Recruiting is one of the most time-intensive functions in any organization. A single hire requires screening 100-250 resumes, conducting 15-20 phone screens (30 minutes each), scheduling 5-8 first-round interviews, and managing feedback from 3-5 interviewers. A recruiter handling 10 open positions simultaneously spends 60-70% of their time on screening calls that yield a 20-30% advancement rate. The math is brutal: 200 hours of phone screens per month to hire 3-4 people.
Async video interviews eliminate the phone screen entirely. Instead of scheduling 30-minute calls with 20 candidates, the recruiter sends all 20 candidates a link to a video interview. Each candidate opens the link in their browser, sees the preset questions, and records their answers on their own schedule. No app install. No calendar coordination. No phone tag. The recruiter reviews all 20 responses in 2-3 hours instead of 10 hours, watches at 1.5x speed, and advances the top candidates to live interviews. Time savings: 60%.
V100's API provides the video infrastructure for this workflow: candidate recording, transcription, AI analysis (keyword matching, communication assessment, sentiment), structured data for comparison, hosting on CDN, and viewer analytics. You build the recruiter dashboard, the candidate experience, ATS integrations (Greenhouse, Lever, Workday), and the scoring workflows. This guide walks through the complete architecture.
The Video Interview Market Opportunity
The video interview market is projected to exceed $1.5 billion by 2027, growing at 25% annually. HireVue, the market leader, processes over 30 million interviews per year and was acquired by The Carlyle Group for reportedly over $200 million. Spark Hire, myInterview, and Vidcruiter each serve tens of thousands of employers. Despite this growth, market penetration is still low. Only 15-20% of employers use video interview tools, which means 80% of the market is still using phone screens.
The opportunity is not in building a better HireVue. HireVue serves enterprise customers with complex compliance requirements and deep ATS integrations. The opportunity is in the mid-market (50-500 employees) and in specific industry verticals (healthcare, hospitality, retail, staffing agencies) where hiring volume is high but budgets for enterprise tools are limited. A video interview tool priced at $79-199 per recruiter per month (versus HireVue's enterprise pricing) captures the mid-market segment.
Vertical specialization is the strongest competitive strategy. A video interview tool built specifically for healthcare recruiting includes compliance features for HIPAA, supports credential verification workflows, and integrates with healthcare-specific ATS systems. A tool for retail and hospitality handles high-volume hiring (hundreds of candidates per week) with automated screening and bulk invitation workflows. These vertical features are difficult for horizontal platforms to justify building but are exactly what buyers in those verticals need.
Architecture: Async Video Interview Pipeline
Interview pipeline
Step 1: Interview Template (Questions and Time Limits)
The interview template is the foundation of the async interview experience. A recruiter creates a template with 3-7 questions, each with an optional time limit (1-3 minutes per question is standard). The template also includes an optional intro video where the recruiter or hiring manager introduces themselves and the role, which personalizes the experience and improves candidate completion rates.
Templates are reusable. A recruiter creates one template for "Customer Success Manager" and uses it for every CSM hire. Templates can be shared across the recruiting team so that all positions of the same type use consistent questions, making cross-candidate comparison fair and structured.
Question design matters significantly. The best async interview questions are behavioral ("Tell me about a time when...") or situational ("How would you handle...") rather than yes/no or factual. Behavioral questions generate richer video responses that give the recruiter more signal about the candidate's communication style, thought process, and experience depth. V100 provides question library templates organized by role type, but the recruiter can customize or create their own.
