50 ChatGPT Prompt Examples You Can Copy and Paste (2026)

Act as a line editor. Edit the draft below for clarity while protecting my voice. Run a grammar check, fix sentence structure and paragraph flow, cut filler words, prefer active voice, and simplify writing wherever a sentence runs past 25 words. Keep my tone and style. Return the edited draft, then list the 5 biggest changes and why.
Draft: [paste draft]

What it does: Edits instead of generating, so the result still sounds like you. It is the editing prompt worth reaching for most often.

Pro tip: Paste one paragraph you love and add "match this rhythm." The edits stay closer to your natural voice.

You are a B2B copywriter. Write a cold outreach email of 120 words max to [audience, e.g. clinic owners] about [offer]. Open with their pain point, mention one specific detail about their work, make one ask, and end with a closing question they can answer in one line. Give 3 subject line options, each under six words. Keep it a professional email, not a sales pitch.

What it does: Forces a hook, a single CTA, and a subject-line limit, the same before/after fix shown earlier.

Pro tip: Describe [offer] as an outcome. "Cut no-shows by 20%" beats "our booking software" every time.

Here is a tricky email I need to send: [paste draft or thread]. The relationship: [boss / client / colleague]. Do a tone rewrite in 3 lengths (2 lines, 5 lines, full) and 3 tones (warm, neutral, firm). Change tone only; keep every fact and deadline. Flag any sentence that could read as passive-aggressive.

What it does: One reusable prompt covers decline emails, apologies, customer support replies, and push-back messages.

Pro tip: The same approach works for a cover letter. Paste the job description and compare the warm and firm versions side by side.

Create a blog outline for "[topic]" aimed at [audience]. Give an H2/H3 section structure, a target length of 150-200 words per section, and real examples to use in each section. Then write 5 hook lines for the intro. After I approve the outline, repurpose it into: 3 LinkedIn posts, a carousel script, a 60-second video script, a thread, and a newsletter opening.

What it does: One of the most reusable prompts for content creation. You outline first, draft second, then turn one idea into six assets.

Pro tip: For writing a book, swap "blog outline" for "chapter outline" and draft one chapter at a time. Outline-first is why ChatGPT prompts for writing a book work at all. For stronger openers, see viral hooks.

Write 10 headline options for [article or product] aimed at [audience]. Use 5 different headline formulas (how-to, number, question, curiosity gap, benefit-first), 2 per formula. Keep each under 60 characters, include one hook word per headline, and mark the 3 you would test first with a one-line reason.

What it does: Makes each option use a different structure, so you do not get ten versions of the same line.

Pro tip: Ask it to score each headline 1-10 for clarity and curiosity. Keep the ones that score high on both.

Write sales page copy for [product] using problem, solution, results. Audience: [who]. Include one proof point per claim, answer the top 3 FAQ objections, and close with one CTA. Then compress the whole page into an email blurb of 150 words for my newsletter.

What it does: Builds a full page and a short promo from one brief, so the claims stay consistent.

Pro tip: Paste real customer quotes as proof. Never let the model invent testimonials or numbers.

My essay topic: [topic]. Here is my draft: [paste]. Act as a strict tutor. Challenge my thesis, then check evidence, counterarguments, paragraph development and the conclusion. Do not rewrite anything. List what is weak, suggest a different approach where needed, and note missing [citation style] formatting.

What it does: Gives feedback without writing for you, which keeps the work yours and plagiarism-free.

Pro tip: Add "ask me one question before each suggestion" so you actually learn from it.

Act as a senior recruiter. Here is one vague resume bullet: [paste]. Before rewriting, ask me up to 5 questions to uncover metrics (numbers, scale, time saved, revenue). Then write 3 versions matched to this job posting: [paste]. Keep every claim honest and mark anything you had to assume.

What it does: The most useful of the resume prompts. It asks you questions before it writes, so the metrics are real.

Pro tip: Run it on one bullet at a time. Whole-resume rewrites come back generic.

