Most candidates treat a HireVue link like a lie detector for smiles. HireVue itself stopped using visual analysis in new assessment models in early 2020 and said so publicly in January 2021; later that year it also removed speech inputs such as tone and pauses from new models so that scoring, when AI is used at all, tracks what you say. After this piece you will know what is and is not scored today, how to budget a 90-second answer, and how to treat the invitation email as the only reliable spec for your invite.
This guide is firm-agnostic. Consulting employers use several one-way video platforms, and the same brand name can mean either a recorded screen for humans or a scored assessment. For BCG's own one-way format, use the dedicated BCG one-way video interview guide.
The three things candidates confuse
A HireVue invite usually sits early in the funnel: after the application, before live case and fit rounds. Firms use one-way video to standardize prompts, cut scheduling friction, and replace a short phone screen. HireVue's own candidate FAQ frames OnDemand interviews and video-based assessments as early steps that help recruiters make a first cut before person-to-person interviews, not as the final offer decision.
Three different products get collapsed into one rumor:
| What you received | Who watches | What typically gets "scored" |
|---|---|---|
| On-demand recording on HireVue (or similar) | Recruiters or assessors review the video | Human judgment against a competency rubric |
| HireVue video-based assessment with AI models | Software highlights competencies; humans still decide | Language content related to role competencies, not face or body language |
| Firm assessment on another vendor (for example SHL) | Firm-defined process | Whatever that firm's careers page states for that market |
HireVue is explicit that it does not use the recording for facial recognition identity matching, and that for the vast majority of jobs people still make the hiring call after later interviews. Separately, when AI assessment models are in play, HireVue's 2021 technical write-up states that AI scores what is said and does not interpret facial expressions, body language, emotions, or surroundings.
So the folklore about "smile more for the algorithm" is solving the wrong problem. The useful preparation is a clear story that survives as a transcript.
What HireVue published about scoring (timeline)
Read the company's own chronology rather than prep-site archaeology.
| Date | HireVue's stated change | Candidate takeaway |
|---|---|---|
| Early 2020 | Visual analysis removed from new assessment models (announced 12 Jan 2021) | Face and body language are not inputs to new AI models |
| 12 Jan 2021 | Public commitment and ORCAA audit release alongside the visual-analysis decision | Treat "facial analysis scores you" as outdated for current models |
| 30 Sep 2021 | New models drop speech inputs (tone, pauses); focus on transcribed language and NLP | Content and structure matter more than theatrical delivery |
That last point is the one most consulting candidates still miss. Even if your invite includes an AI assessment layer, the documented direction of the product is toward language and job-related competencies such as communication, problem solving, team orientation, and initiative — HireVue lists these categories on its candidate FAQ — not toward reading your microexpressions.
What HireVue does not publish for your specific invite: whether your employer enabled AI scoring, which competency model was trained, or the exact pass threshold. Invitation emails and firm careers pages are the primary sources for those details. Prep blogs that invent a universal "HireVue scores X percent on tone" are guessing.
Consulting employers that name one-way video (primary pages)
Platform names vary by office and program. What matters is that several consulting and professional-services employers publish an asynchronous video step on their own careers sites.
Bain & Company states that applicants to its First Forward program answer several questions in video format through the HireVue platform after the application deadline, with no formal live interviews for that program.
PwC UK Early Careers describes an on-demand video interview on the SHL platform: questions appear on screen, answers are recorded, and assessors review them afterward. That page also states that PwC does not use AI to assess candidate performance in its Early Careers selection process. Same format family (one-way video); different vendor; explicitly human review.
Deloitte UK Early Careers includes recording video responses inside its job simulation step, alongside written and ranking tasks. The firm also documents accessibility options for pre-recorded or live video interviews.
None of those pages requires you to optimize for facial scoring. They require you to answer timed prompts clearly enough that a reviewer (or, where AI is used, a language model) can map your story to published competencies.
Format: use the invitation, not a rumor range
HireVue's candidate guidance is that most video-based assessments include five to eight questions, give several minutes to respond to each, and often finish in under 20 to 30 minutes if you budget about 45 minutes end to end for setup. HireVue also notes unlimited practice questions before the real interview starts, and that practice takes are not seen by the employer.
Beyond that baseline, the invitation owns the numbers: prep seconds, recording minutes, retry count, deadline, and whether a game or written prompt is bundled. I treat any blog that claims a single "standard consulting HireVue" as marketing.
Practical defaults when the email is vague:
- Block 45 quiet minutes.
- Assume one take per question unless the foyer screen says otherwise.
- Aim to finish each answer with 10 to 15 seconds left on the clock. Cut-off endings score worse with humans than slightly short answers.
Ned's 90-second budget (worked example)
Consulting one-way prompts are mostly past-behavior and motivation: conflict, leadership without authority, a hard deliverable, why this firm, why consulting. Situational judgment ("what would you do if…") shows up too. Full cases are rare on the pure video screen; save deep case muscle for later rounds and case interview prep.
