ChatGPT Prompts for Copywriters: 34 for Landing Pages, Ads & Sales Copy (2026)
34 copy-ready ChatGPT prompts for copywriters: landing pages, ads, sales emails, product copy, headlines, and where AI shouldn't lead.
Researched with AI assistance, reviewed and edited by Tapabrata Biswas.

In this article
- 01How to prompt as a copywriter
- 02Frameworks that make AI copy convert
- 03Headlines and hooks
- 04Landing pages and sales pages
- 05Ad copy that earns the click
- 06Email sales sequences
- 07Product descriptions and ecommerce copy
- 08Value propositions, USPs, and taglines
- 09Voice, tone, and editing
- 10Objection handling and CTAs
- 11Where not to let ChatGPT lead
- 12What this post does not cover
- 13Sources
Most copywriting prompt lists hand you "write me a sales page" and then wonder why the output reads like every other AI page on the internet. The trouble is that copy converts on two things a chatbot doesn't have: your customer's real language and a claim you can actually prove. A vague prompt skips both and gives you fluent hype you then have to gut. ChatGPT earns its place in a copywriter's day not by writing the winner, but by taking the first draft, the variations, and the structural grunt work off your plate: the landing page skeleton, ten headline angles, the email sequence outline, the ad in three frameworks. The prompts below are built for that, each with a clear job, a place to drop your real offer and voice, and a format you can paste and edit.
They run across the assets a conversion copywriter actually ships, from headlines through landing pages, ads, sales emails, and product copy, and they work the same in ChatGPT, Claude, or Gemini. One quick boundary: if you're writing blog posts, captions, and video scripts to build an audience, that's content work, and our prompts for content creators fit better; if you're planning campaigns and positioning, the prompts for marketing do. This is about copy that asks for the sale. There's also an honest section at the end on where AI has no business leading, because in copywriting that line matters more than most.
How to prompt as a copywriter
A prompt that saves you an hour and one that wastes it differ by four things: role, context, task, and format. Tell the model who to be, give it the offer and buyer it can't see, state the exact asset you want, and name a framework so the copy has a spine. Leave the context out and it fills the gap with hype, the problem behind writing a clearer prompt in any field. And one habit specific to copy: paste a real paragraph of your brand voice, because that's the single biggest lever on whether the draft sounds like you.
Here is the difference. The vague version:
Write me a sales page for my product.
And the version that actually helps, with role, offer, audience, framework, and format:
Act as a direct-response copywriter. My offer is [product] for [audience], and the core transformation is [from pain to result]. Here is a paragraph in my brand voice to match: [paste]. Write the top section of a landing page using the PAS framework: a headline, a subhead, and two short problem-agitate-solution paragraphs. Use my customer's language, not marketing hype, and don't invent any statistics or results.
The first gives you slop. The second gives you a real starting draft in your voice. Every prompt below is written in that second style, and the two habits that matter most are pasting your real voice-of-customer language and never letting the model make up a fact or a testimonial.
Frameworks that make AI copy convert
Base ChatGPT has no persuasion methodology, so left alone it writes grammatically clean copy that follows no structure and lands nowhere. Naming a framework fixes that in one line. The four below cover most of what a copywriter ships, and each prompt tells the model when the framework fits so you're not forcing a VSL structure onto a Facebook ad.
Act as a direct-response copywriter. Write ad copy for [offer] to [audience] using the PAS framework. Open on the specific problem in my customer's own words, agitate the cost of leaving it unsolved, then present [offer] as the solution with one concrete proof point I give you: [proof]. Keep it under 120 words, problem-aware reader, no hype adjectives.
Write the opening of a long-form sales page for [offer] to cold traffic using AIDA. Attention: a headline that names the desire or fear, no clickbait. Interest: two lines on why this is different. Desire: paint the after-state concretely. Action: one clear CTA. Here's my offer and audience: [paste]. Don't fabricate benefits I didn't give you.