Write a cover letter for [role] at [company]. Job description: [paste]. Open with one specific problem the company faces [from the job post, news or product], connect it to one achievement of mine [paste], and keep it under 250 words. Do not open with "I am writing to apply."

What it does: Leads with the company's problem instead of your biography, which recruiters notice.

Pro tip: In the same chat, ask for 5 LinkedIn headline options so your job-search story stays consistent.

Help with interview preparation for [role]. Draft a 90-second answer to "Tell me about yourself" (about 220 words) using: current role, one proof point, why this job. Then reshape these two stories [paste] into STAR behavioral answers and suggest 3 questions I should ask the interviewer.

What it does: Takes timing into account. It gives you an opening statement plus STAR stories in one pass.

Pro tip: Read it aloud with a timer. If you go over 90 seconds, ask it to "cut 15% without losing the proof point."

I am negotiating an offer for [role] in [city]. Before writing anything, help me define my BATNA, my walk-away point, and a market salary range (tell me which sources to check; do not invent numbers). Then list the objections I will likely hear, rank my concessions, and only then draft the negotiation email.

What it does: Sets your leverage before any wording, the same backwards method described above.

Pro tip: Check the salary range yourself before you send anything. Models sound confident about pay data that may be out of date.

Draft a short, gracious resignation letter to [manager] with my last day [date] and an offer to help with handover. Then create a first 90 days plan for my new role as [role]: goals for days 1-30, 31-60 and 61-90, plus the 5 people I should meet first.

What it does: Handles the ending and the new start together, because every career move involves both.

Pro tip: Keep the resignation letter under 120 words. The details belong in a conversation, not in writing.

Write 5 LinkedIn headline options for a [role] targeting [next role], each under 220 characters. Then write a recommendation letter template I can send to [referee], with placeholders for [3 achievements] and the role I am applying for, under 200 words.

What it does: Makes it easy for a referee to say yes, and keeps your profile aligned with your applications.

Pro tip: Say who you want the headline to attract, recruiters or clients. The wording is very different for each.

Give me 5 business ideas that fit [my skills, budget and market]. For each: a SWOT analysis with non-generic items, the riskiest assumption, and kill criteria (the result that would make me drop it within 30 days). Rank them by how cheaply I can test them.

What it does: Turns brainstorming into a filter, so you only keep business ideas that survive being tested.

Pro tip: Add "reject any idea that needs more than [amount] to test." It removes the fantasy options.

My product: [description]. Current price: [$X], set with cost-plus. Compare that against value-based pricing for this customer persona: [persona]. Propose a 1.5x price test for one segment, the metric that proves it worked, and how long to run it.

What it does: Treats pricing as an experiment you can measure, not a guess.

Pro tip: Include your current conversion rate. Without it, the model cannot size the test sensibly.

Write a 20-minute customer discovery interview script for [audience] about [problem]. Open questions only, no pitching. After I paste my interview notes [paste], turn them into a customer FAQ and a one-sentence value proposition.

What it does: Takes you from conversations to positioning without losing what customers actually said.

Pro tip: Ask it to quote customers' exact words in the FAQ. Their phrasing is your best copy.

Browse these competitor URLs: [url 1], [url 2], [url 3]. Build a comparison table of positioning, pricing, target audience and one weakness each. Use only claims you can see on the pages and mark anything unverified. Then suggest where our value proposition can differ.

What it does: Bases the comparison on pages the model actually read, not on memory.

Pro tip: Turn browsing on first. If it is off, the model will fill the table from memory.

Using this product info [paste], write a one-page business plan, then a 10-slide pitch deck script (one headline and 3 bullets per slide). Keep every number consistent with my 24-month financial model [attach]. Finish with a 90-day go-to-market plan and quarterly OKRs.

What it does: Builds every investor document from one source, so the numbers never disagree.

Pro tip: Reuse the same chat for your investor update email. The context carries over.

Turn this business process into an SOP: [describe the steps]. Format: purpose, owner, tools, numbered steps, quality checks, common mistakes. Then list which 3 steps could be automated first and what tool would do it.

