What Is an Asynchronous Interview? How to Choose Text, Audio, or Video

An asynchronous interview is a screening step where the candidate answers preset questions on their own schedule and a recruiter reviews those answers later. "Asynchronous" describes timing, not medium, so the answers can be written, spoken, or recorded on video. Most teams blur the two ideas, default to video because the software offers it, and end up judging lighting and living rooms when they meant to judge answers.

What an asynchronous interview is and where it fits

The asynchronous interview meaning is simple: the candidate and the recruiter never need to be available at the same moment. You'll also see it called a one-way interview or an on-demand interview. The asynchronous video interview is the version most people picture, but it's one subtype among several.

There are three common response formats:

• Written answers to open questions, submitted through a form or the applicant tracking system
• Audio answers, recorded in a browser, given over a phone line, or collected by an AI voice interviewer that asks structured questions out loud
• Video answers, recorded on a webcam or phone camera, usually with a time limit per question

That means you're making two separate decisions. First, live or async? That's a scheduling and capacity question. Second, text, audio, or video? That's an evidence question, and the job should answer it rather than your tool's feature list.

In the funnel, an asynchronous interview usually sits between minimum-qualification review and a live structured interview. Knockout items like work authorization, required licenses, and shift availability come first. The async step replaces the first recruiter phone screen, not the whole interview process. If your real bottleneck is booking dozens of 15-minute screening calls, async timing fixes that. The medium is a separate choice.

The weak version sends every applicant a six-question video interview before anyone has checked whether they can work the shift. The strong version filters on hard requirements first, sends a short async screen to the people who clear them, and books a human conversation for those who pass.

Text, audio, video, or live: a format decision table

Here's the rule worth putting on the wall: use the least intrusive format that captures the job-relevant evidence. Move to a live conversation when you need follow-up questions or when you need to sell the role to the candidate.

Start with what each format gives you.

Then weigh what each format costs.

A worked example. For a warehouse picker, "Can you work 6 a.m. starts, including Saturdays?" is a text question. Nobody needs to record a video to say yes. "Tell me about a time you noticed a safety problem at work and what you did" is different. A form usually gets you two flat sentences, while a spoken answer gives you the sequence, the reasoning, and whether they told anyone.

That's why audio is often the better default for first screens. It collects richer evidence than a form without putting the candidate's face or home in front of the reviewer. If you want the mechanics, here's how an AI voice interview works compared with recorded video.

Audio does not remove bias, though. It removes visual inputs and leaves others in place: accent, speech rhythm, a noisy line, or a speech difference related to disability. The fix is the same for every format. Score the content of the answer against written anchors, and ignore the delivery.

Video makes sense when the visual is the evidence: a sales rep who will run client demos on camera, a tutor who teaches over video, a spokesperson. In those cases one-way video interviews test something the job requires. For a cashier or a forklift operator, they mostly test the candidate's webcam.

If you already pay for asynchronous video interview software, check whether it lets you turn the camera off for some questions. If it doesn't, treat that as a gap in the tool rather than a reason to make video the default.

Live stays the right call for senior hires, candidates with competing offers, and roles where the follow-up question is the real test, such as consulting or sales management.

How the asynchronous interview process works

A good async screen is built backward from the job. Here's the sequence that holds up in practice.

1. Define what the job needs. Pick 3-5 criteria from the actual work, such as reliability for early starts, safety awareness, or handling an upset customer. If a criterion can't be tied to the job, drop it.
2. Write questions and scoring anchors together. Each question maps to one criterion. Describe what a weak, acceptable, and strong answer sounds like before any invitation goes out, so reviewers aren't inventing the standard while they listen.
3. Choose the format per question. Mixing is fine and often best: availability and certifications in text, two situational questions in audio.
4. Send a clear invitation. Say what the step is, how long it takes, who reviews it, whether AI is involved, when it's due, and how to request another format.
5. Let candidates respond. Include one practice question that isn't scored, and decide your retake policy in advance.
6. Review against the anchors. Score each answer separately and write a one-line evidence note that quotes or paraphrases the candidate. "Seemed confident" is not evidence. "Described stopping work and reporting a loose pallet to the shift lead" is.
7. Make a human decision and set a human next step. Every candidate who passes should speak with a person before an offer.

Software can carry the transcription and summary work in step 6, but AI candidate screening tools vary a lot in what they score and who decides. Kira-AI, for example, runs the first screen as a structured AI voice interview with no camera, then gives the recruiter a transcript, a summary, and scores linked to specific answers. It doesn't analyze emotion or personality and doesn't reject candidates automatically. The recruiter makes the decision.

