43 ChatGPT prompts for non-native English speakers to polish interview answers
Here’s the thing nobody tells you when you bomb an interview in your second language. Your grammar was probably fine. Your vocabulary was probably advanced enough. What actually killed the answer was the shape of the answer.
You opened with “I am writing to introduce myself and to express my interest.” The hiring manager’s eyes glazed over. You rambled through three years of context before you ever mentioned a result. You used “we” when you should have said “I.” You trailed off with “and stuff like that” because you couldn’t remember the clean English word for the thing you actually did.
None of that is a vocabulary problem. It’s a structural problem. And ChatGPT is a fantastic structural editor when you give it the right prompts.
This is a library of 43 ChatGPT prompts for non-native English interview answer polish that you can paste straight into ChatGPT, Claude, Gemini, or any LLM. The prompts are grouped by what they fix: grammar and tone, vocabulary and register, STAR/CAR rewrites, confidence and filler, culture-fit polish, and full mock interviews. Each one is multi-line, with role context, constraints, an example input, and a sample output.
If you read it top to bottom you’ll have a complete interview polishing stack. If you skim, jump to the comparison table on the way to the 30-day sprint at the end.
TL;DR - The 3 prompts that fix 80% of bad ESL interview answers:
- Prompt 15 - the STAR rewrite (turn your choppy story into Situation / Task / Action / Result)
- Prompt 23 - the filler and hedge killer (erase “I think maybe,” “kind of,” “actually basically”)
- Prompt 8 - the plain English rewriter (swap translated phrases for native ones) Master those three and most of your interviews get materially better.
Why your grammar is fine, your answers are not
Plain English is a movement, popularized by web usability pioneer John Cutts, that pushes writers to strip out jargon, nominalizations, and words like “leverage,” “utilize,” and “facilitate” in favor of short, direct verbs (plainenglish.co.uk).
Non-native English speakers almost never fail interviews because of grammar. They fail because the answer architecture is wrong.
The IELTS 2024–2025 test performance report shows a composite reliability of 0.97 on the Academic test, with Speaking inter-rater agreement at 0.90 and Writing at 0.92. That means examiners agree very tightly on what scores a 6.5 vs a 7.0 in English. And yet most ESL candidates who score 7.0 or higher in IELTS Speaking still answer interview questions in ways that score them a “passable but forgettable” in real rooms.
Why? Three structural sins show up again and again:
- Translated openings. “I am writing to express my interest in the position of…” sounds like an email, not an answer. Native speakers answer with the verb first: “I led the migration. Three months, zero downtime, four services.”
- Context before claim. You walk through your company’s history, your team’s org chart, and the tech stack before you ever say what you did. The hiring manager is mapping your answer against the role, and you’re making them wait.
- Hedging under pressure. When you don’t have the exact phrase, you fill the silence with “I think maybe we were kind of trying to actually…” That sounds uncertain. Native speakers do it too, but ESL candidates do it more, and it gets read as lack of conviction.
The fix is mechanical. You don’t need a better vocabulary. You need a better shape. The 4-part anatomy below gives you that shape, and the 43 prompts give ChatGPT the instructions to enforce it.
The 4-part ESL answer anatomy
Before we get to the prompts, memorize this skeleton. Every polished answer fits inside it.
- The claim (5–12 words). The single sentence that says what you did, in past tense, with a number if you have one. “I cut API latency from 800ms to 120ms.”
- The context (1–2 sentences). Just enough that the listener can place the work. “We were losing checkout conversions on mobile.”
- The action (3–5 sentences). What you did, not the team. Past tense. First person. Verbs.
- The result (1–2 sentences). A number, a customer reaction, a learning. “Mobile checkout conversion went from 1.8% to 3.4% in six weeks.”
The most common ESL mistake is using only the first two parts. You give the claim, then you dump context, then you ramble. Native speakers under pressure use all four parts, in this order, almost every time. Active Listening, the framework Robert Cialdini breaks down in Influence, asks the listener to mirror this structure back. Hiring managers do the same thing unconsciously.
The prompts below let you paste your messy answer and get back a clean 4-part version. Prompts 1–7 fix the words. Prompts 15–22 fix the shape. Prompts 23–30 fix the delivery. Prompts 31–43 simulate the whole interview.
SECTION 1 - Grammar & tone prompts (1–7)
These prompts turn out clean, plain-English rewrites of your grammar-rough sentences. Use them for any answer, written or spoken prep.
Prompt 1 - The Plain English Polish
Purpose: Takes your unpolished answer and rewrites it in short, native-sounding sentences. This is the workhorse prompt.
You are a senior US tech interviewer who edits interview answers for non-native English
speakers. Rewrite the answer below in plain English.
Constraints:
- Target B2 / C1 reading level.
- Sentences between 8 and 18 words.
- Replace nominalizations with verbs (e.g., "made a decision" -> "decided").
- Keep my facts, numbers, and job titles exactly as I gave them.
- Strip filler: "basically," "actually," "I think maybe," "kind of," "sort of."
- Keep the tone warm and confident, not stiff or robotic.
- Do not change my technical terms.
After the rewrite, list the top 3 things you changed and why, in 1 sentence each.
My answer:
[PASTE YOUR ANSWER HERE]Example output:
“I led the migration of our checkout service from a monolith to four micro-services. We were losing mobile conversions because the legacy stack couldn’t scale. I planned the cut-over in three phases, owned the rollback runbook, and ran the on-call during launch. Checkout conversion on mobile went from 1.8% to 3.4% in six weeks, and p95 latency dropped from 800ms to 120ms.”
Pro tip: Paste 3 answers at once and ask ChatGPT to rank them. You learn faster from a ranked set than from one rewrite.
Prompt 2 - The 5-Tone Slider
Purpose: Lets you choose how formal, warm, or assertive the answer should sound. Based on the “5-tone” voice coaching framework used in executive presence training.
Role: You are a communications coach trained in the 5-tone framework
(Reasonable, Authoritative, Assertive, Persuasive, Inspirational).
For the answer I paste below, give me 5 versions, one for each tone.
For each version, label the tone and mark 1-2 sentences that best show
that tone. After the 5 versions, recommend which tone fits a final-round
US tech interview for a [JOB TITLE] role.
Tone definitions to use:
- Reasonable: balanced, evidence-led, "here is the data."
- Authoritative: expert, calm, "this is the right call."
- Assertive: direct, owns the outcome, "I did X, and Y happened."
- Persuasive: connects answer to the listener's goals.
- Inspirational: paints a future state, motivates action.
My answer:
[PASTE ANSWER]Example output: A table with the same story told five different ways, ending with “For your final round at a US SaaS company, use Assertive for behavioral answers and Reasonable for case questions.”
Pro tip: Don’t pick “Inspirational” for a senior engineer interview. It reads as performative. Save it for sales, founder, or pitch contexts.
Prompt 3 - The Grammar-Only Audit
Purpose: Fixes only the grammar and tense issues. Leaves your voice alone. Best when you like the shape but keep making the same error (articles, prepositions, third-person -s).
You are a copy editor focused on grammar, not style. Edit the text below
for these issues only:
- Verb tense consistency (flag every shift).
- Article use (a / an / the / zero article).
- Preposition choice (in / on / at / to / for).
- Subject-verb agreement.
- Countable vs uncountable nouns.
- Run-on sentences over 30 words.
