In a survey of professional copywriters, 66% said AI has improved their work and about 78% use it at least monthly — yet 52% still worry about AI displacing paid workSource ↗
Among 688 content professionals, AI adoption rose from 65% to 95% in two years, but 'suggest edits' is now the top use and only about 10% use AI to write a complete piece — and those who use it to write everything report weaker resultsSource ↗
85% of marketers now use AI tools for content creation, per CoSchedule's State of AI in Marketing survey of 1,005 marketing professionalsSource ↗
The U.S. Copyright Office confirmed in January 2025 that purely AI-generated text is not copyrightable — prompts alone do not make you the authorSource ↗
writingClaudeChatGPT

Turning a brief and product facts into a first draft you rewrite

The blank page is the slowest part of most projects, and a first draft is the single most delegated task in copywriting. The trick that separates useful drafts from generic slop is grounding: you feed the AI only the real product facts, the audience, and the voice, so the draft is raw material you sharpen rather than a hallucinated pitch you have to fact-check line by line.

Prompt
You are a drafting assistant to a copywriter working on {{deliverable}} for {{brand}}. Write a first draft I will edit — not final copy.

Brand voice and rules (follow exactly): {{brand_voice}}

Product facts — the ONLY claims you may make (do not add features, benefits, or numbers beyond these): {{product_facts}}

Audience and what they care about: {{audience}}

Draft the {{deliverable}} to this structure and length: {{format}}

Rules:
- Make no claim that isn't directly supported by the product facts above. Where a stronger claim would help, insert [CLAIM TO VERIFY: describe it] instead of writing it as fact.
- Do not invent statistics, testimonials, awards, or "studies show" phrasing.
- Lead with the customer's problem in their words, not with the brand.
- Write in the specified voice. Avoid cliches like "in today's fast-paced world," "unlock," and "elevate."
- After the draft, list every claim a reviewer should substantiate before this runs.

Fill in your details and the prompt updates live — then copy.

creativeChatGPTClaude

Generating headline and subject-line variants for A/B testing

You rarely get the winning line first. Testing needs volume across genuinely different angles, and producing fifteen distinct subject lines by hand is slow and tends to circle the same idea. This is where AI earns its keep as a variant machine — as long as each option stays inside your character limit and makes no claim your facts don't support.

Prompt
You are helping a copywriter produce test variants of {{asset_type}} for {{brand}}.

The offer or message, and the facts behind it (use only these — invent no new benefit or number): {{offer_and_facts}}

Hard constraint: each variant must be {{limit}} or shorter. Count and show the character length of each.

Produce 15 variants, three each across these five angles, and label every one with its angle:
- Direct benefit
- Curiosity / open loop
- Objection or risk reversal
- Social proof (only if the facts include a real proof point)
- Specific / number-led

Rules:
- Every variant must be defensible from the facts I gave you. If an angle has no supporting fact (e.g. no real social proof), write "no supporting fact provided" instead of fabricating one.
- No clickbait that the copy can't pay off. No ALL CAPS, no more than one emoji, no "clinically proven" or superlatives unless the fact supports them.
- Vary sentence shape, not just word swaps.

Fill in your details and the prompt updates live — then copy.

analysisClaudeChatGPT

Building a message map from customer research and reviews

Good copy comes from what customers actually say, not from what the brand wishes they'd say. You've got the raw material — review exports, support tickets, interview notes, voice-of-customer quotes — but synthesizing it into a usable message map is tedious. AI is genuinely strong at pattern-finding across messy text, provided you make it work only from the material you paste and quote it back to you.

Prompt
You are a messaging strategist helping a copywriter. Below is raw customer research for {{product}}. Build a message map using ONLY this material.

Audience segment in focus: {{segment}}

Research (reviews, interview notes, support tickets, survey answers): {{research}}

Produce:
1. Primary value proposition — the one benefit customers themselves emphasize most, with two supporting quotes pulled verbatim from the research.
2. Three supporting messages — each with the customer language behind it and a real proof point from the research (not from your own knowledge).
3. Top three objections or hesitations, quoted from the research.
4. Voice-of-customer word bank — the actual phrases customers use, so the copy can mirror them.

Rules:
- Use only the pasted research. Do not add benefits, statistics, or objections that aren't in it.
- Attribute each quote to its source line so I can verify it.
- Where the research is thin on something (e.g. no pricing objections appear), write [NOT IN RESEARCH] rather than filling the gap.

Fill in your details and the prompt updates live — then copy.

communicationChatGPTClaudeGemini

Adapting one approved message into channel-native copy

A campaign lives across email, social, ads, and a landing page, and each channel has its own length, tone, and rules. Once the core message is approved, rewriting it five ways is exactly the mechanical adaptation AI handles well — the claim set is already signed off, so the risk of fabrication is low and your job becomes matching each channel's native voice.