import { V100 } from 'v100-sdk';
const v100 = new V100('YOUR_API_KEY');
// Create an interview template
const template = await v100.interviews.createTemplate({
title: 'Customer Success Manager — First Screen',
introVideo: './hiring-manager-intro.mp4',
questions: [
{
text: 'Tell us about yourself and why you are interested in this role.',
timeLimit: 120, // 2 minutes max
prepTime: 30, // 30 seconds to think
retakes: 1 // Allow 1 retake
},
{
text: 'Describe a time you turned an unhappy customer into a successful one.',
timeLimit: 180,
prepTime: 30,
retakes: 1,
keywords: ['churn', 'retention', 'escalation', 'NPS', 'upsell']
},
{
text: 'How do you prioritize when managing 50+ accounts simultaneously?',
timeLimit: 120,
prepTime: 30,
retakes: 1,
keywords: ['segmentation', 'health score', 'automation', 'risk']
},
{
text: 'What tools and processes do you use to track customer health?',
timeLimit: 120,
prepTime: 30,
retakes: 0
}
],
settings: {
deadline: '2026-04-15', // Candidates must complete by
language: 'en',
transcription: true,
aiAnalysis: true
}
});
// Send invitations to candidates
await v100.interviews.invite({
templateId: template.id,
candidates: [
{ email: 'candidate1@email.com', name: 'Alex Rivera' },
{ email: 'candidate2@email.com', name: 'Jordan Park' },
{ email: 'candidate3@email.com', name: 'Sam Okafor' }
],
emailTemplate: 'friendly', // Warm, branded invitation email
reminderDays: [3, 7] // Send reminders at 3 and 7 days
});Step 2: Candidate Experience (Record in Browser)
The candidate experience determines your completion rate. A complicated or intimidating process loses candidates before they record a single answer. The best-performing async interview tools achieve 70-80% completion rates. The worst-performing ones (requiring app downloads or account creation) see 30-40% completion.
The candidate clicks a link in the invitation email. The link opens a web page (not an app) that works in Chrome, Firefox, Edge, and Safari. The page shows the role title, the hiring manager's intro video, and the list of questions. The candidate clicks "Start Interview," grants camera and microphone permissions, and records their answers one question at a time. Each question is displayed as text on screen while the candidate records. A countdown timer shows the remaining time. If retakes are allowed, the candidate can re-record before moving to the next question.
V100 handles the recording via the browser's native getUserMedia API. Camera and microphone only (no screen sharing for interview recordings). Each answer is recorded as a separate video and uploaded to V100 for processing. If the candidate's internet connection drops during upload, the upload resumes automatically when connectivity returns. The candidate's recordings are stored locally until upload is confirmed.
Preparation time is an underrated UX feature. Before each question's recording starts, the candidate gets 30 seconds (configurable) to read the question and think about their answer. A "Ready" button lets them start early if they do not need the full prep time. This reduces anxiety and produces better-quality responses. Nervous candidates who are forced to start recording immediately give worse answers, which is bad for the candidate and bad for the recruiter who needs signal.
Step 3: AI Processing (Transcribe, Analyze, Score)
Once all answers are recorded and uploaded, V100 processes each response through the transcription and analysis pipeline. The output is structured data that powers the recruiter dashboard.
Transcription converts each answer to text with timestamps. This allows recruiters to read answers instead of watching videos when they want to scan quickly, and it enables keyword search across all candidates for a position.
Keyword matching checks each transcript against the keywords defined in the interview template. If the recruiter specified that a strong answer to "Describe a time you turned an unhappy customer into a successful one" should include terms like "churn," "retention," "escalation," or "NPS," the AI flags which keywords appear in each candidate's response. This is not a measure of answer quality. It is a signal that the candidate has domain-relevant vocabulary, which correlates with relevant experience.
Communication analysis evaluates the transcript for clarity (sentence structure, coherence), conciseness (answer length relative to content density), and relevance (whether the answer addresses the question asked). These are text-based analyses of the transcript, not video-based analyses of body language. This distinction matters. Analyzing body language via video has been shown to introduce bias related to cultural communication norms, disability, and neurodivergence. V100 deliberately focuses on content analysis rather than appearance or behavioral analysis.
A note on AI bias in recruitment
AI analysis of job interviews is a sensitive area with real ethical and legal implications. HireVue faced significant backlash and regulatory scrutiny for using facial expression analysis in hiring decisions, and ultimately discontinued that feature in 2021. Illinois, Maryland, and several other jurisdictions have passed laws regulating AI in hiring.
V100's AI analysis focuses on transcript content (keywords, relevance, clarity) rather than video-based behavioral analysis (facial expressions, eye contact, voice tone). This is a deliberate design choice. Content-based analysis is more defensible legally, less prone to bias, and more useful to recruiters who care about what the candidate said, not how they looked while saying it.