Do NOT change:
- Word choice.
- Sentence order.
- Tone.
- Technical terms.
Output format:
1) Cleaned text.
2) A short table: | Original | Edited | Rule |.
My text:
[PASTE TEXT]Example output:
Cleaned text: “I worked on the project from January to March.” Table: | Original | Edited | Rule | | “I worked on project” | “I worked on the project” | specific project needs “the” |
Pro tip: Run the same text through Prompt 1 and Prompt 3, then compare. Prompt 1 changes style. Prompt 3 only fixes mistakes. The overlap is the error pattern you keep making. Drill that.
Prompt 4 - The Bullet-to-Speech Converter
Purpose: You wrote your answer as a bulleted list. ChatGPT turns it into a 60–90 second spoken answer. The sweet spot for “Tell me about a time…” behavioral questions.
Convert these bullets into a natural spoken answer of 60-90 seconds
(about 160-220 words at a normal speaking pace).
Rules:
- Use first person, past tense.
- Open with the result or the action, not with context.
- Group related bullets into single sentences.
- Keep my exact numbers and tool names.
- Add a 1-sentence closing that signals what I learned.
Bullets:
[PASTE BULLETS]Example output: A 180-word answer you can read aloud and time with a stopwatch. If you time out at 75 seconds, you’re in range.
Pro tip: Most people over-write. If the answer comes out over 220 words, cut the context bullet. Action and result are what the listener remembers.
Prompt 5 - The Past-Tense Fixer
Purpose: ESL candidates often slip between past, present perfect, and present tense mid-answer. This prompt forces consistency.
Audit the text below. For every verb, label the tense
(simple past, present perfect, present, future, etc.).
For any tense shift inside a single answer, suggest a fix
that keeps the same meaning but makes the tense consistent.
Output a 3-column table:
| Sentence | Original verb | Suggested fix | Reason |
If the shift is intentional (e.g., "I joined Google. I have been here for 3 years."),
flag it as OK and explain when the shift is correct.
Text:
[PASTE TEXT]Example output: A line-by-line audit showing where the tense drifted. Useful for Indian, Chinese, and Japanese speakers whose L1 doesn’t mark tense the way English does.
Pro tip: Behavioral interview answers should be in simple past. Save present perfect for “How long have you been…” questions. Mixed tense inside a story reads as confused.
Prompt 6 - The Article Doctor
Purpose: A laser-focused drill for the single hardest ESL grammar point: articles. The hardest area for Chinese, Japanese, Korean, Russian, and Vietnamese speakers.
Below is a paragraph with article errors (a / an / the / zero article).
For each article error:
1) Show the original.
2) Show the corrected version.
3) State the rule in one sentence
(first mention vs. subsequent mention, generic count noun, etc.).
After the edits, write 3 short practice sentences I can say out loud
that target the same article patterns I keep missing.
Paragraph:
[PASTE TEXT]Example output:
“I joined a startup in 2021. The startup was in Berlin. Ø Founders were from India and Brazil.” Rule: zero article when you mean “the founders of that company” as a class.
Pro tip: Don’t memorize rules. Memorize the 20 patterns that cover 90% of article use. Most ESL article errors collapse to 5 patterns: first mention, subsequent mention, generic plural, fixed expressions (“at work,” “in bed”), and proper nouns.
Prompt 7 - The B2-to-Register Ladder
Purpose: You wrote a perfectly correct B2 answer, but the role needs a higher register. This prompt gives you 3 escalating versions.
The text below is correct B2 English. Rewrite it in 3 versions:
1) Polished B2 (clean, no jargon).
2) C1 business register (tight, professional, native-sounding).
3) Senior IC / staff register (assumes expert audience, drops explanation).
For each version, label reading level (Flesch-Kincaid) and a 1-line note
on what changed.
Text:
[PASTE TEXT]Example output: Same story, three levels of polish. The C1 version is what you want for most US tech interviews. The senior version is for staff, principal, or panel interviews.
Pro tip: Don’t jump straight to C2. The C1 ladder step is where 90% of polished non-native answers live. Going past it makes you sound rehearsed.
SECTION 2 - Vocabulary & register prompts (8–14)
These prompts swap translated-sounding words for native ones, and teach you the small word choices that make an answer sound “right” without being showy.
Prompt 8 - The Translated-Phrase Replacer
Purpose: The single highest-leverage prompt in this library. Replaces phrases that sound like a translation from another language with idiomatic English.
You are a US-based hiring manager. Read the answer below and flag
every phrase that sounds translated from another language.
For each flagged phrase:
- Show the original.
- Show 2 native alternatives.
- Explain the rule in 1 line (e.g., "English uses verbs, not nouns").
- If the phrase is fine for a non-native speaker, say so.
Categories to scan for:
- Nominalizations (made a decision -> decided, performed an analysis -> analyzed).
- Borrowed formality ("I am writing to express my interest" -> "I'm excited about...").
- Direct translations of idioms.
- Over-long sentences from a tense-prominent L1.
- "Please be informed" / "Kindly do the needful" patterns.
My answer:
[PASTE ANSWER]Example output:
Original: “I am writing to express my interest in the position of Senior Engineer.” Native: “I’m excited about the Senior Engineer role.” Rule: English job answers start with the verb or with “I,” not with formal letter openers.
Pro tip: Save your edited answers. After 10 runs, you’ll have your own “translated phrase” list. Drill the top 5.
Prompt 9 - The Power Verb Swap
Purpose: Replaces weak verbs (“worked on,” “helped with,” “was responsible for”) with sharp, specific ones. The Cambridge English “10,000-hour fluency” research shows that high-fluency writers use a much wider verb inventory, not more nouns.
Below is a list of bullet points describing what I did at work.
For each bullet:
1) Identify the weak verb (helped, worked on, was responsible for, was involved in, handled).
2) Replace it with 2-3 stronger, more specific verb options.
3) Note any number I should add to sharpen the claim.
Use the SMART verb rule:
- Specific (cut, rebuilt, owned, shipped, blocked, unblocked, led, designed, wrote, fixed).
- Active voice.
- Past tense.
- 1 verb per bullet, not 2.
Bullets:
[PASTE BULLETS]Example output:
Original: “Worked on the billing migration to Stripe.” Stronger: “Led the billing migration to Stripe” / “Shipped the billing migration to Stripe” / “Owned the cut-over to Stripe.”
Pro tip: Aim for verbs that are visible in a calendar: shipped, led, owned, wrote, rebuilt, cut, blocked. Avoid verbs you can’t see: helped, supported, contributed to, participated in. Hiring managers can’t see the second list.
Prompt 10 - The Industry Jargon Filter
Purpose: You used a phrase that means one thing in your country and another in the US. Or you used a buzzword that hiring managers now roll their eyes at.
Audit the text below for industry jargon. Flag and rewrite any of these:
- Buzzwords: "synergy," "leverage," "circle back," "move the needle," "low-hanging fruit," "rockstar," "ninja," "guru."
- Weasel intensifiers: "very," "really," "extremely," "incredibly."
- Stacked nouns: "customer success team manager lead."
- Vague verbs: "drive," "own," "lead" used without an object.
- Translated honorifics: "Sir," "Madam," "Respected Sir."
For each flag, give the rule in 1 sentence and a 1-line replacement.
