Coril
You open the email. The recruiter is asking for a one-way video interview. You will read questions on screen, get 30 seconds to prepare, then record your answer into a camera that will not react.
There is no live interviewer. No follow-up question to clarify what they meant. No nod, no smile, no anything. Just your face on a black rectangle, a timer, and an AI that scores it before a human watches.
That is what makes one-way video interviews feel harder than a live Zoom call. Most prep content treats one-way as a slightly weirder Zoom interview, but it is a different prep problem.
The good news lives in the scoring. The AI is not judging your likeability, your charisma, or your face (HireVue dropped facial analysis in 2021). It is scoring your content, your structure, and your pace. That is a narrower target than a human interviewer applies, and narrower targets are easier to hit once you know the shape.
Use this guide as your single reference for HireVue interview tips, one-way video interview tips, and prep for any asynchronous video interview format your employer sends, with role-specific patterns for nursing, retail, banking, and grad-scheme one-way interviews across HireVue, Modern Hire, Spark Hire, Willo, and VidCruiter.
Why one-way video interviews feel harder than live
Live interviews give you a human mirror. You ask a question with your eyes, the interviewer smiles or frowns, you adjust. The conversation has a shape because two people are building it together. One-way recorded interviews remove the mirror entirely. You talk into a camera that does not react, and the platform's silence is what makes the format different.
Real candidate language captures this exactly. "Answers disappear into the void." "Talking to a bot for your first interview can feel incredibly awkward." These pain points show up across Reddit, Quora, and platform reviews. The cognitive load is real: you are performing without feedback, cannot tell if you are landing, and will replay every awkward pause in your head for three days afterward whether the AI cared or not.
The reframe that changes everything: the platform is not judging your likeability or making a human-style holistic read. It is scoring content, structure, and pace against a fixed rubric. Once you know what is actually being measured, the void becomes a structured exam instead of a vibes-based audition.
Cross-industry: the format is the same whether you are a nurse applying to a hospital chain, a retail manager applying to Target, a graduate applying to Shell, or a banker applying to BMO. The questions vary; the prep problem does not.
Our video interview preparation guide covers live Zoom and Teams interviews, including connection stability, real-time eye contact, and reading the interviewer's reaction. This guide owns the asynchronous-recorded format that does not let you do any of that.
What the AI actually scores (and what was dropped in 2021)
What the AI actually scores is the most-searched question in this category and the one competitors answer worst. Each platform has its own marketing language, but three scoring layers hold up across the dominant ones (HireVue, Modern Hire, Spark Hire, Willo, VidCruiter). The question of how does HireVue AI scoring work is the most-cited version of that query because HireVue is the largest platform, but the answer applies industry-wide.
Verbal content
Keyword matching to the job description. Specific competency examples named clearly. Did you mention the skills the role listed? Did you give a concrete example, or did you stay generic? The model rewards specificity because specificity is harder to fake.
Answer structure
STAR-shaped responses (situation, task, action, result) land harder than ramble answers. The AI is trained to recognize the four beats and credit answers that hit them. Skip a beat (most common: skipping Task to jump from Situation directly to Action) and the structure-scoring layer marks it down.
Vocal pace and confidence cues
Steady tempo beats rushed or halting delivery. Long pauses get read as hesitation. Fast-talking gets read as anxious. The model does not score filler words in isolation; it scores the overall pace pattern they create.
What is no longer scored across the category: facial expressions. The industry inflection point was HireVue publicly discontinuing facial analysis in January 2021 after an independent audit determined it added negligible predictive value to the model. SHRM and Fortune both covered the change at the time. Modern Hire, Spark Hire, Willo, and VidCruiter operate on similar audio-and-text scoring without facial analysis. The AI is reading your words, your structure, and your pace, not your face.
Visual scoring is much smaller than candidates fear. The platform sees your face but the score is mostly your words and your pace. That is the load-bearing reframe of this whole guide.
This is not a guide to gaming the AI. No specific phrase tricks. No deliberate filler-word suppression beyond natural pacing discipline. The honest read is that mechanical scoring rewards naturally well-structured, clearly-paced answers, which is exactly what voice practice builds.
The near-universal Q1 in one-way formats is tell me about yourself. The AI scores this answer harder than any other because it is the calibration baseline for the rest of the interview. Our tell me about yourself guide covers the answer architecture that lands; in one-way format, that architecture is what the AI scores first.
Setup (the first 10 seconds the algorithm reads)
Production care is a tone-scoring signal. A clean recording reads as care; a messy one reads as lack of preparation.
Lighting
Front-facing window during daylight is the cheapest and best option. A ring light at face level works for evening recordings. Never lit from below (looks horror-movie) or behind (silhouettes you against a window).
Audio
External mic beats built-in laptop mic for the AI's tone-scoring layer. A USB lavalier, AirPods Pro, or any decent headset removes the room reverb and keyboard noise that built-in mics pick up. The model reads cleaner audio as more confident delivery.
Eye-line
Camera at eye level, not the standard laptop-on-desk down-tilt. Put your laptop on a stack of books, or use an external webcam at monitor height. Down-tilt makes you read as defensive even when you are not.