What it does: One of the best ChatGPT prompts for business operations. Knowledge in someone's head becomes a written process anyone can follow.

Pro tip: Record yourself doing the task and paste the transcript. You will catch steps you forget to mention.

Create a brand voice guide for [brand]. Include one customer persona (goals, fears, words they use), one core angle, 5 voice traits with do/don't examples, and a list of banned phrases. Format it as a one-page table I can paste into every future marketing prompt.

What it does: Gives every later asset the same personality, so nothing reads as filler.

Pro tip: Paste three pieces of past copy you liked. The guide gets noticeably sharper.

Write 3 ad variations for [product] aimed at [persona], each capped at 30 words and each testing a different hook (pain, outcome, curiosity). Adjust each for Meta and for Google Search limits. Then give an A/B test plan: which single variable changes per round and what result declares a winner.

What it does: Produces ads you can test, not just ads that read well.

Pro tip: Change one variable per round. Test two at once and the result tells you nothing.

Here is my SEO content brief: [paste]. Extract 15 content ideas from the audience questions in it, then map them into a 30-day social media content calendar as a table: date, platform, format, hook, and which existing asset it repurposes.

What it does: Turns one piece of research into a month of posts.

Pro tip: Ask for a "repurpose from" column. It stops the calendar inventing brand-new work every day.

Write a LinkedIn post under 80 words about [one idea]. One idea only, first line under 12 words, no hashtags in the body, end with a question. Then write a cold outreach message under 100 words to [name, role] that references this specific thing about their work: [detail].

What it does: Uses tight word limits to force one clear idea and a personal reference.

Pro tip: If you cannot fill in [detail], do not send the message yet.

Build a product launch plan for [product] launching on [date]: pre-launch, launch day, and 2 weeks after. Then write the product description from benefits outward (benefit, proof, feature) and outline a 5-email nurture sequence that warms subscribers across 3 weeks.

What it does: Puts the launch steps in order so launch day converts attention you already built.

Pro tip: Give it your list size and past open rate so the plan matches your real audience.

Write 3 meta descriptions for a page about [topic]. Primary keyword: [keyword]. Rules: 140-155 characters, keyword placement within the first 60 characters, one clear reader benefit, active voice, no quotation marks. Show the character count after each.

What it does: Sets character limits and a reader benefit, so the copy cannot drift into fluff.

Pro tip: Check the character counts yourself. Models regularly miscount.

Here is an error and the code around it: [paste error, logs and code]. Do not write a fix yet. First explain the most likely root cause, list 2 other possible causes, and tell me which values or logs to check to confirm. Only after I confirm, propose the smallest fix.

What it does: You understand the bug before anything changes, so it does not come back.

Pro tip: Always attach the logs. Debugging with logs cuts out most of the guessing.

Explain this function line by line for a developer new to the codebase: [paste]. For each line, say what it does and why it might be there. Then list any hidden logic, side effects or assumptions that are not documented.

What it does: A narrow scope gives a better explanation than a summary of the whole codebase.

Pro tip: Keep it to one function per request. Bigger chunks get skimmed.

Refactor this code for [readability / speed / safety; pick one]: [paste]. First write unit tests that capture the current behaviour. Then refactor in small stages, show a diff per stage, and explain the tradeoff each change makes.

What it does: Replaces a vague "improve code" with one goal and a safety net of tests.

Pro tip: Run the tests yourself between stages. Do not accept "all tests pass" without seeing it.

Act as a senior [language] engineer doing a code review. Review this diff: [paste]. Group comments by severity (bug, risk, naming, style). For each, quote the line, explain the issue and suggest a fix. End with the 3 changes you would insist on before merging.

What it does: The persona brings up naming and risk issues a generic review misses.

Pro tip: Tell it your team's conventions first, or it will enforce its own.

Write a regex for [language or flavour] that matches [pattern description]. It must match: [list]. It must NOT match: [list, including the strings currently failing]. Explain each part of the pattern and give 3 edge cases I have not listed.

What it does: Real passing and failing strings make it test the pattern instead of guessing.

Pro tip: Paste the result into a regex tester before trusting it.