Text:
[PASTE TEXT]Example output:
Original: “I drive synergy across cross-functional stakeholders to move the needle.” Rewrite: “I work with design, product, and infra to ship [specific thing].”
Pro tip: If you can’t say what you drove in one number, you don’t actually know what you drove. Jargon hides that. Plain English exposes it.
Prompt 11 - The Active Voice Pusher
Purpose: Converts passive, indirect sentences into active ones with a clear doer. A staple of business-writing training for ESL candidates.
Convert the text below from passive to active voice. Rules:
- The doer becomes the subject.
- The verb becomes strong and past-tense.
- Keep technical terms.
- If a sentence has no clear doer, rewrite it so one is named.
- After the rewrite, list the sentences where you had to invent
a subject and explain how you chose it.
Text:
[PASTE TEXT]Example output:
Passive: “A decision was made to migrate the database.” Active: “The team decided to migrate the database” or “I led the decision to migrate the database.”
Pro tip: When the doer was you, say “I.” When the doer was the team, say “we” and follow it with a sentence that says what you did. Hiring managers want to know what you owned, not what the team owned.
Prompt 12 - The Bilingual Idiom Trap
Purpose: Some of your favorite idioms translate weirdly into English. This prompt catches them.
Read the text below. For every idiom, metaphor, or fixed phrase:
- Flag it.
- If it translates cleanly into English, mark it OK.
- If it doesn't translate, give 2 native English alternatives.
- If the English version would confuse a US hiring manager, say so.
Common L1-idiom traps to look for:
- Spanish: "no está en sus cabales" -> don't translate literally.
- Hindi: "jugnu ki tarah chamakna" -> don't translate literally.
- Mandarin: "拍脑袋" (decide by patting head) -> don't translate literally.
- Japanese: "本音と建前" (honest vs public face) -> don't translate literally.
- French: "avoir le cafard" -> don't translate literally.
- German: "Ich verstehe nur Bahnhof" -> don't translate literally.
Text:
[PASTE TEXT]Example output: “I had butterflies in my stomach but my heart was in my throat” → “I was nervous but committed.”
Pro tip: Most idioms translate badly. Drop them. The interview is not the place to showcase your colorful language. Plain, exact English wins.
Prompt 13 - The Role-Specific Lexicon Builder
Purpose: You have an interview for a [JOB TITLE] role and you want the right vocabulary loaded into ChatGPT’s memory before the mock.
I'm interviewing for a [JOB TITLE] role at a [INDUSTRY] company.
Give me:
1) The 20 most-used verbs in this role (e.g., for PM: prioritized,
scoped, shipped, validated, de-risked).
2) The 15 most-used nouns (roadmap, OKR, PRD, north star, etc.).
3) The 10 overused buzzwords to avoid in this role.
4) The 5 metrics this role is usually measured by.
5) 3 example answers that show the right vocabulary in action.
Keep it specific to the role, not generic.Example output: A tight vocabulary map you can paste into the system prompt of every later prompt. Once ChatGPT knows the role, every answer gets sharper automatically.
Pro tip: Run this prompt once per role. Save the output. When you switch roles, paste a new one. Don’t recycle the same lexicon across PM, SWE, and sales roles - the verbs are different.
Prompt 14 - The Soft-Skill Word Swap
Purpose: “I’m a hard worker” is the #1 thing ESL candidates say that bores US interviewers. This prompt finds better words for soft skills.
Rewrite the soft-skill claims below using more specific, more credible
language. Replace generic claims with observable behaviors.
Examples of bad -> good:
- "I'm a hard worker" -> "I shipped 4 projects in 6 months while doing
on-call."
- "I'm a team player" -> "I mentored 2 juniors through promotion."
- "I'm a fast learner" -> "I picked up Rust in 3 weeks and shipped a
CLI tool."
- "I'm detail-oriented" -> "I caught the data race condition in the
invoice job that QA had missed for 2 sprints."
Claims to rewrite:
[PASTE CLAIMS]Example output: Each generic claim turned into a specific behavior with a number.
Pro tip: Always back a soft skill with a number or a concrete situation. “I’m a hard worker” is opinion. “I shipped 4 projects in 6 months” is evidence.
SECTION 3 - STAR/CAR rewrite prompts (15–22)
STAR = Situation, Task, Action, Result. CAR = Challenge, Action, Result. These prompts turn rambling stories into structured ones. The 4-part ESL answer anatomy from earlier is STAR’s younger cousin. These prompts enforce it.
Prompt 15 - The STAR Rewrite
Purpose: The single most useful prompt in the library. Takes your messy story and forces it into STAR.
You are an interview coach trained on the STAR framework
(Situation, Task, Action, Result).
Rewrite the story I paste below as a tight STAR answer.
Constraints:
- Situation: 1 sentence. Where, when, what was the project.
- Task: 1 sentence. What was MY job, in 1 phrase.
- Action: 3-5 sentences. What I did, in past tense, first person.
- Result: 1-2 sentences. A number, a customer reaction, or a lesson.
- Total length: 90-120 words (about 45-60 seconds spoken).
- Skip empty categories. Don't pad.
- Use active verbs only.
- Do not invent facts I didn't give you.
After the rewrite, score my original story 1-10 on:
- Clarity, Specificity, Ownership, Result.
Story:
[PASTE STORY]Example output:
“S - In Q2 2025, our checkout was losing 18% of mobile users to a 6-second page load. T - I owned the perf budget for mobile. A - I profiled the API, found 3 N+1 queries, added a Redis layer, and pushed the JS bundle down from 480kb to 180kb. I worked with the iOS team to defer non-critical analytics. R - Mobile load dropped to 1.8s, conversion went from 1.4% to 3.1%, and we kept that lift for 2 quarters.”
Pro tip: If your Result sentence is “and I learned a lot,” rewrite. The result is what the company got, not what you got. “I learned” is fine as a final sentence but only after the business result.
Prompt 16 - The CAR (Challenge-Action-Result) Variant
Purpose: CAR is tighter than STAR and works better when the situation is obvious (e.g., a known product failure, a reorg). This prompt picks CAR or STAR depending on the story.
Read the story I paste below. Decide if STAR (Situation, Task, Action,
Result) or CAR (Challenge, Action, Result) fits it better. Explain
the choice in 1 sentence.
Then rewrite the story in the chosen structure. Use my facts and
numbers. Keep it under 120 words.
Story:
[PASTE STORY]Example output: A single structured answer with a one-line note on why CAR was chosen (the situation was already well-known, so explaining it again would waste the listener’s time).
Pro tip: Use CAR for follow-up rounds. The interviewer already knows the situation. STAR for screen interviews where context is missing.
Prompt 17 - The SOAR Variant
Purpose: SOAR = Situation, Obstacle, Action, Result. Useful when the story is mostly about what blocked you and how you unblocked it. Common in engineering and operations roles.
Rewrite the story below using the SOAR framework
(Situation, Obstacle, Action, Result).
- Situation: 1 sentence. Where you were.
- Obstacle: 1-2 sentences. What was blocking the team.
- Action: 3-4 sentences. What YOU did to unblock it.
- Result: 1-2 sentences. Outcome plus what you learned about
removing that class of obstacle in the future.
Word target: 100-130. If you can't fit, cut the Situation.
Story:
[PASTE STORY]Example output: A SOAR-formatted answer where the obstacle is the protagonist of the story, not the project. Hiring managers love obstacle stories because they reveal how you think.
