-way interview with an AI interviewer, Nova, asking the questions. Here is exactly what a candidate does and what is recorded:
- Nova asks each question aloud. The candidate answers in their own words, at a time that suits them, from the link in their email.
- Video by default, audio if you prefer. By default the browser records each answer on camera. A job can be set to audio only instead. The longest answer defaults to 180 seconds and you can set it between 30 and 600.
- The audio of every answer is transcribed. Scoring uses the transcript. Video is evidence for your team to watch and is never sent to an AI model.
- One take per answer. There is no re-recording, as in a real interview. Candidates get a practice question and a microphone check first, so nerves are not spent on the real thing.
- A notice before anything is recorded. The candidate sees what is recorded, what the checks are and who sees it, and can ask for an accommodation.
- Scores come from what is said. Appearance, facial expressions, accent and tone of voice are not scored. Each answer is scored against the job's competencies with a written explanation. Under three answered questions, the report says there is not enough evidence to recommend anything.
- Integrity notes for a person. Where something deserves a second look, a flag is tied to the answer it happened on. You play it back and decide.
- Recordings are kept for 90 days by default. A daily job deletes them. Candidates can ask for deletion.
This addresses the two biggest weaknesses of one-way video: reviewing time (transcripts, scores and reasons mean you can start from the answers that matter) and reviewer bias toward presentation (scores ignore delivery). It does not fix everything. There are still no follow-up questions in the standard interview. If you need them, a Live AI interview is a booked real-time conversation with follow-ups.
Infyva never rejects anyone. Recruiters review, share a report link that hiring managers open without an account, and decide. Our Spark Hire comparison looks at how a video-first tool differs from this approach, including where it is stronger.
Designing the interview itself
The format matters less than the questions. A few choices decide whether you get useful recordings:
- Number of questions. Five to eight is a common range for a first round. Fewer than that and one weak answer decides too much. More and candidates drop out.
- Answer time. Give enough time for a real example, typically one to three minutes. Very short limits reward quick talkers.
- Question type. Behavioral questions ("tell me about a time...") and situational ones ("what would you do if...") both work. Avoid trivia.
- Practice. A practice question removes the fear of the unknown and lets people fix a bad microphone before it matters.
- Deadline. A window of a few days is usual. A same-day deadline excludes people with jobs.
Questions to ask any vendor
- Is the video analyzed by AI, and for what? If it is, what in the analysis relates to the job?
- Can a reviewer get from any score to the answer behind it?
- How long are recordings kept, and who can delete them?
- What does the candidate see before recording starts?
- What can a candidate do if they cannot use a camera?
- Has an independent bias audit been done, where one is required?
We answer these for Infyva on our AI and fairness page, and we say plainly what we do not have: no SOC 2 report and no independent bias audit yet.
If you are the candidate
A short note for people on the other side of the camera. Test your microphone and sit somewhere quiet. Use the practice question. Answer with one specific example, what you did and what happened. It is fine to pause and think. If you cannot use a camera, ask the employer for another way before you start.
When one-way video is a good fit
- Many applicants for each opening, and not enough recruiter hours.
- Candidates in several time zones, or working shifts.
- Roles where a short, structured first round predicts more than a resume does.
It is a weaker fit when:
- The role depends on live rapport or improvised conversation from the first minute.
- Your pool is small enough that a short call is faster for everyone.
- You cannot offer an alternative for candidates who cannot record.
Reviewing recordings without wasting a week
The cost of one-way video moves from scheduling to reviewing. A few practices keep it manageable. Review answer by answer rather than candidate by candidate, so you compare everyone's reply to question one before moving to question two. Use the same written standard for each question. Let two people rate independently if the role matters, and talk about the gaps. And stop watching once you have enough: if the first ten candidates already give you a strong short list, you do not need to watch the other ninety before calling anyone.
With Infyva, the transcript and the written reasoning sit next to each recording, so a reviewer can read first and play only the answers that matter. A hiring manager gets a link that opens without an account.
What it costs with Infyva
Infyva plans include credits, not seats. An AI interview uses 1 credit, used when a candidate actually answers. Every new company starts with a 14-day Pro trial with 20 credits. Starter is $39 a month for 30 credits and Pro is $149 a month for 150. See pricing or run the demo to take an interview yourself.
Frequently asked questions
What is a one-way video interview?
An interview where the candidate records answers to set questions on their own time, and the employer watches later. Nobody is on the other end while the candidate answers. It is also called an asynchronous or on-demand video interview.
Is Infyva's AI interview a one-way video interview?
The standard interview is one-way. Nova asks each question aloud and the candidate answers in their own words. By default the browser also records each answer on video for your team to watch, and a job can be set to audio only. Scoring uses the transcript, not the video.
Can candidates re-record an answer?
No. Infyva allows one take per answer, as in a real interview. Candidates get a practice question first and a microphone check before the interview starts.
How long are recordings kept?
By default, 90 days after the interview, and a daily job deletes them. Candidates can also ask for deletion. Details are on the data retention page.
Is the video analyzed by AI?
No. Video is evidence for your team to watch and is never sent to an AI model. The scoring model sees the transcript only.