AI Prompts for Interview Preparation

AI prompts for interview preparation work when you feed real job facts and demand a fixed artifact: mock interview script, STAR story bank, role research brief, or thank-you email. You leave with labeled ROLE / TASK / FORMAT shapes you can paste into GPT-5.5 Instant, GPT-5.6, Claude Sonnet 5, or Gemini 3.5 Flash, plus honesty rules that block invented employers, fake metrics, and company claims you never verified. This guide is for job seekers prepping for hiring loops. It is not a prompt-engineering job interview cheat sheet. Soft tip: PromptMake /text can turn a rough prep ask into a structured scaffold at https://promptmake.net/text before you paste into your chat model.

Who AI prompts for interview preparation help

You have a posting, a resume fact sheet, and a calendar invite, and you need the model to turn those inputs into drills you can rehearse out loud. The patterns fit people who freeze on behavioral questions, career changers who need STAR stories from transferable work, and managers who want a shared prep template for reports who interview. Students with thin work history still need the honesty rule: the model can reframe projects and coursework. It cannot invent years of employment.

Skip this page if you want prompt-engineering interview questions for a PE role, coding whiteboard solutions, or salary negotiation scripts as the main goal. Those jobs need different prompts. This article stays on hiring prep for a target role: company and role research, behavioral and situational drills, STAR story drafting, and short follow-up emails after the call.

Treat the model as a rehearsal partner with constraints. You supply the posting, your facts, the interview format you know, and banned claims. The model proposes questions, tighter story structure, and email drafts. You reject anything you cannot defend with a real example from your career.

Core prompt pattern for interview prep

Strong AI prompts for interview preparation give four inputs before any tone request: the role target, your proof facts, the interview format, and the honesty boundary. Role target is the posting, company name, and level you chase. Proof facts are employers, titles, dates, tools, and metrics you own. Interview format covers screen, hiring manager, panel, or case style if you know it. The honesty boundary forbids invented employers, titles, dates, degrees, metrics, and company facts you did not paste from a trusted source.

Paste those four blocks near the top. Put format rules next. Put banned phrases at the end so they survive a long paste. GPT-5.5 Instant, GPT-5.6, Claude Sonnet 5, and Gemini 3.5 Flash all follow labeled blocks well. Vague "help me prepare for my interview" prompts produce generic fluff because the model has no fact pool, no question type, and no forbidden list.

Aim for one artifact type per thread. Mixing mock interviews and thank-you emails in the same chat muddies length and tone. Run separate passes if you need both. Keep a short PREP_FACT sheet outside the chat: target title, company, posting text, interview stage, known interviewer titles, and your resume facts. Update that sheet when you learn a new detail from a recruiter. Feed it into every prompt that touches prep.

Role, task, format skeleton

Copy this skeleton and fill the brackets with your material:

ROLE: You are an interview coach for [target title] at [company type or name]. You draft from JOB_POSTING and CANDIDATE_FACTS only. You never invent employers, titles, dates, degrees, metrics, or company claims.

TASK: Turn the inputs into a [mock interview script | STAR story bank | role research brief | follow-up email] for [interview stage]. Keep every claim tied to CANDIDATE_FACTS or TRUSTED_SOURCE. Mark gaps with [NEED FACT].

FORMAT: Use the section headers listed under OUTPUT_SHAPE. Short sentences. Bullets preferred for questions and stories. Cap each STAR answer at [N] words unless I raise the limit.

JOB_POSTING: [paste]

CANDIDATE_FACTS: [paste resume notes, projects, metrics you can defend]

TRUSTED_SOURCE: [paste company About page excerpt, product page notes, or recruiter email you verified]

INTERVIEW_STAGE: [recruiter screen | hiring manager | panel | case | unknown]

OUTPUT_SHAPE: [list required sections]

RULES: No soft claims like "passionate leader" or "results-driven." If a fact is missing, write [NEED FACT] instead of guessing. Never invent interviewer names.

REMINDER: Never invent career facts or company metrics. Use [NEED FACT] for gaps.

That reminder line stops confident fiction. Models love tidy numbers and named products. Your rule forces a gap marker you can fill from payroll, dashboards, or the company site.

Paste-ready shapes for mock interviews, STAR, research, and follow-ups

Mock interview pattern adds: "Write 8 questions matched to JOB_POSTING. Mix 3 behavioral, 3 role-specific, 1 situational, 1 closing question I should ask them. For each question, add a 2-line coaching tip that points to which CANDIDATE_FACTS I should use. Do not write full answers yet. Flag any question that needs a fact I did not supply with [NEED FACT]." Run Instant or Flash for the first set. Pick the hard questions. Ask for a second pass that scores your spoken answer against the tip.

