Overview
The AI Interview Practice Coach simulates a real job interview experience. Upload your resume and a job posting, configure interview parameters, and face an AI interviewer who asks role-relevant questions. Record your responses (text or video), and receive comprehensive post-interview evaluation with per-question feedback, dimension scoring, and actionable coaching tips.
Unlike generic interview question banks, this tool generates questions specific to YOUR resume against THIS job posting, evaluating gaps, strengths, and areas where the interviewer would probe deeper.
New in v2 Phase 3: Custom Persona Builder
Advanced subscribers can now design their own interviewer personas beyond the curated presets. Pick a face from D-ID's live gallery, choose a voice, dial in tone (warmth / rigor / pace), and pick the focus areas the interview should emphasize. No prompt engineering required — the builder is fully guided. Custom personas stay private to your account, are reusable across sessions, and can be edited or deleted anytime. Token cost per interview is unchanged.
New in v2 Phase 3.5: 5 More System Personas + Job-Posting Precedence
The system-persona roster now includes an HR Generalist (Morgan Alvarez, first-screen), a PMO Manager (Brian U.), an SVP of Finance (Peter F.), a Chief Operating Officer (Taylor Reid), and a Chief Executive Officer (Samira Okafor) — alongside the six industry presets. Useful when practicing for a specific interviewer type rather than an industry.
Also: the LLM question generator now treats the job posting as the ground truth when it conflicts with the persona. The persona provides voice and style; the job posting drives what's asked. Picking a Healthcare interviewer for a marketing role now produces marketing questions asked in that interviewer's style — not healthcare questions.
New in v2 Phase 1.1 + Phase 2: D-ID Video Interviewer, Split Pricing, Persona Thumbnails
D-ID video interviewer: Advanced subscribers now face a lifelike video avatar that speaks each question. The six industry personas (Alex Chen for Tech, Jamie Morgan for Finance, Dr. Sarah Patel for Healthcare, Michael Davies for Legal, Priya Rao for Consulting, Jordan Riley for Creative) now appear on camera, not just as voices. Videos are generated in parallel with the interview setup so there's no waiting per question. If D-ID is unavailable for any question, the persona's voice keeps the interview going.
Split pricing: Base cost covers the AI (1/1/2/2/3 tokens for 10/15/30/45/60 min) regardless of persona. If you pick a persona, a surcharge (+0/+1/+1/+3/+3) covers the D-ID video + HD TTS generation. You can disable the persona anytime to pay only the base cost — the live cost display under the duration picker shows the breakdown and a one-click "Disable persona" link.
Thumbnails: The persona picker now shows each interviewer's headshot — the same face you'll see on video during the session.
New in v2 Phase 1: Industry Personas + Duration-Scaled Pricing
Launched earlier the same day. Six industry-matched personas with tone + focus areas + industry-specific question banks; duration-based pricing (10 min = 1 token, 60 min = 6 tokens with persona).
How It Works
- Setup — Paste or upload your resume and a job posting. Configure interviewer role, difficulty, and question count.
- Question Generation — AI analyzes your resume vs. the job requirements and generates categorized questions (behavioral, technical, situational, brain-teaser).
- Interview Room — Questions appear one at a time with an AI avatar. Record your responses via microphone with live transcription.
- Evaluation — AI scores all responses on relevance, depth, structure, confidence, and communication.
- Results — Overall score, per-question feedback, coaching tips, and optional STAR method analysis.