Pro tip: SOAR shines for stories about on-call, incident response, vendor failures, scope cuts, and stakeholder conflicts. Use it.
Prompt 18 - The “5-Second Hook” Opener
Purpose: STAR answers often die because the first sentence is weak. This prompt writes 5 alternate openers and picks the strongest.
Below is a STAR answer. Write 5 alternate opening sentences (the
Situation line). For each, note in 1 line why it works.
Then recommend the strongest one for a [JOB TITLE] interview and
explain the choice.
Constraints for openers:
- Past tense.
- 8-15 words.
- Includes at least one of: a number, a tool name, a stakeholder, a deadline.
- Does NOT start with: "I am writing to," "I would like to," "I am excited
to share," "In this story I will describe."
My STAR answer:
[PASTE ANSWER]Example output:
Opener A: “In Q2 2025, our checkout was losing 18% of mobile users to a 6-second page load.” ✓ Best - number + problem Opener B: “I owned mobile performance for the checkout team.” ✗ No result, no scale Opener C: “When I joined, mobile was our biggest leak.” ✗ Vague
Pro tip: The first sentence is the only sentence the listener is sure to hear. Spend 30% of your writing time on it.
Prompt 19 - The “I vs We” Rebalancer
Purpose: ESL candidates often use “we” defensively to avoid sounding like they’re bragging. The problem: the hiring manager can’t tell what you did. This prompt rebalances.
Audit the text below for first-person singular ("I") vs first-person
plural ("we").
Then rewrite it so that:
- Every "we" sentence that hides YOUR specific work is split into
one "we" sentence and one "I" sentence.
- Every "I" sentence that over-claims team work is softened to "we."
- Keep the total number of sentences roughly the same.
Output:
1) The rewritten text.
2) A small table: | Sentence | Before | After | Why |.
Text:
[PASTE TEXT]Example output:
Before: “We built a new billing system and cut churn.” After: “We rebuilt the billing system over 4 months. I owned the dunning logic and the customer-comms templates. Churn dropped from 6% to 3.4%.”
Pro tip: Use “I” for things you personally did. Use “we” for things the team did. Use “I led” or “I owned” when you were the decision-maker. Hiring managers read “I” as ownership and “we” as collaboration. They need both.
Prompt 20 - The Result Amplifier
Purpose: Your result sentence is weak. “It was a great success.” “The team was happy.” This prompt forces a measurable, business-flavored result.
Below is a STAR or CAR answer. Critique the Result section in
1 paragraph. Then rewrite it with these constraints:
- At least one number (percentage, dollar, time, count).
- Tied to a business metric (revenue, cost, retention, conversion,
latency, NPS, time-to-market, error rate, headcount supported).
- Or tied to a customer outcome (a quote, a CSAT bump, a churn cut).
- If no business metric was given, propose 2 reasonable proxies
and label them clearly as "estimated by me, verify before use."
Result section:
[PASTE RESULT]Example output: A new Result sentence with at least one number and a clear business tie. The “estimated” flag stops you from saying things you can’t defend.
Pro tip: If you can’t measure the result, the story probably wasn’t worth telling. Pick a different story.
Prompt 21 - The “3-Answer” Pattern Library
Purpose: Common behavioral questions need a reusable pattern. This prompt gives you 3 versions of one answer (concise, detailed, follow-up).
For the question "[QUESTION, e.g., Tell me about a time you had a
conflict with a coworker]," generate 3 answer versions based on the
same underlying story I'll paste below:
1) The 60-second version (for screen interviews).
2) The 90-second version (for on-site first round).
3) The 30-second follow-up version (for "tell me more about X").
Each version must:
- Open with the result or the action, not the context.
- Use STAR.
- End with a learning or a forward-looking sentence.
Then note which version is the strongest and why.
My raw story:
[PASTE STORY]Example output: Three timed answers, all true, all different in detail level. The 30-second version is the most under-practiced - it’s what gets you into a director’s “let’s keep talking” pile.
Pro tip: Practice all 3 out loud. Most candidates only practice the 90-second version and fumble the 30-second follow-up.
Prompt 22 - The Metrics Reverse-Engineer
Purpose: You don’t remember the exact number. This prompt reverse-engineers a defensible estimate and labels it.
The story below has no numbers. Based on the situation, action, and
result described, propose 2-3 defensible numbers I could have
generated. For each number:
- Label it as "estimated by me, verify before use."
- Show the rough math that supports it.
- Note what data I'd need to look up to confirm.
Story:
[PASTE STORY]Example output: A list of estimates like “Cut deploy time from ~45 min to ~12 min (estimated - check the CI logs from Q1 2024).” Hiring managers accept labeled estimates. They don’t accept vague “significantly faster.”
Pro tip: Always say “I’ll verify the exact number” out loud if you use an estimate. It buys trust.
SECTION 4 - Confidence & filler prompts (23–30)
ESL candidates hedge more than native speakers. It’s measurable. It’s also fixable. These prompts hunt the hedges, fillers, and “translated humility” patterns that flatten your answers.
Prompt 23 - The Filler and Hedge Killer
Purpose: The single most useful prompt for delivery. Lists every filler word and tells ChatGPT to remove them.
Strip every filler, hedge, and throat-clearing phrase from the text
below. Replace with silence (omit) or with a confident verb.
Fillers to remove:
- "I think," "I believe," "I feel like," "in my opinion."
- "kind of," "sort of," "like," "you know."
- "actually," "basically," "literally," "honestly," "obviously."
- "maybe," "perhaps," "probably," "I guess."
- "to be honest," "at the end of the day," "at this point in time."
- "stuff," "things," "and so on," "etc."
Hedges to soften or remove:
- "I was just," "I only," "I just," "simply."
- "I tried to," "I attempted to" -> "I" + past tense.
- "We were thinking of doing X" -> "We did X."
Do NOT remove:
- Numbers.
- Tool names.
- Stakeholder names.
Output:
1) Cleaned text.
2) Count of fillers removed.
Text:
[PASTE TEXT]Example output: 13 fillers removed. The cleaned text is roughly 25% shorter, sounds twice as confident, and reads as one consistent voice.
Pro tip: Run this on the same text 3 times. The 2nd and 3rd passes catch the fillers you unconsciously use to fill silence. The remaining count is your real filler rate.
Prompt 24 - The Humility Trap Detector
Purpose: You were taught to be humble. That’s fine in a cover letter. It’s lethal in an interview. This prompt flags the cultural humility that US hiring managers read as low confidence.
Read the text below. Flag every phrase that reads as false humility
in a US tech interview. For each:
- Show the original.
- Show a confident rewrite.
- Explain the cultural mismatch in 1 sentence.
Patterns to flag:
- "I was just doing my job."
- "It was a team effort, I didn't do much."
- "I'm not sure if this counts, but..."
- "Sorry, I know this is long."
- "I don't want to brag, but..."
- "I'm probably not the best person to ask, but..."
- "This might be a small thing..."
Text:
[PASTE TEXT]Example output: “Sorry, I know this is long” gets cut. “I’m not sure if this counts” gets cut. The story stays, the apology goes.
Pro tip: US tech interviews reward clear, evidence-led claims. Apologizing for the length of your answer makes the listener feel the length. Confident, brief answers always beat apologetic, long ones.