Here is a reproducible timing budget for a 90-second answer. Invented company, invented numbers — use your own facts.
Prompt: Tell me about a time you influenced a decision without formal authority.
Weak take (what I stop listening to): 40 seconds of team history, 35 seconds of "we collaborated," 15 seconds of "it went well." No decision, no personal moves, no outcome a reviewer can write on a score sheet.
Strong take (same 90 seconds):
| Segment | Seconds | What you say |
|---|---|---|
| Situation | 12 | "In my internship at Northline Logistics, the ops lead wanted to delay a carrier switch that was costing about $18k a month in mid-route fees." |
| Task | 8 | "I owned the analysis but not the decision. My job was to get the switch approved before month-end." |
| Action | 50 | Three moves with "I": rebuilt the cost model with the finance analyst; ran a 20-minute walkthrough with the skeptical warehouse manager using two weeks of exception data; proposed a two-week pilot instead of a full cutover. |
| Result | 20 | "Pilot cleared; full switch landed two weeks later. Mid-route fees fell about 40% in the next month, roughly $7k saved, and the ops lead asked me to own the next carrier RFP." |
That shape is transcript-friendly. A muted video still carries the story. A face-optimized performance with empty content does not.
Ned's rule. If your answer would still earn points as a plain transcript with the camera off, you are ready. If the only thing you practiced was eye contact and smiling, you prepared for a scoring system HireVue publicly retired.
How reviewers actually mark these
When I score early screens, I am not grading cinematography. I am asking four questions:
- Did you answer the prompt asked, or a neighboring prompt you preferred?
- Can I name your personal contribution in one sentence?
- Is there a result I could write on a sheet (number, decision, or clear before/after)?
- Would I want this person in a client workshop tomorrow — calm, structured, specific?
Keyword stuffing does not help. HireVue states that video-based assessments consider whole-context responses, not isolated keywords. Rehearsed scripts that ignore the question fail the first check above, which is enough to lose the round with a human reviewer.
Dress, lighting, and a quiet room still matter for human watchers. They do not resurrect facial analysis. Treat tech setup as hygiene: charge the device, test camera and mic in the practice foyer, put notes as bullet cues off-camera, look near the lens rather than at your own face on screen.
Practice this before your deadline
Build three stories that cover conflict, delivery under pressure, and leadership without a title. Speak each one aloud to a timer at 90 seconds. Then cut each story until the Action and Result occupy about 70 percent of the clock.
On CoachNed, run the same stories through behavioral practice with follow-ups, or take the five-minute first rep at /start. Everything is open for seven days, no card; then $120 for a recruiting season or $49 a month.
Frequently asked questions
Does HireVue still use facial analysis?
No for current assessment models. HireVue announced in January 2021 that it had removed visual analysis from new models after research showed language added more predictive value, and its September 2021 update reiterated that AI does not score facial expressions, body language, emotions, or surroundings.
Does every HireVue interview use AI scoring?
No. Firms can use the platform as a recording tool for human review, as an AI-assisted assessment, or both depending on the role. Some consulting employers publish one-way video on other vendors entirely. Your invitation and that employer's careers page are the sources of truth.
How many questions are in a HireVue assessment?
HireVue's candidate FAQ says most video-based assessments include five to eight questions with several minutes to respond to each. Exact counts, prep time, and retries are set per invitation.
Can I retake a question if I stumble?
Only if the employer enabled retries. HireVue's practice foyer typically shows how many attempts you get for real questions. Assume one take unless the screen says otherwise, and use practice questions — which HireVue says are not shared with the employer — to settle nerves first.
Is a HireVue screen the same as a consulting case interview?
No. One-way video is usually a behavioral and motivation gate. Live case interviews come later and decide most offers. Prepare stories for the video; keep case drills for the rounds that follow.
Sources
- Industry leadership: new audit results and decision on visual analysis — HireVue's January 2021 announcement removing visual analysis from new assessment models. Checked 2026-09-24.
- The latest leap in HireVue's assessment technology — September 2021 statement that AI scores what is said, not visuals, and that speech inputs (tone, pauses) were removed from new models. Checked 2026-09-24.
- HireVue candidate FAQ — assessment length, question counts, competencies, facial recognition clarification, and human decision-making. Checked 2026-09-24.
- Bain First Forward program — Bain names HireVue video questions for that program. Checked 2026-09-24.
- PwC UK Early Careers selection process — on-demand video interview via SHL with human assessor review; firm states it does not use AI to assess Early Careers performance. Checked 2026-09-24.
- Deloitte UK Early Careers assessment — job simulation includes recorded video responses. Checked 2026-09-24.
CoachNed is not affiliated with HireVue, McKinsey, BCG, Bain, Deloitte, PwC, or any firm named above.