Write email copy for [offer] using the Before-After-Bridge framework. Before: where my customer is stuck now, in their words. After: the specific result they want. Bridge: how [offer] gets them there. Audience and offer details: [paste]. Keep it to 150 words, one CTA, and a subject line that matches the Before.
Draft a long sales page outline for [offer] using the PASTOR framework (problem, amplify, story, transformation, offer, response). Give me the section-by-section beats with a one-line note on what each must accomplish, so I can write the copy into a proven structure. My offer, audience, and the real customer story I can use: [paste].
Headlines and hooks
The headline does most of the work, so it's worth generating a lot of angles and killing most of them. ChatGPT is genuinely good here, because volume and variation are exactly what it's fast at, and picking the one that's true and specific is exactly what you're for. Give it the offer and ask for angles, not one perfect line.
Act as a copywriter. Write 15 headline options for [offer] aimed at [audience]. Mix the angles: benefit-led, problem-led, curiosity, specific-number, and objection-led. Keep every one honest and specific to my offer, no vague hype like 'unlock your potential', and flag the three you'd test first and why.
Here's my current headline: [paste]. It's underperforming. Give me 10 sharper rewrites that keep the same promise but make it more specific, more concrete, or more clearly aimed at [audience]. For each, tell me in a few words what lever you pulled, so I can learn the pattern, not just copy the line.
Write 10 scroll-stopping first lines (hooks) for a [platform] post or ad about [topic] to [audience]. Each should earn the next line without clickbait or a fake promise. Range from a bold claim to a specific question to a surprising number, and keep them in my voice: [paste sample].
Give me 8 subject-line and preview-text pairs for an email about [offer] to [audience]. Vary the approach: curiosity, benefit, urgency, and plain-spoken. No spammy caps or fake 're:' tricks. Note which two you'd A/B test against each other and what each one is testing.
Landing pages and sales pages
A landing page has to carry one argument from headline to CTA without a wasted line, and that structure is something the model drafts well once you give it the offer and the proof. Feed it real specifics and a framework, and treat what comes back as a skeleton to sharpen, not a page to ship.
Draft a full landing page for [offer] to [audience] using AIDA. Sections: hero (headline, subhead, CTA), the problem, the solution and how it works, benefits tied to my customer's real goals, objection handling, social proof placeholders where I'll paste real testimonials, and a closing CTA. Offer details and voice sample: [paste]. Mark every place I need to add a real proof point.
Write the 'how it works' section for [offer] in three or four steps. Make each step concrete and outcome-focused, not feature-focused, so the reader can picture using it. Here's what actually happens when someone buys: [paste]. Keep it skimmable and in my voice, and don't add steps that aren't in what I gave you.
I need a benefits section that doesn't read like a feature list. Here are my product's features: [paste]. For each, translate it into the specific benefit and the deeper outcome my customer actually wants, using the 'so what' test. Write it as short punchy lines, and cut any benefit I can't back with something real.
Here's my landing page draft: [paste]. Act as a conversion copywriter and critique it: where does it lose the reader, where's the argument weak, where am I talking about myself instead of the customer, and where's the CTA unclear? Give me a prioritised list of fixes, most impactful first, not a full rewrite.
Ad copy that earns the click
Paid copy lives or dies on the first line and a tight promise, and platform limits force the discipline. ChatGPT is useful for generating variations fast so you have something to test, as long as you hold it to real claims and your customer's language. Give it the platform, the angle, and the offer.
Write 5 Meta ad variations for [offer] to [audience]. Each needs a scroll-stopping first line, a short body that names one real benefit, and a clear CTA. Vary the angle across the five (problem, desire, objection, social proof, curiosity). Keep them honest and specific, no 'game-changing' or 'revolutionary', and note the angle for each.
Write 3 Google search ad sets for the keyword [keyword], each with 3 headlines (30 characters max) and 2 descriptions (90 characters max). Match the intent behind the search, lead with the benefit, and include the offer and a CTA. Here's my offer: [paste]. Stay within the character limits exactly.
Write a LinkedIn ad for [B2B offer] to [job title / industry]. Professional but not stiff, lead with the business problem they'd recognise, name the outcome, and keep it credible with no hype. One CTA. My offer and who it's for: [paste].