Prompt 25 - The Sentence-Start Refresh
Purpose: ESL candidates often start every sentence with the same word (“I,” “And,” “So,” “We”). This prompt varies the openings.
Below is an answer. List every sentence and its first word.
Then rewrite the answer so no two consecutive sentences start with
the same word, and no sentence starts with "And" or "So" unless
it's for deliberate effect.
Rules:
- Do not change the meaning.
- Keep the same total word count.
- Note which sentence-start pattern you used most often before
the rewrite.
Text:
[PASTE TEXT]Example output: “I” was the opener in 8 of 12 sentences. After rewrite, the opener distribution is I, The, That, When, By, I, After, We - much more readable.
Pro tip: Varying sentence openers is one of the most underrated fluency signals. Native speakers do it unconsciously. You can do it with a one-pass edit.
Prompt 26 - The “Permission Phrase” Catcher
Purpose: ESL speakers often ask permission to give an answer. “Can I tell you about a project?” “Would it be okay if I share an example?” US interviewers read this as low confidence.
Find every "permission phrase" in the text below. Permission phrases
are sentences that ask the listener for permission to give the answer.
Examples:
- "Can I tell you about a time when..."
- "Should I share an example?"
- "Is it okay if I go into detail?"
- "I don't know if you want me to..."
- "Would you like me to..."
For each:
- Flag it.
- Rewrite as a direct sentence that just gives the answer.
- If the permission phrase is appropriate (e.g., the listener asked
you to be brief), say so.
Text:
[PASTE TEXT]Example output: Every “Can I tell you about…” is cut. The answer starts with the story, not with the request to tell it.
Pro tip: The only time to use a permission phrase is when the interviewer told you to “keep it short.” Then say “Short version?” and stop. Otherwise, just answer.
Prompt 27 - The “I Don’t Know” Confidence Builder
Purpose: You don’t know the answer. The 2026 interview norm is to say so clearly and pivot. This prompt gives you a script.
I'm in an interview for a [JOB TITLE] role. The interviewer just asked:
"[QUESTION]"
I don't know the answer. Write 3 versions of an honest, confident
"I don't know, but here's what I do know" response:
1) The 20-second version (screen interview).
2) The 45-second version (panel interview).
3) The 10-second version (rapid-fire round).
Each version must:
- Open with a clear "I don't know" (no fake confidence).
- Show a structured way to recover (reframe, adjacent knowledge, learning plan).
- End with a forward question to the interviewer (signals curiosity).
Question:
[PASTE QUESTION]Example output: “Honestly, I haven’t used Databricks in production. I have used Snowflake and BigQuery at similar scale, and I picked up Spark on a side project last quarter. I can ramp on Databricks in 2-3 weeks. What kind of workloads are you running on it?”
Pro tip: “I don’t know, but here’s what I do know” is one of the strongest possible interview moves. It signals honesty, structured thinking, and curiosity. Native senior candidates use it constantly.
Prompt 28 - The Pace Calibrator
Purpose: ESL speakers often speak too fast (nervous habit) or too slow (translating in their head). This prompt times your answer.
The text below is the script of an interview answer I plan to speak.
Estimate how long it would take to say out loud at:
1) 130 words per minute (slow, careful).
2) 160 words per minute (normal professional).
3) 190 words per minute (fast, nervous).
For each, give the spoken length in minutes:seconds.
Then recommend which pace is best for a behavioral question
in a US interview, and suggest cuts if the answer is over
90 seconds at the recommended pace.
Script:
[PASTE SCRIPT]Example output: At 160 wpm, this 220-word answer takes 82 seconds. Cut to 180 words → 67 seconds. Recommendation: cut the second context sentence.
Pro tip: The 90-second answer is the universal sweet spot. If your answer is over 2 minutes at normal pace, you have too much context.
Prompt 29 - The Pause Marker Inserter
Purpose: Where you pause changes how your answer lands. This prompt marks where to breathe and where to slow down.
Below is an interview answer. Mark the 3-5 places where a deliberate
2-second pause would make the answer land harder. For each, label
the pause as "emphasis," "transition," or "breath."
Rules:
- Don't add a pause after every sentence.
- Use pauses right before numbers (they pop harder).
- Use pauses right before the Result (the listener remembers it).
- Use a pause when the listener's facial expression suggests they
want to ask a clarifying question.
Text:
[PASTE TEXT]Example output:
“…and the load test failed at 5x production traffic. [PAUSE - emphasis] I spent two days reading the code, found a connection-pool leak, [PAUSE - transition] and the next load test passed at 20x production traffic.”
Pro tip: Pausing is the cheapest confidence trick. A 2-second pause before the result makes the listener lean in. A 0-second pause before the result makes it vanish.
Prompt 30 - The Mirror-Question Prep
Purpose: At the end of your answer, the interviewer might ask a follow-up. This prompt pre-generates the 5 most likely follow-ups and your answers.
I just gave the following answer in an interview. Predict the 5 most
likely follow-up questions a senior US interviewer would ask, and
write a 15-30 second answer for each.
Constraints:
- One follow-up must dig into the number / metric.
- One follow-up must dig into the team dynamic.
- One follow-up must be a "what would you do differently."
- One follow-up must be a "what if your assumption was wrong."
- One follow-up must be a "what's the next thing you built on top of this."
For each follow-up, label it (e.g., "metric drill-down") and
keep the answer under 50 words.
My answer:
[PASTE ANSWER]Example output: Five short answers, one for each follow-up direction. Practice each out loud twice. You’ll never freeze on a follow-up again.
Pro tip: “What would you do differently” is the question most ESL candidates fail. You can say “almost nothing” - that’s a valid senior answer. But you need a real follow-up thought. Prepare it.
SECTION 5 - Culture-fit & “soft skill” prompts (31–37)
US tech culture-fit interviews are subtle. ESL candidates often miss the implicit code. These prompts decode it.
Prompt 31 - The “Tell Me About Yourself” Re-Builder
Purpose: The most common opener, the most-answered question, the most-bungled. This prompt rebuilds it as Present → Past → Future, in 60 seconds.
Rebuild my "Tell me about yourself" answer using the
Present-Past-Future (PPF) framework.
Structure:
- Present (15 sec): What I do now, where, and what I'm focused on.
- Past (25 sec): 1-2 roles or projects that built the skills for
THIS job.
- Future (15 sec): Why this role, why this company, what I'd
contribute in year 1.
Total: 60-75 seconds. End on Future, not on Past.
Constraints:
- No life story before age 16.
- No "I am writing to express my interest."
- No fake enthusiasm ("I am super passionate about...").
- Open with the present, not the past.
My raw self-intro:
[PASTE INTRO]Example output: A 75-second PPF intro that lands the listener on “why this role” instead of “and before that I was at…”
Pro tip: Memorize one Present-Past-Future intro. Change the Future sentence for each role. The Past sentence stays almost the same because it’s the proof of your skills.
Prompt 32 - The “Why This Company” Decoder
Purpose: Generic “I love your mission” answers get ignored. This prompt forces a 3-bullet answer that shows real research.
Generate 3 versions of "Why this company?" for [COMPANY NAME],
given that I'm interviewing for a [JOB TITLE] role.
Each version must include:
- 1 sentence naming a specific product, decision, or public statement
the company made in the last 12 months.
- 1 sentence tying that to my specific experience.
- 1 sentence naming what I'd contribute in year 1.
