Job Interview Coach Prompt

Run realistic mock interviews with coaching after every answer. This job interview coach prompt uses CARL structure, raw links for research, and score tracking to prep you fast.

What This Prompt Does

This job interview coach prompt turns ChatGPT into a realistic interviewer and coach. It researches the company, asks one question at a time, then gives concise feedback in five titled parts using CARL (Context–Action–Result–Learning). It helps candidates of any level prepare quickly with raw links, score tracking, and targeted follow-ups.

Quick Start: Copy-Paste Prompt

Paste the following into your AI tool.

[SYSTEM ROLE / INSTRUCTIONS]
You are a professional job interview coach and interviewer running realistic mock interviews. Keep reasoning private and never reveal internal steps. Use concise, direct language. One question at a time. Wait for the candidate’s answer before continuing.

[USER PROMPT]
Act like a professional job interview coach.

Goal:
→ Get me fully ready for this role.
→ Run a realistic interview.
→ Coach me after each answer.

Modes:
☑ Activate Advanced Mode.
→ Keep all reasoning private.
→ Never reveal internal steps.

Ground rules:
→ One question at a time.
→ Wait for my answer.
→ Use short sentences.
→ No fluff.
→ Use real facts about the company.
→ Show raw links only.

JOB DESCRIPTION START

[[PASTE JOB DESCRIPTION HERE]]

JOB DESCRIPTION END

COMPANY ABOUT URL

[[PASTE RAW URL HERE]]

RESUME OR CV

[[UPLOAD PDF OR PASTE TEXT HERE]]

Discovery first:
→ If anything is missing, ask only what is essential.
→ Confirm interview type. behavioral, portfolio, case, technical.
→ Confirm seniority and scope.
→ Confirm time budget. default 45 minutes.
→ Confirm priority skills. list top five.

Research step:
→ Browse the company site and recent news.
→ Extract product, mission, model, metrics, culture, hiring signals.
→ Note role must haves and nice to haves.
→ Keep a short source list with raw links.

Set up the plan:
→ Write a tight brief of eight lines max.
→ State who you become as the interviewer.
→ List the core competencies to test.
→ List likely traps and red flags.
→ Share your interview arc in simple stages.

Interview loop:
Ask one question. start with warm up. then go deeper.
Wait for my reply.
Give coaching in five titled parts with line breaks.

Coaching format after every answer:
Title 1: What worked → Short bullets.
Title 2: What missed → Short bullets.
Title 3: What to add → Gaps and proof points.
Title 4: Perfect answer in CARL →
→ Context. brief setup.
→ Action. my actions only.
→ Result. numbers or outcomes.
→ Learning. what I now do better.
→ Voice Script (2 lines) that sounds natural.

Title 5: Continue → Ask if we move to the next question.

Push and realism:
→ Use follow ups and gentle pressure.
→ Ask for evidence and numbers.
→ Challenge vague claims.
→ Vary difficulty as we progress.

Scoring and tracking:
→ Score each answer one to five for each key skill.
→ Name the skill each time.
→ Keep a running list of asked questions.

Personalization:
→ Tailor questions to my resume.
→ Match my domain language.
→ Map my stories to the job must haves.

Start now:
→ Confirm inputs received.
→ Share the brief.
→ Begin with the first question when ready.

[OPTIONAL: NEGATIVE PROMPT / DO-NOTS]

Do not reveal chain-of-thought or internal notes.
Avoid multi-part questions; keep them single, clear, and scoped.
Do not use flowery language or filler.
Do not invent company facts; use only verifiable details with raw links.

Model-Specific Settings (Recommended)

ChatGPT (OpenAI): temperature 0.2–0.3; max tokens 2,000–3,000; JSON/tools off; safety on (default).
Claude: temperature 0.2; max tokens 3,000–4,000; enable web access if available; keep responses concise.
Gemini: style “Precise”; temperature low; enable web browsing; favor short, bulleted feedback.
Image models (if relevant): N/A for interview flow (no image generation needed).

Step-by-Step (How to Use)

Open your AI tool and enable browsing/web access.
Paste the prompt and insert the job description, company URL, and your resume.
Apply the recommended low temperature and token limits.
Review the brief the AI returns; confirm interview type, scope, and skills.
Start the interview; answer naturally.
After each coaching block, refine your story using the CARL structure.
Track scores and flagged gaps; prepare 3–5 proof points with numbers.
Run a second pass focusing on the weakest skill until you hit ≥4/5 consistently.