Prompt
You are helping a copywriter repurpose one approved message across channels for {{brand}}. The claims below are already approved — keep the claim set identical everywhere; adapt only tone, length, and format.

Approved core message and claims: {{core_message}}

Brand voice: {{brand_voice}}

Call to action: {{cta}}

Produce channel-native versions, respecting each limit:
- Email: subject line (under 50 characters) + preview text (under 90) + 100-word body
- LinkedIn post: 120-180 words, no hashtags stuffed at the end
- Google Search ad: 3 headlines (30 characters each) + 2 descriptions (90 characters each)
- Landing page hero: headline (under 60 characters) + subhead (under 120)

Rules:
- Do not introduce any claim, number, or benefit not in the approved message. If a channel needs a proof point I didn't provide, mark [NEED PROOF] rather than inventing one.
- Match the platform's native rhythm — LinkedIn is conversational, search ads are terse and literal.
- Show the character count for every length-limited field.

Fill in your details and the prompt updates live — then copy.

planningClaudeChatGPT

Reverse-engineering a creative brief from a messy kickoff

Half of copywriting problems are actually brief problems: a client sends rambling kickoff notes, a Slack thread, and a call transcript, and somewhere in there is the real objective. Turning that mess into a structured brief is pattern work AI does quickly, and forcing it to mark what's missing surfaces the questions you need to ask before you write a word.

Prompt
You are helping a copywriter turn a messy client kickoff into a structured creative brief for {{project_type}}.

Raw input (call notes, emails, Slack, transcript): {{kickoff_notes}}

Produce a one-page brief with these fields, using only what's in the input:
- Objective (what this copy must achieve)
- Audience (who it's for, in their situation)
- Single most important message
- Proof points (only those stated in the notes)
- Tone and voice
- Mandatories and legal/compliance notes (required disclaimers, claims that need substantiation, brand rules)
- Deliverables and formats
- Primary call to action
- Success metric

Rules:
- Fill each field only from the notes. Where the notes don't answer a field, write [ASK CLIENT] — never guess an objective, metric, or audience.
- At the end, list the 5 most important open questions to resolve before drafting.
- Flag any claim in the notes that will need substantiation before it can be used in copy.

Fill in your details and the prompt updates live — then copy.

automationClaudeChatGPT

Codifying a brand voice into a reusable profile from approved copy

The complaint clients have about AI copy is that it sounds like everyone. You can fix that once and reuse it: feed the model several pieces of copy the client already approved, have it extract the actual patterns into a voice profile plus a reusable prompt, and every future draft starts closer to on-brand. It compounds — build it once, paste it before every job.

Prompt
You are a brand-voice analyst. Below are {{count}} pieces of copy that {{brand}} has approved and published. Analyze only these samples and build a reusable voice profile.

Approved samples: {{samples}}

Produce:
1. Voice attributes — 4 to 6 traits (e.g. "dry, not jokey"), each with a do example and a don't example drawn from or contrasted with the samples.
2. Mechanics — typical sentence length and rhythm, person (first/second), contractions, punctuation habits (e.g. em dashes, no exclamation marks).
3. Vocabulary — words and phrases the brand uses, and words it clearly avoids.
4. A reusable system prompt — a paste-ready paragraph I can put before future drafting so any model matches this voice.

Rules:
- Base every guideline on a pattern actually present in the samples. Do not import generic "brand voice" advice or invent rules the samples don't show.
- Quote a short example from the samples for each attribute so I can verify it.
- If the samples are inconsistent on something, say so rather than picking one.

Fill in your details and the prompt updates live — then copy.

Common questions from copywriters

Will AI replace copywriters?

The evidence points to the job changing rather than disappearing. In surveys, most copywriters use AI monthly and 66% say it has improved their work, even as 52% worry about displacement, and the writers who hand everything to AI report weaker results, not better ones. Strategy, judgment, a defensible claim, and a distinct voice are the parts clients pay for — and those are exactly what AI can't supply on its own.

Do I have to tell clients I'm using AI?

Check your contract and NDA first — some clients explicitly restrict AI use or require disclosure, and quietly ignoring that is a real risk. Surveys suggest most copywriters don't proactively disclose, but norms are tightening. The safe default is to agree on AI use up front, especially since AI-generated copy carries copyright and confidentiality implications your client should understand.

Is it safe to paste a client's brief or product details into ChatGPT or Claude?

Not into consumer accounts for anything confidential. Free and personal tiers may retain and train on what you paste, so unreleased products, pricing, launch dates, customer data, or an NDA-covered brief shouldn't go in. De-identify research before pasting, and use an enterprise or zero-retention tier for sensitive client work.

Can my client copyright copy that AI helped write?

Only the human-authored parts. The U.S. Copyright Office confirmed in January 2025 that text generated purely from prompts isn't copyrightable, and prompts alone don't make you the author. Your editing, selection, and arrangement are what create protectable authorship — so if a client needs to own the final copy, meaningful human rewriting matters, not just generation.

Related professions