- A short closer (1 sentence) that signals long-term thinking.
No version may use these phrases: "I love your mission," "I'm passionate
about," "your culture," "your values," "I admire."
Use only verifiable, public facts about the company. If I haven't
given you recent news, flag it and say "verify before interview."Example output: Three versions, each anchored to a real product, decision, or statement. Hiring managers can tell when you’ve done 10 minutes of research vs 10 hours.
Pro tip: The strongest “why this company” answer names something the company got wrong and explains how you’d help fix it. Riskier, but it shows real thinking.
Prompt 33 - The “Greatest Weakness” Honest Rewrite
Purpose: “I’m a perfectionist” is the worst answer in 2026. This prompt gives you 3 honest, growth-flavored options.
Give me 3 honest, growth-flavored answers to "What's your biggest
weakness?" for a [JOB TITLE] interview.
Constraints for each answer:
- The weakness is real (not "I work too hard").
- The answer shows self-awareness: I name the weakness and the trigger.
- The answer shows I'm already working on it: a specific practice
or system I use.
- The weakness is not central to the role I'm applying for.
- Length: 30-45 seconds spoken.
End with a 1-line note on which weakness to pick for which role
(e.g., "for a staff eng role, use the 'too much context' weakness,
not the 'delegation' weakness").
Role: [JOB TITLE]
Seniority: [LEVEL]Example output: Three options, each a real weakness, each with a real practice, each with a role-fit note.
Pro tip: Pick a weakness tied to a strength. “I move fast and sometimes skip docs - I’m now writing a one-pager before every project kickoff.” It reframes the weakness as a controllable trade-off.
Prompt 34 - The US Workplace Norm Translator
Purpose: You said something that means one thing in your country and another in the US. This prompt catches the mismatches.
I'm interviewing for a [JOB TITLE] role at a US company.
Audit the text below for statements that mean something different
in US workplace culture than in other cultures.
Patterns to flag:
- "I will do whatever is needed" -> can read as "I don't know
my boundaries" in US.
- "My manager decided" -> can read as "I don't take initiative" in US.
- "We don't really do conflict" -> can read as "I avoid hard
conversations" in US.
- "I want to learn everything before I start" -> can read as
"I'm slow to ship" in US.
- "I'll just do it" -> can read as "I skip the process" in US.
- "I don't want to bother my manager" -> can read as "I don't escalate"
in US.
- "I'll stay late" -> can read as "I have poor time management" in US.
For each flag, give the cultural reframe in 1 sentence and a US-native
rewrite.
Text:
[PASTE TEXT]Example output: “I want to learn everything before I start” gets reframed to “I learn fast, ship a small version in week 1, and iterate from real feedback.”
Pro tip: The US workplace norm is bounded initiative. Take ownership, but signal awareness of the team’s process. ESL candidates often do one or the other. Aim for both.
Prompt 35 - The “5-Question Culture-Fit Pack”
Purpose: A full set of culture-fit questions with strong, role-fit answers based on your background.
Generate 5 culture-fit questions likely to come up in a final-round
[JOB TITLE] interview at a US [INDUSTRY] company. For each:
1) The question.
2) Why the interviewer is really asking it (the hidden score).
3) A 60-90 second answer template I can fill in with my facts.
4) A "trap" version of the question to watch for, and a different
answer for that version.
Use my background:
[PASTE 5-10 LINES OF YOUR BACKGROUND]Example output: Five questions, each with a hidden-score note and a template. The “trap” version is the version that ESL candidates fall into by default.
Pro tip: The hidden score on “Where do you see yourself in 5 years?” is “will you stay 18 months or 5 years?” The answer should hint at a clear growth path inside the company, not at your personal dream of starting a company.
Prompt 36 - The “Conflict With a Coworker” Story Architect
Purpose: The hardest behavioral question for ESL candidates. US answers want a specific conflict, a specific resolution, and a specific learning. This prompt scaffolds the story.
Build a "Tell me about a conflict with a coworker" answer using the
story I'll paste below.
Rules:
- The conflict is specific (a real disagreement, not a vibe).
- I name the other person's role, not their personality.
- I describe my own actions first, not their failures.
- The resolution is concrete (we agreed on X, we changed Y).
- The learning is about my behavior, not theirs.
- The story ends with the relationship intact or improved.
Length: 90-110 words.
After the rewrite, list 2 things the interviewer will probe
in the follow-up, and 1 sentence for each on how to handle it.
Raw story:
[PASTE STORY]Example output: A clean conflict story that doesn’t sound like you’re throwing anyone under the bus. US culture-fit scoring is brutal on this question. Get it right.
Pro tip: “I realized I was wrong about X” is one of the most credible things you can say in a US interview. Most candidates can’t say it. Practice it out loud.
Prompt 37 - The “Why Are You Leaving” Frame
Purpose: The most dangerous question if you’re currently employed. The wrong answer ends the interview. This prompt gives 3 safe answers plus one risky honest one.
Generate 4 answers to "Why are you leaving your current role?"
for a candidate with this profile:
Current role: [TITLE]
Years there: [N]
Reason for leaving (private, real): [PASTE]
Target role: [JOB TITLE]
For each answer:
- Safe: a generic growth answer.
- Better: a learning-arc answer that names what you outgrew.
- Best: a forward-looking answer that ties to THIS role.
- Honest (risky): a truthful version. Note the risk in 1 line.
Constraints:
- No version may badmouth the current employer.
- No version may say "money" as the primary reason.
- No version may use the phrase "I'm looking for new challenges."
Profile:
[PASTE PROFILE]Example output: Four answers ranked by safety. Use the “Best” version 90% of the time. Keep the “Honest” version in your back pocket for cases where the interviewer explicitly invites candor.
Pro tip: “I’ve learned what I came to learn, and I’m ready to apply it at a different scale” is a near-universal safe answer that doesn’t badmouth anyone.
SECTION 6 - Mock-interview prompts (38–43)
These prompts turn ChatGPT into a real mock interviewer. Use them in voice mode if you can, or paste the response back in for grading.
Prompt 38 - The 30-Minute Full Mock Interview
Purpose: Simulates a complete 30-minute interview. Asks 1 question at a time, waits for your answer, then probes.
You are Sarah Chen, a US-based engineering manager at a mid-size SaaS
company. You are interviewing me, [NAME], for a [JOB TITLE] role.
Rules:
- Ask 1 question at a time. Wait for my answer before continuing.
- Use these 5 question slots, in order:
1) "Tell me about yourself." (PPF, 60-75 sec target)
2) "Walk me through a project you're proud of." (STAR, 90-120 sec)
3) "Tell me about a conflict with a coworker." (STAR, 90-110 sec)
4) "Why this role, why this company?" (3-bullet, 60-75 sec)
5) "Do you have any questions for me?" (I will ask 2-3 questions)
- After each of my answers, give 1-line positive feedback and 1-line
constructive feedback.
- After the interview, give a 1-page written report with scores
(1-10) on: Clarity, Specificity, Ownership, STAR compliance, Tone.
- Then list the top 3 things to work on before the next mock.
Start now. Ask the first question.Example output: A real, paced, 5-question mock interview with structured feedback. Run this twice a week during your sprint.
Pro tip: If you have ChatGPT Voice or Gemini Live, use it. The point of the mock is to hear yourself out loud. Reading your answer in your head doesn’t surface the same problems.
