I lead marketing for [COMPANY AND OFFER]. Build a copy-paste prompt library for [PRIMARY MARKETING GOAL] aimed at [TARGET AUDIENCE]. Our brand voice and constraints are: [BRAND CONTEXT]. The channels, assets, and data currently available are: [AVAILABLE CHANNELS AND DATA]. Create 10 distinct prompts I can use in ChatGPT, Claude, or Gemini for day-to-day marketing work. Cover these jobs: customer research synthesis, positioning, campaign brief, landing-page messaging, email sequence, social-content plan, SEO content brief, paid-ad concepts, sales enablement, and performance analysis. For each prompt: 1. Give it a clear job-based title. 2. Write a ready-to-paste prompt with only the inputs that task actually needs, using capitalized square-bracket placeholders. 3. Tell the AI the output format, decision criteria, and limits. For example, a campaign brief should name audience insight, single-minded proposition, proof, channel role, message hierarchy, CTA, and measurement; an analysis prompt should separate observed facts from hypotheses and recommend the next test. 4. Include a short “Best when” note and one practical follow-up prompt. Make the prompts specific to my context, not a list of interchangeable marketing clichés. Do not invent customer evidence, performance results, competitor claims, testimonials, legal approvals, or product capabilities. Where evidence is missing, instruct the model to label assumptions and suggest the minimum research needed. Before answering, check that all 10 prompts serve different decisions or deliverables, and that every requested claim has a source input or is explicitly marked as an assumption. Output a numbered library, with each entry under 180 words.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [COMPANY AND OFFER] | Describe your company, product or service, pricing model, and the main problem it solves. | Northstar HR sells onboarding software for 100-500 person companies; $8 per employee monthly; it reduces manual new-hire setup. |
| [PRIMARY MARKETING GOAL] | State the outcome you need, such as qualified demos, trial signups, retention, or category awareness. | Increase qualified demo requests from HR operations leaders in Q4. |
| [TARGET AUDIENCE] | Describe the buyer or user, their role, situation, pain points, and buying context. | US HR operations managers at growing SaaS firms who manage onboarding in spreadsheets and multiple systems. |
| [BRAND CONTEXT] | Add your voice, claims you can support, prohibited language, competitors, and relevant positioning. | Practical, calm, and evidence-led; avoid claims about guaranteed retention; competitors include two large HR platforms. |
| [AVAILABLE CHANNELS AND DATA] | List active channels, existing assets, budget limits, analytics, research, and team capacity. | LinkedIn organic, monthly webinar, email list of 4,200 leads, $12,000 paid-search budget, CRM conversion data, and two customer interviews. |
How to use
- Paste the core prompt and replace every bracketed field with your real market, offer, and evidence.
- Save the resulting prompts in a shared document, then test the two most urgent ones against a live brief or dataset.
- Check that each prompt requests a concrete deliverable and does not treat assumptions as customer facts.
- Follow up with: “Rewrite prompts 3 and 7 for a two-person team with no paid-media budget and a 30-day launch window.”
Variations
Campaign Brief
Use this when one campaign needs a decision-ready brief before copy or design begins.
I am planning a [CAMPAIGN TYPE] for [OFFER] to reach [AUDIENCE]. Use these verified inputs: [CUSTOMER EVIDENCE], [CHANNEL AND BUDGET], and [BUSINESS GOAL]. Write a one-page campaign brief with the audience tension, single-minded proposition, supporting proof, message hierarchy, offer, CTA, channel roles, asset list, timeline, and success metrics. Separate facts from assumptions. Do not invent proof points or customer quotes. Before answering, check that the CTA and metrics match the stated business goal.
Content Brief
Use this when assigning an SEO article, guide, or newsletter issue.
Create a content brief for [TOPIC OR KEYWORD] for [AUDIENCE] at [COMPANY]. The search intent and known customer questions are [SEARCH AND RESEARCH NOTES]. Include the reader’s job to be done, angle, working title, outline, information that requires a source, internal links to consider, CTA, and acceptance criteria for a writer. Identify claims that need verification and do not fabricate statistics or expert quotes. Before answering, check that the outline answers the search intent before introducing our product.
Performance Review
Use this when deciding what to change after a campaign has run.
Analyze this [CAMPAIGN OR CHANNEL] performance data: [DATA]. The original goal was [GOAL], and these changes or external factors may matter: [CONTEXT]. Produce a concise review with a metric table, what happened, likely explanations ranked by confidence, segments worth investigating, and three next experiments. State calculation formulas when deriving rates. Do not claim causation from correlation or fill data gaps with estimates. Before answering, verify that each recommendation has a measurable expected signal and a defined decision rule.
Tips
- Use one library prompt per marketing decision; combining research, copy, channel selection, and reporting in one request produces vague work.
- Give the model raw interview excerpts or campaign exports when possible, then require it to distinguish direct evidence from interpretation.
- For regulated, technical, or comparative claims, add your approved claims list and require exact source citations before copy is drafted.
- Keep a short record of which prompts led to usable work, then revise their acceptance criteria rather than only changing the tone.
FAQ
Should I ask AI to create a whole marketing strategy from scratch?
It can structure options and expose missing inputs, but it should not substitute for customer evidence, channel economics, or leadership decisions. Give it research and constraints first.
How many placeholders should a marketing prompt have?
Use only inputs that change the answer. A landing-page prompt usually needs audience, offer, proof, and CTA; it does not need every company detail.
Can these prompts work in Claude, ChatGPT, and Gemini?
Yes. Keep the inputs and acceptance criteria in the prompt, then compare outputs using the same evidence and constraints.