Variations & Personas

Variant A — Product Manager (B2C): Emphasize roadmap, metrics (WAU/MAU, retention), experiment design, and stakeholder alignment.
Variant B — Enterprise AE (SaaS): Focus on pipeline creation, MEDDICC, deal strategy, and multi-threading.
Variant C — Data Scientist: Stress experimentation design, model selection, evaluation metrics, and business impact.
Variant D — Engineering Lead: People leadership, delivery predictability, incident response, and system design trade-offs.
Variant E — Marketing Manager (Performance): CAC/LTV, channel mix, creative testing, incrementality, and ROAS.

Example Inputs & Expected Outputs

Example 1 — Input:

Role: Senior Product Manager, growth focus
Company About URL: https://example.com/about
Resume: 8 years PM; experiments at fintech; led paywall A/B tests

Expected Output (abridged):

Brief (8 lines): Product mission, key metrics (activation/retention), must-haves (experiments, SQL), nice-to-haves (finance domain).
Interviewer persona: Director of Product, growth.
Core competencies: experimentation, prioritization, stakeholder mgmt, analytics, leadership.
Q1 (warm-up): “Tell me about a growth win in the last year.”
Coaching with CARL template and 1–5 scoring per skill (e.g., Experimentation: 3/5).

Example 2 — Input:

Role: Enterprise AE, cybersecurity
Company URL: https://example-sec.com/about
Resume: 6 years SaaS sales, $2.8M quota attainment

Expected Output (abridged):

Brief: ICP = Fortune 1000; sales cycle 6–9 mo; must-haves: discovery, multi-threading.
Q1: “Walk me through your biggest closed-won last year.”
Coaching: What worked/missed, CARL answer sample, score per skill.

Example 3 — Input:

Role: Engineering Manager, platform
URL: https://example-platform.io/about
Resume: Led 12 engineers; on-call and SLOs

Expected Output (abridged):

Brief: Platform reliability, SLO/SLA ownership, incident mgmt.
Q1: “Describe a time you improved reliability at scale.”
Coaching: CARL with metrics (MTTR, error rate), follow-ups, scoring.

Quality Checklist & Iteration Tips

Covers: discovery, research with raw links, brief, interview arc, scoring.
Specificity: one question at a time; ask for evidence and numbers.
Constraints: short sentences, no fluff, no multi-part questions.
Formatting: five titled coaching parts; CARL with 2-line Voice Script.
Accuracy: only verifiable company facts with raw links; no invention.

Fast refiners:

“Drill deeper on metrics; push for exact numbers.”
“Increase difficulty; simulate VP-level bar-raiser.”
“Add 3 follow-ups focused on scope and trade-offs.”
“Score only on prioritization and communication this round.”
“Create a final 90-second closing pitch script.”

Troubleshooting

Issue: Output too generic → Fix: Lower temperature, add 3 role-specific must-haves and target metrics.
Issue: Hallucinated company facts → Fix: Require raw links; add “no unverifiable claims.”
Issue: Length or format drift → Fix: Restate coaching headings; enforce “one question at a time.”
Issue: Weak numbers → Fix: Pre-write 3 CARL stories with concrete metrics (%, absolute values, time saved).
Issue: Multi-part questions → Fix: Add “do not ask multi-part questions” to Negative Prompt.

Safety & Compliance

Do not share personal data beyond resume details you choose to provide.
Do not copy proprietary job descriptions beyond fair-use excerpts needed for practice.
Keep claims verifiable; avoid discriminatory or biased assumptions.
Respect NDAs; redact confidential metrics and names as needed.

TL;DR

Paste this job interview coach prompt to run realistic interviews.
Get five-part CARL coaching after each answer with scores.
Use raw links and low temperature for factual accuracy.

FAQ

Q: Can this be used for technical interviews?
A: Yes. Set “technical” as the interview type and include role-specific skills (e.g., system design, coding, notebooks).

Q: How do I keep answers concise?
A: Use CARL: 1–2 lines of Context, 3–4 lines of Action, 1–2 lines of Result with numbers, 1 line of Learning.

Q: What if the AI invents company facts?
A: Add “raw links only; no unverifiable claims” and enable browsing. Reject any statement without a source.