STAR story bank pattern:

TASK: From CANDIDATE_FACTS, draft 5 STAR stories mapped to themes in JOB_POSTING (conflict, ownership, learning, impact, collaboration). For each story: Situation (2 sentences), Task (1), Action (3 bullets), Result (1 sentence with a number only if I supplied one). Cap each story at 180 words. If a stronger Result wants a metric I did not give, write [NEED FACT].

Role research brief pattern: "TASK: From TRUSTED_SOURCE and JOB_POSTING, draft a one-page brief: Company snapshot (5 bullets), Product or service (5), Likely pain points for this role (5), Questions I should ask (5), Claims I must not invent (list). Cite only TRUSTED_SOURCE. Mark missing items [NEED FACT]. Ban stock phrases like "innovative culture.""

Follow-up email pattern: "TASK: Draft a thank-you email under 120 words for [interview stage]. Sections: Thanks, one specific topic from NOTES, one proof point from CANDIDATE_FACTS that matches that topic, clear next-step line. Tone: plain, warm, no flattery stacks. Ban words like thrilled, excited, and passionate. No invented meeting details."

Step-by-step interview prep workflow

Use one chat thread per interview stage. Dumping a screen, a panel, and a case loop into one long thread blurs question difficulty and story length. The loop below keeps a stable PREP_FACT sheet while you swap only the stage and any new notes from the recruiter. You spend free ChatGPT, Claude, Gemini, or PromptMake runs on structure, then human time on rehearsal out loud. People who skip the fact sheet ask the model to "prep me" and accept invented metrics that collapse when a hiring manager asks for the dashboard story.

Keep PREP_FACT in plain text outside the chat: target title, company, posting, stage, interviewer titles if known, and your proof facts with numbers you can defend. Update it after each recruiter call. Feed it into every prompt that touches questions or stories. The model should never be your career memory.

Step 1: Build the prep fact sheet

List the role and your proof in raw notes. Example: "Acme, Support Lead, 2022-2024. Tools: Zendesk, Looker. Cut first-response time from 8h to 3h after triage rules. Trained 4 new hires. Target: Customer Success Manager at Beta Co. Stage: hiring manager, 45 minutes." Ugly notes beat polished fiction.

Mark uncertain numbers with a question mark. In the prompt, tell the model to keep those as [NEED FACT] or omit them. Guessing "40% faster" when you meant "faster than before" creates interview risk when someone asks how you measured it.

Step 2: Research the role, then generate questions

Paste TRUSTED_SOURCE first. Ask for the research brief pattern before any mock script. Read the "Claims I must not invent" list out loud. Cross out anything you cannot verify on the company site or in the recruiter email.

Then generate questions:

TASK: From JOB_POSTING and INTERVIEW_STAGE, list 10 likely questions ranked Hard, Medium, Easy. For each Hard item, name which STAR theme from my bank should answer it. Do not write full answers yet. Flag questions that need facts missing from CANDIDATE_FACTS with [NEED FACT].

Use that map to decide which stories to draft first. Partial matches get careful wording. Missing skills stay out of your spoken answers unless you gain them before the interview.

Step 3: Draft STAR stories, rehearse, then audit honesty

Run the STAR bank pattern. Speak each story once with a timer. Take the spoken version back into the chat:

TASK: Audit these DRAFT_ANSWERS against CANDIDATE_FACTS. For each story, reply Keep, Soften, or Remove. Soften means the claim overreaches my facts. Propose a safer rewrite for Soften and Remove lines. Never add new achievements. Cap rewrites at 160 words.

Optional scaffold: open https://promptmake.net/text, describe "interview prep with mock questions, STAR stories, and honesty rules," generate once, then paste your posting and facts into the returned structure. Guests get about three runs per day; free accounts get about five. Use a run to shape the prompt, then finish in your chat model.

Save the winning prompt next to the company name and stage so next week's loops reuse the same wrapper.

Mistakes that wreck AI interview prep

Mistake 1: Asking the model to "invent a strong story about leadership" with no facts. Fiction fails when a panel asks for names, timelines, and tools. Paste controlled facts instead.

Mistake 2: Allowing invented metrics. If you did not supply a number, ban numbers in FORMAT. Fake "increased revenue 30%" fails reference checks and live follow-ups.

Mistake 3: Treating company research as free invention. Paste an About page or product notes as TRUSTED_SOURCE. Ban claims outside that paste.

Mistake 4: Using GPT-5.5 Instant for the final honesty audit on a senior loop. Instant is fine for first question lists. Route the audit pass to GPT-5.6 or Claude Opus 5 when stakes are high.

Mistake 5: Memorizing word-for-word AI scripts. Tell FORMAT to produce talking points and STAR bullets, then rehearse in your own voice. Panels notice canned cadence.

Mistake 6: One universal story bank for every company. Keep a master fact sheet, then run posting-specific passes. The master stays honest; the variants reorder emphasis.

Mistake 7: Skipping follow-up emails until the night after. Draft a shell before the call. Fill NOTES right after. Send within a day while details stay fresh.