That said, content-based analysis is not bias-free. Language patterns, vocabulary choices, and communication styles vary by culture, socioeconomic background, and first language. AI scoring should always be one input into the recruiter's decision, not the sole determinant. Build your product with human review as the primary decision mechanism and AI as an assistive tool.
Step 4: Recruiter Dashboard (Side-by-Side Comparison)
The recruiter dashboard is where screening decisions happen. The key UX is side-by-side comparison: the recruiter selects 2-4 candidates and views their answers to the same question simultaneously. This is the fundamental advantage of async video over phone screens. With phone screens, the recruiter speaks to candidates sequentially over days and relies on memory and notes to compare them. With the dashboard, the recruiter watches three candidates answer "Describe a time you turned an unhappy customer into a successful one" back-to-back and immediately identifies the strongest response.
For each candidate, the dashboard shows: the video recording (playable at 1x, 1.5x, or 2x speed), the full transcript, the AI summary, keyword match indicators, and a scoring rubric. The recruiter rates each answer on a configurable scale (1-5 stars, thumbs up/down, or custom rubric) and adds written notes. These ratings and notes are visible to other members of the hiring team who review the candidates later.
// Fetch candidate responses for review
const responses = await v100.interviews.getResponses({
templateId: template.id,
include: ['transcript', 'analysis', 'keywords'],
sortBy: 'ai_score', // AI-ranked shortlist
status: 'completed'
});
// Each response includes per-question data:
// {
// candidate: { name: "Alex Rivera", email: "..." },
// completed_at: "2026-04-02T14:30:00Z",
// answers: [
// {
// question: "Tell us about yourself...",
// video_url: "https://cdn.v100.ai/int/...",
// duration: 98, // seconds
// transcript: "I've spent the last 4 years in customer success...",
// analysis: {
// clarity: 0.88, // 0-1 score
// relevance: 0.92, // How well answer addresses question
// conciseness: 0.75, // Content density relative to length
// keywords_matched: ["retention", "NPS", "churn"],
// keywords_missing: ["escalation", "upsell"]
// }
// }
// ],
// overall_score: 0.85
// }Playback speed control is a small feature with massive impact. Recruiters who review 20+ video interviews per day become efficient at watching responses at 1.5x or 2x speed. A 2-minute response watched at 1.5x takes 80 seconds. Across 20 candidates with 4 questions each, that is the difference between 160 minutes and 107 minutes of review time. Add the ability to skip to the next question with a keyboard shortcut, and a recruiter can screen 20 candidates in under 90 minutes.
Step 5: Team Sharing (Hiring Committee Reviews)
Hiring decisions are rarely made by one person. The recruiter screens candidates, but the hiring manager, team lead, and sometimes peer interviewers all need to weigh in before a candidate advances. In the phone screen model, the recruiter writes up notes and shares them. The hiring manager trusts the recruiter's subjective summary. With async video, the hiring manager can watch the actual candidate responses and form their own opinion.
V100 enables this through shareable review links. The recruiter shares a curated set of candidates (the shortlist) with the hiring team via a link. Each team member can watch the responses, read transcripts, add timestamped comments ("At 1:23, this answer about churn reduction is very strong"), and submit their own ratings. Comments and ratings are visible to all reviewers, creating a collaborative evaluation that is more thorough and less biased than a single person's assessment.
Timestamped comments are a key feature. Instead of saying "I liked Alex's answer to question 2," a reviewer says "At 0:45 in question 2, Alex describes a specific churn reduction strategy that reduced churn by 30% over 6 months. This is exactly the experience we need." Other reviewers can click the timestamp and jump directly to that moment in the video. This makes feedback specific, verifiable, and actionable.
Step 6: Analytics (Time-to-Hire, Completion Rates, Drop-Off)
Recruitment analytics answer three questions. First, is the tool reducing time-to-hire? Track the number of days from job posting to offer acceptance for positions that use video interviews versus those that use traditional phone screens. A 25-40% reduction in time-to-hire is typical. Second, are candidates completing the interviews? Track completion rates by role, question count, and time limit. If a template with 7 questions has a 45% completion rate while a template with 4 questions has a 78% completion rate, the data tells you to use fewer questions. Third, where do candidates drop off? If 30% of candidates abandon the interview at question 3, the question may be too difficult, too personal, or the accumulated recording time may feel too long.