Design a REST API endpoint and database schema for [feature]. Constraints: [database, auth method, expected traffic, existing tables]. Give the endpoint spec (method, path, request, response, errors), the table definitions, and tradeoff notes for 2 alternatives you rejected.

What it does: Asks for the design decisions and their tradeoffs, not just output.

Pro tip: Use the same pattern when you convert code between languages: "explain what changes in behaviour."

Imagine it is [date] and my project "[project]" has failed. List the 10 most likely reasons, ranked by probability. For the top 3, give one early warning sign and one action I can take this week to prevent it. Then turn the result into goals for a 30-day plan.

What it does: Starting from failure shows the risks a normal to-do list hides.

Pro tip: Say who is involved and what the budget is. Most real failures come from people and money.

Turn these raw meeting notes into a meeting summary: [paste]. Format: 3-line summary, decisions made, action items as a table (task, owner, deadline), and open questions. One line per bullet. If an owner is missing, write "UNASSIGNED" instead of guessing.

What it does: Turns messy notes into accountability, with names attached.

Pro tip: Add "draft a follow-up email to attendees" at the end. It saves a second prompt.

Here are the subject lines and first lines of my unread emails: [paste]. Sort them into urgent today, this week, delegate and archive, with a one-line reason each. If two requests collide, build a decision matrix with weighted criteria (impact, deadline, effort) and recommend which goes first.

What it does: You decide what comes first with reasons, not guilt.

Pro tip: Remove personal details before pasting email content into any AI tool.

I use [Excel / Google Sheets]. Columns: [A = ..., B = ..., C = ...]. I need a formula in [column] that returns [expected output] when [condition]. Give the formula, a plain-English breakdown of each part, and one test row to check it.

What it does: Describing the columns and the expected output first means the formula works on the first try.

Pro tip: Add a few sample rows, including an empty cell. Most formulas break on blanks.

Friday review: here is what I planned and what actually happened this week: [paste]. Summarise wins, misses and their causes. Then draft Monday's weekly plan with 2 deep-work blocks, and a 1:1 prep note for [manager] with a status update draft and 2 questions to raise.

What it does: Turns a weekly routine into one repeatable prompt.

Pro tip: Save it as a template and run it at the same time every Friday.

I am learning [topic]. Here is my explanation in my own words: [explanation]. Act as a Socratic tutor. Do not correct me directly; ask one question at a time that exposes gaps in my explanation. If I get stuck twice, give a different explanation using an analogy, then ask me to teach it back.

What it does: Finds your weak spots by making you explain, not by showing you answers.

Pro tip: Say your exam date. It changes how hard the tutor pushes.

Explain [concept] like I'm 5. Then explain it again for a 25-year-old with no background. Then explain it at my level: [e.g. first-year engineering student]. End with a formula breakdown (if there is one) and 3 real-life uses.

What it does: Takes you from beginner to expert in three steps, grounded in real uses.

Pro tip: Notice where the 25-year-old version loses you. That is the part to study.

Quiz me on [topic or chapter] for [exam]. Ask 10 practice questions one at a time, mixing recall and application. Wait for my answer before revealing the correct one, then check my work step by step. At the end, predict 3 likely exam questions and list the common mistakes I made.

What it does: Makes you recall answers instead of rereading, which is what actually sticks.

Pro tip: Paste your syllabus so the predicted exam questions match your course.

Turn this vocabulary list or these notes into Anki flashcards: [paste]. Output CSV with columns: front, back, example sentence, tag. Add a mnemonic for the 5 hardest items and a simple concept map (as a text outline) that compares and contrasts the related ideas.

What it does: Gives you flashcards you can import straight into Anki.

Pro tip: Ask for one fact per card. Cards with two facts get forgotten.

My goal: [skill] by [deadline]. I have [X] minutes a day. Build a 30-day learning plan as a table: day, daily action (under [X] minutes), resource type, and a checkpoint every 7 days. Include a review schedule that brings earlier material back on days 3, 7 and 21.

What it does: Turns a goal and a deadline into daily actions, with review built in.