Benefits and tradeoffs of asynchronous screening

The benefits are practical. Nobody plays calendar tag. A candidate working nights can answer at 11 p.m. instead of taking an unpaid break to call you back. Every candidate gets the same questions in the same order, and the recruiter can review 20 screens in one sitting instead of spreading 20 calls across a week.

The tradeoffs are just as concrete, and good process design works around them.

Candidates are not big fans of one-way formats. The SIOP summary of applicant reactions research reports less favorable reactions to asynchronous digital interviews than to some other online recruitment methods. It also notes that fairness perceptions affect whether candidates find the employer attractive and whether they'd recommend it. Separately, SHRM reports an American Staffing Association study in which 70% of US job seekers preferred in-person interviews. That figure covers interviews in general rather than async specifically, but it tells you where candidate comfort starts.

Other tradeoffs to plan for:

• You can't follow up in the moment. If an answer is vague, "what happened next?" has to wait for a later step.
• Candidates can't ask you anything. A one-way format won't tell someone the pay range, the shift pattern, or whether there's parking, and those answers decide whether good people stay in the process.
• Timed recordings can disadvantage people with speech differences, hearing impairments, or anxiety unless an alternative exists.
• Technology fails. Bandwidth drops, browsers block microphones, and phones run out of storage.
• Proxy signals creep in. Eye contact, background, fluency, and confidence are easy to notice and rarely what the job requires.

SIOP's summary also lists what makes a selection process feel fair: job-related content, a real chance to perform, consistent procedures, an honest explanation of the process, and two-way communication. A one-way screen can cover most of that list. It can't cover the two-way part, which is why the later human conversation belongs in the design from the start.

Setup checklist for a fair asynchronous interview

Run through this before the first invitation goes out. Most of it is standard candidate experience practice, applied to a format where nobody is on the line to smooth over gaps.

• Disclose the format, who reviews the answers, and whether AI transcribes, summarizes, or scores anything.
• Give an honest time estimate, such as "about 10 minutes, 5 questions."
• Include one unscored practice question so the first real answer isn't also the first time they've used the tool.
• Offer an alternative route on request: a phone call-in, written answers, a live call, or extra time. Same questions, same anchors.
• Set a technical fallback. If a recording fails, re-invite the candidate. Never score silence or a broken file.
• Write scoring anchors for every question before launch.
• Require an evidence note for every score.
• Apply the pass/clarify/stop rule below.
• Tell candidates when they'll hear back and give them a contact a person actually reads.
• Book a human conversation for every candidate who passes.

The pass/clarify/stop rule

Every completed screen should land in one of three buckets.

Clarify is the bucket teams skip, and it's where most avoidable rejections happen. Ambiguity is a reason to ask again.

Edge cases that trip teams up

Shift availability and motivation are different kinds of evidence. A candidate who writes "I can't do Sundays" has given you a fact, and a great spoken answer about why they want the job shouldn't override a true must-have. The reverse also holds: a no on a nice-to-have shift shouldn't sink a strong candidate. Decide which criteria are must-haves before anyone reviews.

Some candidates have weak mobile data or no private room. A parent recording in a shared flat, or someone answering from a car between jobs, should be able to switch to a phone line or written answers without explaining why.

A technical failure is a system problem. If an answer cuts off at eight seconds, mark it "not scored," re-invite the candidate, and note what happened. Don't let a reviewer score the fragment.

Treat accommodation and format requests as routine. The candidate gets the same questions in another format and is scored against the same anchors, with no note that suggests they asked for special treatment.

Because a one-way format can't answer questions, put the basics in the invitation: pay range, schedule, location, and what happens after the screen. Then make sure a human reads replies.

What weak and strong setups look like

A weak setup sends video by default, gives no time estimate, scores "energy" and "presence," treats a dropped recording as a non-response, and sends rejections from a no-reply address. Candidates experience it as a black box, and the strong ones with options drop out first.

A strong setup uses text for logistics and audio for judgment questions, explains the process in plain language, scores answer content against anchors with evidence notes, routes ambiguous screens to clarify, and puts every passing candidate in front of a person within a few days.

Key Takeaways

• An asynchronous interview is defined by timing. Answers can be text, audio, or video, and video is only one option.
• Decide live vs async first, then use the least intrusive format that captures job-relevant evidence.
• Audio often beats a form or a video for first screens, but it doesn't remove bias. Score answer content against anchors, not accent or delivery.
• Never score a technical failure, and offer an alternative route for bandwidth, privacy, or accommodation needs.
• Use a pass/clarify/stop rule with evidence notes so unclear answers get a follow-up instead of a rejection.
• A one-way screen can't answer candidate questions, so every candidate who passes should talk to a human next.