Key metrics
Completion Rate
Percentage of invited candidates who complete all questions. Industry benchmark: 65-80%. If below 60%, reduce question count or time limits. Track by role type and source (internal referral vs. job board) for deeper insights.
Time-to-Screen
Average time from invitation sent to recruiter review completed. Async video should reduce this from 5-7 days (phone screen scheduling) to 1-2 days (candidates record on their own time, recruiter batch-reviews).
Advancement Rate
Percentage of video-interviewed candidates advanced to live interviews. Compare with phone screen advancement rates. If video advancement rates are higher, the tool is helping recruiters identify better candidates earlier.
Question-Level Drop-Off
Percentage of candidates who abandon at each question. High drop-off at a specific question indicates the question needs revision. Track this per template to optimize the candidate experience over time.
Recruiter Efficiency
Number of candidates screened per recruiter per day. Phone screens: 6-8 candidates per day. Video reviews: 20-30 candidates per day (at 1.5x speed with keyboard shortcuts). This is the headline productivity metric.
Cost: V100 vs. HireVue vs. Spark Hire vs. Custom
| Feature | HireVue | Spark Hire | V100-Built |
|---|---|---|---|
| Pricing model | Enterprise quote | $149-$499/mo | You set pricing |
| AI transcription | Yes | Limited | 40+ languages |
| AI analysis | Behavioral (controversial) | Basic | Content-focused |
| Customization | Limited | Moderate | Full control |
| Multi-language dubbing | No | No | Yes (V100) |
| White-label | Enterprise only | Higher tiers | Always (your brand) |
| V100 infra cost | N/A | N/A | $500-$5K/mo |
Honest comparison
HireVue and Spark Hire are mature products with years of ATS integrations, compliance certifications, and enterprise customer references. Building on V100 gives you full control over the product, pricing, and vertical specialization, but you are starting from zero on integrations, compliance, and customer trust. If your target market requires SOC 2 certification, GDPR compliance documentation, and pre-built Greenhouse/Lever integrations, budget 3-6 months of development work beyond the core video interview functionality.
The advantage of building on V100 is speed to market for a differentiated product. If you are building a video interview tool specifically for healthcare recruiting, or specifically for high-volume retail hiring, you can ship a vertical-specific product in 2-3 months that serves those segments better than horizontal tools like HireVue ever will.
Revenue Model
| Model | Price | Target |
|---|---|---|
| Per-Interview | $2-$5/interview | Staffing agencies, seasonal hiring, low-volume employers |
| Recruiter Pro | $79-$99/recruiter/mo | In-house recruiters, unlimited interviews, AI analysis, team sharing |
| Team | $149-$199/recruiter/mo | Recruiting teams, ATS integrations, advanced analytics, custom branding |
| Enterprise | Custom | SSO, compliance, SLA, dedicated support, custom AI model training |
What V100 Does Not Do
- • Applicant Tracking System. V100 is video infrastructure. It does not track applicants through the hiring pipeline. You integrate with existing ATS platforms (Greenhouse, Lever, Workday, iCIMS) via their APIs. V100 provides the video data. Your application bridges V100 and the ATS.
- • Behavioral video analysis. V100 does not analyze facial expressions, eye contact, body language, or voice tone for hiring assessment. This is a deliberate choice. Behavioral video analysis in hiring has been shown to introduce bias and has attracted regulatory scrutiny. V100 analyzes transcript content only.
- • Live video interviews. V100's recruitment workflow is asynchronous. Candidates record answers on their own time. If you need live two-way video interviews (panel interviews, technical interviews with screen sharing), that is a different product. Many video interview platforms offer both async and live. The live component requires real-time video conferencing infrastructure.
- • Compliance certifications. V100 provides the infrastructure. SOC 2, GDPR, EEOC compliance, and local hiring regulations are your responsibility to implement at the application level. These certifications are essential for enterprise sales and regulated industries.
- • Candidate sourcing. V100 does not find candidates. It interviews candidates you have already sourced. Candidate sourcing (job board postings, LinkedIn outreach, referral programs) remains your responsibility or your client's responsibility.
Ready to build your video interview platform?
Start with V100's free tier. Create an interview template, send a test invitation, record a practice response, and see the AI analysis in action. No credit card required for the free tier.