Pro tip: Ask it to leave two buffer days. Plans without slack break by week two.

I need to decide [decision]. Before researching, list the 5 research questions whose answers would actually change my choice. For each, suggest the best type of source, a method to answer it, and what is out of scope so the topic stays bounded.

What it does: Aims your research at the decision you have to make, not at collecting everything.

Pro tip: Drop any question whose answer would not change what you do.

We will do this in 3 stages. Stage 1: build a structured literature review on [topic] from these sources [paste or attach], grouped by theme, one line per source. Then stop and wait for me. Stage 2: list the research gaps these sources do not cover. Stage 3: critique the methods of the 3 strongest papers.

What it does: Shows what long, complex prompt examples should look like: separate stages with stops between them.

Pro tip: Review each stage before saying "continue." That is where you catch mistakes early.

Here are my primary sources: [paste] and secondary sources: [paste], kept separate. Run a CRAAP test (currency, relevance, authority, accuracy, purpose) on each as a table, scoring each criterion out of 5. Then run a bias check and suggest search terms for evidence that is missing.

What it does: Checks your sources before you build conclusions on them.

Pro tip: Open every source it flags as weak yourself. The score is a starting point, not a verdict.

Create an annotated bibliography in [APA 7 / MLA / Chicago] for these sources: [paste]. For each: full citation, a 3-sentence summary, one limitation, and how it supports my argument about [topic]. Flag any source with incomplete citation details instead of filling them in.

What it does: Naming the citation style up front means you do not have to reformat afterwards.

Pro tip: The "flag, don't fill" rule stops made-up page numbers and DOIs.

Document this research process so a colleague can reproduce it: [paste the prompts used, sources and steps]. Then convert the findings into a 10-slide presentation outline for [audience]: slide title, 3 bullets, and the one chart each slide needs.

What it does: Makes your research repeatable and ready to present.

Pro tip: Say who the audience is. Executives want the conclusion on slide 1, academics want the method.

Suggest gift ideas for [recipient: age, relationship]. Their hobbies: [list]. Budget: up to [amount]. Needed by: [date]. Give 3 tiers by price (under [X], [X-Y], splurge) with 3 ideas each, and explain why each fits this person specifically.

What it does: Three details and price tiers make the ideas feel handpicked, not scraped.

Pro tip: Add "nothing they'd already own if they love [hobby]." It removes the obvious picks.

Make a 5-day meal plan using what is already in my fridge: [list]. Dinners under 30 minutes, [dietary needs]. Add a short shopping list for missing items with estimated cost, and note which leftovers become the next day's lunch.

What it does: Plans around what you already have, which cuts waste and grocery spending.

Pro tip: The same constraints-first pattern works for a 4-week workout plan or a trip itinerary grouped by neighborhood.

Here are my loose brainstorm notes: [paste]. Turn them into a creative brief (genre, tone, audience, what readers should feel), then a story outline with a clear narrative structure: setup, 3 turning points, climax, resolution. List 3 worldbuilding questions I still need to answer.

What it does: Moves you from a loose brainstorm to a structured story without losing your ideas.

Pro tip: Answer the worldbuilding questions yourself. That is where your story becomes yours.

Write a [wedding toast / retirement speech] for [person], from me, their [relationship]. Details: [3 stories or traits]. Length: about 2 minutes spoken. Give 2 variants, one heartfelt and one funny. Mark the pauses and avoid inside jokes that need explaining.

What it does: Gives you two tones to choose from, or to mix.

Pro tip: Need a matching image? Ask it to write an image-generation prompt with the lighting and style you want.

Act as a prompt engineer. Score my prompt below from 1 to 10 against this checklist: role, context, constraints, format, audience, success criteria. List the gaps. Then rewrite it to fix every gap and explain each change in one line. Finally, give 3 alternative approaches with different framings.
My prompt: [paste]

What it does: Checks any AI prompt example against the checklist before you spend time running it.

Pro tip: Run it on your own saved prompts once a month. The gaps it finds are usually the same ones each time.