Write 4 retargeting ad variations for people who visited my [page] but didn't buy [offer]. Assume they already know the product, so handle the likely reason they hesitated: price, timing, trust, or fit. One objection per ad, addressed honestly, with a nudge to come back. Details: [paste].

Email sales sequences
Email is where copy compounds, and a sequence is repeatable enough that the model can draft the arc while you write the lines that sell. Give it the offer, the trigger, and where the reader is in the journey, and it will shape a sensible sequence you can sharpen email by email.
Draft a 4-email welcome sequence for new subscribers who signed up for [lead magnet] and are a fit for [offer]. Map the arc: deliver value, build trust, introduce the offer, and make the pitch. For each email give me a subject line, the goal, and a short draft in my voice: [paste]. One clear CTA per email, and no fake urgency.
Write a 5-email launch sequence for [offer] opening [date] and closing [date]. Cover the arc: announce, benefits and proof, objection handling, social proof, and last chance. Give each a subject line and a draft. Here's the offer, the audience, and the real deadline: [paste]. Keep the urgency honest and tied to the real close date.
Write a 3-email cart-abandonment sequence for [product]. Email one: a helpful reminder. Email two: handle the top objection [objection] and add one real proof point. Email three: a final nudge. Keep them short, human, and low-pressure, in my voice: [paste]. Don't invent a discount I didn't offer.
Write a re-engagement email to subscribers who haven't opened in [time]. Acknowledge the silence honestly, remind them what they signed up for, give one reason to stay, and make it easy to leave. Warm, not needy, in my voice: [paste]. End with a simple choice, not a guilt trip.
Take this one sales email: [paste] and give me 5 subject-line variations and 3 different opening lines to test. Keep the promise consistent with the body so nothing is bait, and tell me what each variation is testing (curiosity vs benefit vs specificity) so the test actually teaches me something.
Product descriptions and ecommerce copy
Product copy has to answer 'why this one' fast, and at scale it's the kind of repetitive writing where AI saves real time. The trap is that it will happily invent specs and benefits, so the rule here is strict: it only writes from the real details you give it. Feed it the facts and the buyer, and let it handle the shaping.
Write a product description for [product] for [audience]. Here are the real features, materials, and specs: [paste]. Lead with the main benefit, weave the specs in as reasons to believe, and match the tone to my brand: [voice sample]. Keep it to [length], and use only the details I gave you, no invented features or claims.
I have [number] products in the same category that need descriptions. Here's the template that worked for one: [paste], and here are the details for the next product: [paste]. Write its description in the same structure and voice, adapted to its real specifics, so the set stays consistent without sounding copy-pasted.
Write the copy for a product comparison or bundle page: [paste the products and real differences]. Help the buyer choose by making the trade-offs clear and honest, guide the likely best fit for each type of buyer, and write a short line for each option. Don't oversell the premium one; make the choice feel fair.
Value propositions, USPs, and taglines
The value proposition is the hardest line to write and the one worth generating options for. ChatGPT can pressure-test how you're positioning the offer and throw angles you hadn't tried, which is useful even when you reject most of them. Give it the offer, the buyer, and what makes you actually different.
Help me sharpen my value proposition for [offer] to [audience]. What we do differently: [paste]. Give me 8 one-sentence value propositions, each making the specific benefit and the point of difference clear in plain language. Avoid category cliches, and flag the two that are most specific and hardest for a competitor to copy.
I'm struggling to articulate my USP because my competitors say similar things: [paste what they say]. Here's what's actually true and different about my offer: [paste]. Help me find and phrase the angle competitors can't claim, and warn me if what I think is a differentiator is actually table stakes.
Write 12 tagline options for [brand/offer] that capture [the core promise] for [audience]. Range from clear and literal to a little more creative, keep them short and memorable, and don't sacrifice clarity for a clever pun. Note which three are the clearest and which one is the most distinctive.