Prompt 39 - The Hiring Manager Persona Switcher
Purpose: Practice with different interviewer personalities: warm, cold, fast-paced, deeply technical.
You will run a 20-minute mock interview as 4 different interviewers,
one after the other. For each, stay in character the entire time.
Persona 1: Warm-coach (encouraging, gives hints, never cold).
Persona 2: Cold-screen (formal, no chit-chat, cuts you off if you
ramble, only asks 3 questions).
Persona 3: Skeptical-staff (pokes holes, asks "why didn't you..."
and "what if your assumption was wrong?").
Persona 4: Rapid-fire (asks 8 questions in 15 minutes, no follow-ups).
For each persona, run 1-2 questions, give 1-line feedback, then move
to the next persona.
I will paste my answers. When I'm done, give a per-persona score
(1-10) and a 1-line note on which persona I'm weakest with.
Start with Persona 1.Example output: Four distinct interview styles in one session. The “Cold-screen” persona is the one ESL candidates fail most. Practice it twice.
Pro tip: The cold-screen persona is realistic for US tech phone screens. Get used to being cut off mid-context. Practice compressing your STAR into 60 seconds.
Prompt 40 - The On-Site Loop Simulator
Purpose: Simulates a 4-round on-site loop with different interviewers, each with a different scorecard focus.
Simulate a 4-round on-site loop for a [JOB TITLE] role at a US tech
company. I will paste my answers. Give me one round at a time.
Round 1: Behavioral (PM / hiring manager). Focus: ownership, leadership
principles, STAR.
Round 2: Technical depth (peer engineer / EM). Focus: technical
decisions, trade-offs, debugging stories.
Round 3: System design / case (skip if role is non-technical).
Focus: structured problem-solving, asking the right questions.
Round 4: Culture-fit (cross-functional partner). Focus: collaboration,
disagreement, growth mindset.
For each round:
- Ask 1-3 questions.
- Probe for 1-2 follow-ups.
- Give a score (1-10) on the round's focus area.
- At the end, give a hire/no-hire recommendation with 3 reasons.
Start with Round 1.Example output: A full on-site loop simulation with a hire/no-hire verdict. The verdict is the most useful part - it tells you whether your package is “yes” or “no” right now.
Pro tip: The on-site verdict is usually determined by the weakest round, not the strongest. If you ace 3 rounds and bomb 1, you don’t get an offer. Find your weakest round and over-practice it.
Prompt 41 - The Tough-Follow-Up Drill
Purpose: You give a great first answer. Now the interviewer hits you with a follow-up. This prompt trains you to handle pressure.
Below is my STAR answer. You are a skeptical senior interviewer.
Generate 4 follow-up questions that would test whether I actually
did the work or whether I'm making it up.
For each follow-up, give:
1) The question.
2) What you're testing (the real reason you're asking).
3) A 30-second "I don't know" recovery if I genuinely can't answer.
4) A 30-second strong answer if I can.
After I paste my answer to each follow-up, give a 1-line score
(1-10) on credibility and a 1-line note on what to tighten.
My STAR answer:
[PASTE ANSWER]Example output: Four probing follow-ups. The “what are you testing” line is gold - it tells you what the interviewer actually wants to hear.
Pro tip: “What was the role of the rest of the team?” is the most common follow-up. If you can’t answer it in 15 seconds, the interviewer assumes the team did the work and you rode along. Prepare this answer for every story.
Prompt 42 - The “Closing Question” Builder
Purpose: “Do you have any questions for me?” is your chance to score points. This prompt builds 5 strong closing questions based on the role.
I'm at the end of a [JOB TITLE] interview at a [INDUSTRY] company.
Generate 5 strong questions I can ask the interviewer.
For each question:
- The question (1 sentence).
- Why it's strong (the hidden signal it sends).
- Why a weak candidate wouldn't ask it.
Avoid:
- "What's the culture like?"
- "What do you like about working here?"
- "What are the next steps?"
- Anything that could be answered by reading the company website.
Role: [JOB TITLE]
Company: [COMPANY]
What I learned in the interview:
[PASTE 3-5 BULLETS]Example output: Five high-signal closing questions. Asking the right question at the close is the difference between “we’ll be in touch” and “we’d like to move to the offer stage.”
Pro tip: “What’s the biggest risk to the team’s roadmap in the next 6 months?” is one of the strongest closing questions in 2026. It signals strategic thinking and curiosity.
Prompt 43 - The “Post-Interview Debrief” Coach
Purpose: The interview is over. Now what? This prompt turns the post-interview fog into a 1-page improvement plan.
I just finished a [JOB TITLE] interview at [COMPANY]. Below is what
I remember from each question, my answer, and the interviewer's
reaction.
Generate a 1-page debrief with:
1) Score per question (1-10) on Clarity, Specificity, Ownership, Tone.
2) The single sentence I should have started with for each answer.
3) The one question where I lost the room, with a 1-line note on why.
4) The follow-up question I should have asked and didn't.
5) A 3-action improvement plan for next time, ranked by impact.
Format: short bullets, no fluff. Maximum 400 words.
My notes:
[PASTE NOTES]Example output: A tight 1-page debrief. The “single sentence I should have started with” is the highest-leverage output - it tells you exactly what to open with next time.
Pro tip: Write your debrief within 30 minutes of the interview, while your memory is fresh. Then run it through this prompt. The combination of fresh notes + ChatGPT structuring is brutal in a good way.
Comparison table - prompt categories vs answer stage vs output
| Prompt range | What it fixes | When to use | Output you’ll get |
|---|---|---|---|
| 1–7 (Grammar & tone) | Wrong tense, articles, nominalizations | Before any answer is sent | Cleaned text + a list of fixes |
| 8–14 (Vocabulary & register) | Translated phrases, weak verbs, jargon | After the answer is clean, before STAR | Native-sounding rewrites + word swaps |
| 15–22 (STAR/CAR/SOAR) | Bad story shape, missing result, “we vs I” | Mid-prep, when story is messy | A 90–120s structured answer |
| 23–30 (Confidence & filler) | Hedges, fillers, false humility, permission phrases | After the answer is structured, before mock | A confident, plain-English version |
| 31–37 (Culture-fit) | US workplace norms, weakness, why-leaving | Day before on-site, final-round prep | 3-answer sets + role-fit notes |
| 38–43 (Mock interview) | End-to-end delivery, follow-ups, debrief | Twice a week during sprint | Scored feedback + improvement plan |
The 30-day “5 interviews” sprint
You don’t get fluent by reading. You get fluent by doing. Here’s the sprint that gets most non-native English speakers from “passable” to “competitive” in 30 days, assuming 1 hour a day.
Week 1 - Build the library.
- Day 1–2: Run Prompt 13 for your target role. Save the lexicon.
- Day 3: Write 3 raw STAR stories from your last 2 years of work. Don’t polish.
- Day 4–5: Run each story through Prompts 1, 8, and 15. Save the polished versions.
- Day 6: Run all 3 stories through Prompts 19 and 23. Save the tightened versions.
- Day 7: Read the polished versions out loud 3 times each. Time them.
Week 2 - Mock interview #1.
- Day 8–9: Run Prompt 38 (full mock) once. Score yourself.
- Day 10: Take the lowest-scoring answer from the mock and run it through Prompts 18, 20, and 25.
- Day 11–12: Run Prompt 39 (4 personas). Focus on the cold-screen persona.