Framing
Head-and-shoulders, not full-body, not face-fills-screen. Your eyes should sit about a third of the way down from the top of the frame. Background clean and uncluttered, neutral wall preferred over a visible bedroom or kitchen.
What to wear maps to framing. Head-and-shoulders framing means the top half of your outfit is the entire visual signal. Match the level you would dress for an in-person interview at that company, then go one notch up because the camera flattens.
Tactic: do a 10-minute test recording before the real session. Watch it back at 1.5x speed and you will catch every setup flaw in 90 seconds. Lighting from the wrong angle, audio echo, eye-line off, framing too tight, background distractions. The test recording is the cheapest debugging step in the whole prep.
Setup matters more in nursing-chain HireVue and large-retail one-way screens (where production polish reads as professionalism) than in startup video screens (where polish can read as overprepared). One-way often slots in the pipeline right after the phone screen as a first cut before live rounds; our phone screen tips guide covers how to read where the format sits in the broader process.
Structure and pacing (where practice on camera unlocks)
STAR-heavy because mechanically scored. The AI reads situation, task, action, result beats and credits answers that hit all four cleanly.
Where STAR breaks
The most common failure mode is skipping Task. Candidates name the situation, then jump straight to what they did, never naming what specifically they were responsible for. The structure-scoring layer marks the gap down. Name your task explicitly even when it feels redundant.
Pacing and takes
Pacing target: 80 to 90 percent of allotted time. If you have two minutes per question, aim for 95 to 110 seconds. Hit the time without rushing or stalling. Thinking time before recording is typically 30 seconds to one minute, also employer-configurable. Use it to write three bullet points on paper (situation, task, action plus result) instead of trying to script the whole answer.
How many takes you get is the employer's choice, not a platform default. Most one-way platforms let the employer configure 0 to 3 retakes per question, with some giving unlimited and some giving exactly one. Always read the platform instructions before clicking record; the widely-shared "three takes always" rule is not real. The re-record paradox: candidates with three takes who use all three often score lower on tone than candidates who nail it in one or two, because the AI reads the third attempt as anxious. Aim to make take one the take.
Practice on camera is the only unlock for the no-feedback recording dynamic. Silent reading and mental rehearsal do not build the muscle of holding composure while talking to a black rectangle that does not react. Only on-camera repetition does.
The void anxiety lives in the gap between knowing what you want to say and being able to say it cleanly into a camera that gives you nothing back. That gap closes with reps, not with reading more articles. Rehearse on camera before the platform timer with an AI interviewer that asks the question and waits. The dynamic is the closest analog to HireVue or Modern Hire you can get without the actual employer's pipeline on the line.
Our AI practice methodology guide covers the broader pattern. For one-way prep specifically, voice practice IS the rehearsal, in the same category as the real thing rather than a different one.
Strategy by role and platform
The structure recommendation is the same across verticals (STAR-shaped, 80 to 90 percent time use). The question content varies by role. Knowing which questions to expect lets you preload three or four prepared answer architectures before the recording session.
Healthcare (nursing chains, NHS, hospital systems)
Typical questions: "describe a patient safety incident and how you handled it," "tell me about yourself," "why this unit or specialty," situational nursing-judgment scenarios. The AAAE structure (acknowledge, assess, act, escalate) from clinical scenarios lands well in the one-way format because it gives the AI a clear four-beat structure to score.
Retail and customer service (Target, Walmart, Lowe's)
Typical questions: conflict resolution scenarios, "describe a difficult customer interaction," scheduling availability and flexibility, "what would you do if" situational questions. Target-style retail formats typically give short answer windows (one to two minutes) with a brief prep period per question.
Banking and grad schemes (BMO, Shell, BAE Systems, Diageo, WTW)
Typical questions: behavioral STAR-heavy, "why this company" (the AI scores specificity here harder than almost anywhere else), competency-based questions from the published grad-scheme rubric, light brain-teaser scenarios. Grad-scheme one-way formats typically use a brief prep window with answers in the two-to-three-minute range across a handful of questions. Standard HireVue interview questions for banking tend to repeat across employers, so searching for any specific employer's question bank before recording is worth the 10 minutes.
Government grad schemes and civil service
Structured competency questions tied to the role's published behaviours framework. Often paired with a written or psychometric test before or after the video portion. Answer the named behaviour explicitly ("the behaviour here is making effective decisions, so let me name the decision I made and the reasoning") rather than leaving the AI to infer the match.
Cross-platform: HireVue dominates large enterprise and grad-scheme hiring. A Modern Hire interview is most often what you encounter in high-volume retail (and Modern Hire's interface is more candidate-friendly than HireVue's for first-time recorded interview takers). Spark Hire skews SMB and mid-market. Willo and VidCruiter are emerging across both. The scoring logic is similar across all of them, but each platform exposes slightly different employer controls on retake counts and time limits.
Nursing chains in particular use HireVue and Modern Hire heavily as a first-round screen before any live conversation. Our healthcare interview questions guide covers the substance of nursing interview answers; this guide covers the recorded-format mechanics that those answers have to land inside.
The camera does not soften you the way a human interviewer does. Setup, structure, and on-camera rehearsal close that gap. The platform timer starts when the question loads; the rehearsal already happened.