Voice, tone, and editing
The fastest wins with AI in copywriting are often on copy you already have: tightening it, matching a voice, and generating test variants. This is where the model is a genuine force multiplier rather than a first-drafter, because you're steering it with real material. Give it your copy and a clear instruction.
Here are three paragraphs I've written in my brand voice: [paste]. Analyse the voice: sentence length, tone, vocabulary, rhythm, and the habits that make it distinctive. Then write me a short voice guide I can paste into future prompts so any draft comes back sounding like this, not like generic AI copy.
Tighten this copy without losing the meaning or the voice: [paste]. Cut the filler, the hedging, and any hype adjective that isn't earning its place, and make the sentences punchier. Show me the tightened version, then a one-line note on the biggest thing that was slowing it down.
This copy is flat: [paste]. Punch it up without adding hype or false claims. Add rhythm with sentence-length variation, make the concrete details more vivid, and strengthen the verbs. Keep every factual claim exactly as true as it is now, and don't invent specifics to make it sound better.
Rewrite this piece of copy in three distinct tones for the same audience: [paste]. Version one plain and direct, version two warmer and more conversational, version three bolder and more confident. Keep the offer and the claims identical across all three so I'm testing tone, not message.
Objection handling and CTAs
Copy that ignores the reader's real hesitation doesn't convert, and naming objections out loud is a skill the model can help you rehearse. It's also good at generating CTA variations once the argument is made. Give it the offer and the likely reasons a buyer stalls.
List the top 6 objections a [audience] buyer would have about [offer], ranked by how often they'd stop the sale. For each, write one honest paragraph of copy that addresses it without being defensive, and tell me where on the page or in the sequence it should go. Here's the offer: [paste].
Write a FAQ section for [offer] that does real objection-handling work, not filler. Turn the hard questions (price, time, fit, what if it doesn't work) into honest answers that move the reader toward yes. Real details and my refund or guarantee terms: [paste]. Don't promise anything I didn't state.
Give me 10 CTA button and line variations for [offer] beyond 'Buy now' and 'Sign up'. Match the CTA to the reader's state of mind at that point in the page, make the value or the next step clear, and note which two you'd test first. Keep them honest about what happens on click.
Where not to let ChatGPT lead
Here is the honest part, and in copywriting it's the line that protects both your results and your client. ChatGPT is a fast drafter and a poor judge of what's true, so the claims and the research stay yours.
Keep three things in your own hands. The facts, because it will invent benefits, statistics, and even testimonials that read as real and are fabricated, and in most markets an unsubstantiated claim is an FTC problem and a fake review is a serious one, so every number and every quote has to be real and yours. The research, because the line that actually converts usually comes from a real customer review, a support ticket, or a sales call, and the model can't run that mining for you; feed it the voice-of-customer and it shapes it well, ask it to guess and it gives you an average. And the voice, because base AI copy has a recognisable cadence, and if you ship it unedited your pages start to sound like every other page that skipped the editing. Used inside those lines, it clears the first draft, the variations, and the grunt work so you can spend your time on the angle, the proof, and the test. Outside them, it produces confident copy that says nothing true and sounds like everyone else.
What this post does not cover
This is a working set of prompts and how to use them safely, not a copywriting course or a guide to any specific tool like a landing-page builder or an email platform. The prompts are starting points to adapt, and they assume you'll paste your own real offer, audience, and voice, supply the real proof, and edit and test every output. Nothing here replaces the customer research, the judgment, and the craft that actually make copy convert. If prompting itself is new to you, the basics of prompting is the place to start.
Sources
- Prompt engineering guide, OpenAI (the role, context, task, and format principles behind the prompts here)
- AIDA (marketing), Wikipedia (the attention-interest-desire-action model behind the framework prompts)
Frequently asked questions

Written by
Tapabrata Biswas
Tech Researcher
I test AI productivity tools and research home-automation gear the way most people use them. Not in a lab, but on an ordinary desk with an ordinary internet connection. The only test that matters: does it save you time?
Share the Post with Your Besties
Get the plain-English tech brief
One email a week on AI tools and smart-home tech. No jargon, no hype.