Key Features
Interview Room
- AI interviewer persona with text-to-speech (premium: D-ID video persona)
- Live speech transcription via Web Speech Recognition
- Configurable countdown timer before recording starts
- Question categories — behavioral, technical, situational, brain-teaser
- Skip and review — skip difficult questions, review completed answers
Evaluation
- Multi-dimension scoring — relevance, depth, structure, confidence, communication, technical accuracy (0-10 each)
- Per-question feedback — specific feedback on each answer with key points covered vs. missed
- STAR analysis (advanced) — evaluates Situation, Task, Action, Result structure
- Filler word detection — configurable sensitivity from flexible to hyper-strict
- Communication analysis — clarity, pace, confidence indicators
Settings
- Interviewer role — HR Recruiter, Hiring Manager, Director, VP, Owner, Team Member
- Difficulty — 1-5 slider affecting question complexity
- Evaluation strictness — 1-5 slider for scoring tolerance
- Question count — 5, 8, 10, or 15 questions
- Follow-up depth — 0, 1, or 2 follow-up questions per main question
- Brain teasers — toggle analytical puzzle questions
- Video recording — record via webcam for body language analysis
Post-Interview
- Coaching tips — prioritized, actionable improvement areas
- Coaching Q&A chat — ask follow-up questions about your performance
- PDF/DOCX export — full assessment report with all scores and feedback
- Cross-tool links — jump to Speech Coach for delivery practice, Career Gap Analyzer for skill gaps
Bias-Aware AI Interviewing
A default-off "Apply EEOC-aligned anti-bias rules" toggle in the setup form prepends a strict instruction set to the AI: evaluate the candidate strictly on role-relevant skills and stated competencies, ignore protected-class attributes (race, gender, age, disability, religion, national origin, marital/parental status, military/veteran status) and their proxies (school name, location, names suggesting cultural origin, gap patterns), and avoid "cultural fit" as a scoring factor. Aligns with U.S. EEOC, UK Equality Act 2010, EU Equal Treatment Directive, and Canadian Human Rights Act frameworks. Useful for hiring teams who use the tool to draft interview question banks for live candidate sessions.
Use Cases
- Job seekers preparing for a specific interview with a known job posting
- Career changers practicing how to frame transferable experience
- Students preparing for their first professional interviews
- Professionals preparing for promotion interviews at higher levels
- Recruiters testing interview question quality before candidate sessions
- Career coaches using the tool with clients for structured practice
From Demo to Production
This demo provides text-only interviews with 5 questions at basic difficulty. A production deployment would add:
- AI video persona — D-ID talking-head video interviewer for realistic simulation
- Video recording + body language analysis — GPT-4o vision analyzes gestures, eye contact, posture
- Follow-up questions — AI probes deeper based on initial answers
- Interview history — track improvement across multiple practice sessions
- Custom question banks — company-specific or role-specific question libraries
- Panel interviews — simulate multiple interviewers with different personas
- Mock assessment centers — combine interview, presentation, and group exercises
Real-World Challenges
| Challenge | Why It's Hard |
|---|---|
| Speech recognition accuracy | Accents, technical jargon, and quiet environments affect transcription quality. Production needs noise cancellation and accent adaptation. |
| Answer evaluation subjectivity | What counts as a "good" answer varies by industry, role level, and company culture. Evaluation rubrics need calibration per context. |
| Video analysis ethics | Body language analysis must avoid cultural bias (eye contact norms vary globally) and disability discrimination. |
| Real-time video persona | D-ID video generation takes 5-15 seconds per question. Production needs pre-generation or streaming approaches. |
| Candidate anxiety simulation | Practice interviews can't fully replicate real interview pressure. Production could add time pressure and unexpected curveballs. |
Cost Estimates (Platform Deployment)
| Component | Starter | Growth | Enterprise |
|---|---|---|---|
| AI API (question gen + evaluation) | $100-400/mo | $400-1,500/mo | $1,500-5,000/mo |
| D-ID video persona API | $200-800/mo | $800-3,000/mo | $3,000-10,000/mo |
| Speech-to-text (if not browser) | $50-200/mo | $200-800/mo | $800-3,000/mo |
| Total monthly | ~$350-1,400 | ~$1,400-5,300 | ~$5,300-18,000 |
ROI Definition
- Primary metric: Interview pass rate improvement — candidates who practice pass 2-3x more often
- Secondary metric: Time-to-hire reduction — better-prepared candidates reduce interview rounds
- Break-even: Within 1 month for staffing agencies; within 1 quarter for enterprise L&D
- Concrete example: If 20 candidates per month use the tool and interview pass rate improves from 25% to 50%, that's 5 additional hires per month without additional recruiting spend
Technology Stack
- AI Model: OpenAI GPT-4o-mini (demo) / GPT-4o (advanced)
- Speech: Web Speech Recognition API (browser-native, free)
- Video Persona: D-ID Talks API (premium only)
- Video Analysis: GPT-4o Vision (premium only)
- Backend: Next.js API routes (serverless)
- Frontend: React multi-step interview flow
Want This for Your Business?
White-label deployment for staffing agencies, career services, universities, and enterprise L&D. Includes custom question banks, branded interview rooms, and integration with your ATS. A full deployment typically takes 4-6 weeks and starts at $8,000.
This demo uses GPT-4o-mini for question generation and evaluation. Speech recognition uses your browser's built-in capability. No audio or video data is stored on our servers.