Mistake 8: Confusing this page with prompt-engineering interview questions. If you interview for a PE or LLM ops role, use a different library. This guide preps you for hiring loops as a candidate in a named job family.

Model notes for interview prep (mid-2026)

ChatGPT often defaults to GPT-5.5 Instant for fast chat. Instant fits brainstorming question lists, first-pass STAR outlines, and three follow-up email options when you already locked facts. Keep prompts short: ROLE, TASK, FORMAT, FACTS, RULES. Skip long chain-of-thought slogans.

GPT-5.6 (Sol in API naming as of mid-2026) fits harder edit passes: honesty audits, contradiction checks between stories and the fact sheet, and careful mapping of answers to posting requirements without overclaiming. Give goal, constraints, and format. Drop "think step by step" padding on reasoning-class models.

Claude Sonnet 5 handles long pastes of postings plus multi-story banks in one message. Claude Opus 5 fits senior loops when you need a strict Soften/Remove audit. Gemini 3.5 Flash works for quick research briefs and question lists when you want speed. Gemini 3.1 Pro fits denser role research when TRUSTED_SOURCE is long.

Prompting split that holds: Instant and Flash get RTF plus a short few-shot STAR sample when you need a house style. GPT-5.6 and Opus get goal + constraints + format, with an explicit "never invent" rule and a [NEED FACT] token. All of them need your PREP_FACT sheet in the message. None replace a human check of dates, titles, and company claims.

For regulated fields (finance, healthcare, government contractors), keep license IDs and clearance details out of generative edits. Type those from your records. Ask the model for structure help on behavioral stories only.

Build interview prep prompts with PromptMake

Write the rough ask in plain words: target title, interview stage, honesty rule, whether you need a mock script, STAR bank, research brief, or thank-you email. Open https://promptmake.net/text and generate a structured prompt once. Expect labeled sections you can fill with JOB_POSTING and CANDIDATE_FACTS.

Edit company names, interviewer titles, and metrics yourself. PromptMake cannot know your career. Paste the filled prompt into Instant for drafts or GPT-5.6 / Opus for audits. Keep free-tier runs for scaffolding, not five synonym retries of the same weak ask.

Workflow that sticks: PREP_FACT sheet → PromptMake scaffold → fill posting and facts → Instant or Flash draft → GPT-5.6 or Opus honesty audit → rehearse out loud → short follow-up email. Store one template per company and stage so you do not rewrite ROLE and RULES from scratch each week.

FAQ

What are the best AI prompts for interview preparation in 2026?

The best AI prompts for interview preparation lead with ROLE and honesty rules, paste a job posting and a candidate fact sheet, then demand FORMAT for a mock script, STAR bank, research brief, or thank-you email. Add a second audit prompt that marks Keep, Soften, or Remove against your facts. Match Instant or Flash for drafts and GPT-5.6 or Claude Opus 5 for the audit when the loop is senior.

Can AI run a full mock interview for me?

Yes when you name the stage, paste the posting, and ask for questions plus coaching tips before full answers. Speak your replies out loud, then paste a transcript or summary for feedback. The model should score against your CANDIDATE_FACTS, not invent stronger stories. Treat blank-slate "interview me" chats as high risk for generic questions that miss the posting.

How do I prompt AI for STAR method answers?

Paste CANDIDATE_FACTS and ask for Situation, Task, Action, Result with hard word caps. Allow numbers only when you supplied them. Map each story to a theme in the posting, then run an audit that Softens any claim that overreaches your facts. Rehearse the bullets in your own words so you do not sound scripted.

Should I use GPT-5.5 Instant or GPT-5.6 for interview prep prompts?

Use Instant for question lists, first STAR outlines, and multiple email options. Use GPT-5.6 when you need a careful honesty audit, conflict checks against the fact sheet, or tighter mapping to a senior posting. Run the same facts through both only when you measure quality for a recurring interview pipeline.

How do I stop AI from inventing interview stories?

State the ban in ROLE and again in a REMINDER line. Forbid new employers, titles, dates, degrees, certifications, and metrics. Require [NEED FACT] when a stronger Result wants a number you did not give. Follow with an audit prompt that compares DRAFT_ANSWERS to CANDIDATE_FACTS story by story.

Can PromptMake help with AI prompts for interview preparation free?

Yes. PromptMake /text turns a rough interview-prep idea into a labeled prompt you can aim at ChatGPT, Claude, or Gemini. Guests get about three generations per day; registered free users get about five. Fill in your own posting and facts, then paste into Instant for drafts or GPT-5.6 for audits.

How should I prompt AI for a post-interview follow-up email?

Paste short NOTES from the call, one proof point from CANDIDATE_FACTS that matches a topic you discussed, and the interviewer name if you have it. Cap the email at 120 words. Ban flattery stacks and invented meeting details. Ask for two tone options, pick one, and send within a day while the conversation is still fresh.