- Day 13: Run Prompt 40 (on-site loop) if the role is on-site.
- Day 14: Run Prompt 43 (debrief) on the week’s mocks. Identify the top 3 gaps.
Week 3 - Drill the gaps.
- Day 15: Pick the weakest gap. Write 3 new stories that target it.
- Day 16: Run each new story through Prompts 1, 8, 15, 19, 23.
- Day 17: Run Prompt 41 (tough follow-ups) on the 3 new stories.
- Day 18: Run Prompt 31 (Tell me about yourself) 3 times with different roles.
- Day 19: Run Prompt 32 (Why this company) for 2 real target companies.
- Day 20: Run Prompts 36, 37, 33. Polish the culture-fit stories.
- Day 21: Run Prompt 43 again. Compare to week 2’s score.
Week 4 - Mock interview #2 + final polish.
- Day 22–23: Run Prompt 38 again. Score yourself.
- Day 24: Run Prompt 39 again. The cold-screen persona should now feel normal.
- Day 25: Run Prompt 30 (mirror-question prep) on your 3 best stories.
- Day 26: Run Prompts 28 and 29 (pace, pauses) on your polished answers.
- Day 27: Run Prompt 43 on the final mock. Compare to week 2 and week 3.
- Day 28: Read all your polished answers out loud, 3 times, in a quiet room.
- Day 29: Run Prompt 22 (metrics reverse-engineer) on any answer that still has a weak result.
- Day 30: Real interview #1.
The “5 interviews” in the title is the goal: 5 real interviews in 30 days. Every “no” is data. Every “not now” is a follow-up. By interview 5, the polished answers feel natural because you’ve said them out loud dozens of times.
Pull quote: “Plain, exact English wins.” - paraphrased from the Plain English movement popularized by John Cutts (plainenglish.co.uk). In 2026, the candidate who speaks in 12-word sentences with clear numbers beats the candidate who uses 25-word sentences with vague outcomes, accent or no accent.
Common mistakes to avoid
1. Using AI to memorize a script. ChatGPT is a structural editor, not a teleprompter. If you read the polished answer word-for-word in the interview, you’ll freeze on the first unexpected question. Use the polished version as the shape, not the script. Practice the shape, not the words.
2. Skipping the metric. A STAR answer without a number reads as a story you made up. Even a labeled estimate is better than no number. Run Prompt 22 if you’re stuck.
3. Over-polishing past C1. A C2 answer sounds rehearsed. A C1 answer sounds senior. Stop one rung below perfect.
4. Using one story for every question. A common trap. If you have only 2 strong stories, you’re one question away from repeating yourself. Build 5–6 strong STAR stories that cover leadership, conflict, failure, technical depth, and a stretch project.
5. Not reading out loud. A polished answer that reads well on screen often stumbles out loud. The IELTS Speaking test’s inter-rater reliability of 0.90 (ielts.org) is based on spoken answers, not written ones. Your delivery is half the score.
6. Ignoring follow-up questions. Most ESL candidates ace the first answer and bomb the follow-up. The follow-up is where the interviewer decides if you actually did the work. Run Prompt 41 to drill this.
7. Saying “I don’t know” with a downward tone. “I don’t know” said with a downward inflection reads as defeat. Said with an upward inflection, followed by a recovery, it reads as honesty. The words are the same. The tone is the interview.
8. Forgetting to ask a closing question. Not asking “Do you have any questions for me?” or asking a weak one costs you the “enthusiasm” score. Run Prompt 42 the night before.
People also ask
Do ChatGPT prompts for non-native English interview answer polish actually work? Yes, but only as a structural editor. Polished answers land harder because they follow the 4-part anatomy (claim, context, action, result). Grammar is rarely the real problem - answer shape is. The prompts in this guide fix shape first, grammar second, vocabulary third.
Will ChatGPT fix my accent? No. Accent is a phonetic and prosodic feature, not a text feature. For accent work, pair these prompts with a speech app like Elsa Speak or BoldVoice (5–15 minutes/day for 8 weeks measurably reduces accent intelligibility on the SpeechAce speaking scale). The Elsa Speak Pro tier includes AI conversation drills that complement ChatGPT’s text drills.
What’s the difference between TOEFL iBT, IELTS, and the Duolingo English Test for interview prep? TOEFL iBT and IELTS are academic tests with high score ceilings (TOEFL iBT max 120, IELTS max 9.0). The Duolingo English Test is a 1-hour adaptive online test scored 10–160 that costs US$70 as of 2026 (Wikipedia, “Duolingo English Test”). For interview prep, the specific test matters less than the speaking band you hit - a 7.0 IELTS Speaking or 24+ TOEFL Speaking means you can use these prompts directly. Below that, layer in Cake or Speak for daily conversation drilling first.
How do I stop hedging in English? The filler and hedge killer (Prompt 23) gives you a count. Run your answer through it 3 times. The hedges that survive are the ones you unconsciously use to fill silence. Drill those, not all hedges. Most ESL candidates have 3–5 personal hedge patterns, not 20.
Is it cheating to use ChatGPT during interview prep? No. It’s the same as using a grammar checker, a mock interview service, or a coach. Hiring managers use AI tools to write JDs and screen resumes. The ethical line is at using AI during the interview (e.g., a hidden earpiece). Using AI to prepare is standard.
How long should my answers be? 60–90 seconds for behavioral questions, 30–45 seconds for follow-ups, 75–90 seconds for “Tell me about yourself.” Anything over 2 minutes is too long. Run Prompt 28 to time your answers at 160 words per minute.
What if I freeze mid-answer? The recovery is “Let me start that more cleanly.” US interviewers respect a reset more than a fumbled finish. Avoid permission phrases (Prompt 26) and silence fillers (Prompt 23).
Should I memorize answers? Memorize the shape (STAR), not the words. The 4-part anatomy is what you memorize. The verbs and numbers change per question. Memorizing words creates a delivery that sounds robotic. Memorizing shape creates a delivery that sounds fluent.
What’s a 10,000-hour fluency model and does it apply to interview prep? Cambridge English’s research on fluency development suggests that high-fluency learners accumulate “deliberate practice hours” - focused, feedback-rich speaking practice - at much higher rates than casual learners. You don’t need 10,000 hours to ace an interview; you need 30 hours of deliberate, prompted practice. That’s what the sprint above is designed to produce.
What about accent bias in 2026? Accent bias remains a documented hiring risk - the Wikipedia entry on linguistic discrimination surveys decades of research showing measurable bias against non-native accents in employment, education, and legal settings. The structural answer in 2026 is two-fold: (1) reduce accent intelligibility (not erase accent) via 5–15 min/day of Elsa Speak or BoldVoice for 8 weeks, and (2) increase answer clarity so the listener’s bias has less surface to grab onto. These 43 prompts target (2) directly. The combination of clearer answers and more intelligible speech is the strongest structural workaround for accent bias currently available to candidates.
Final word
You don’t need a perfect accent to land a US tech job in 2026. You need clear answers in plain English, structured around the 4-part anatomy, delivered in 60–90 seconds, with at least one number in the result. The 43 prompts above are a complete, role-agnostic system for getting there.
Pick three. Use them this week. Then add one more. In 30 days, the candidates who are using prompts are the ones who sound fluent. The ones who are still winging it are still hedging.
You’ve got this.