5 reusable marketing prompts you should save instead of rewriting
Use these AI prompt templates for homepage copy, buyer research, ad angles, launch planning, and campaign reviews — with variables you can edit and save in your prompt library.
Who this is for
This page is for marketers, creators, business owners, and entrepreneurs who use ChatGPT or Claude for recurring marketing work and want better AI prompts they can reuse instead of rewriting from scratch.
Specifically: solo growth marketers, fractional marketers, founder-marketers, creators, small business owners, early-stage SaaS operators, and service-business owners doing their own marketing.
The wedge is recurring marketing work — anything you find yourself rewriting from scratch each week or each campaign.
The core problem
Most marketers' AI workflow looks like this:
The version of the homepage hero prompt that worked in March is buried in a ChatGPT chat you can't find. The cold email angle that pulled 8% replies last quarter is a screenshot in your camera roll. The customer-research synthesis prompt that cracked your Q3 ICP is somewhere in Notion.
So every campaign, you retype from scratch. Every quarter, every launch, every new ICP. The prompt was an asset. You just didn't save it.
The cost isn't writing the prompt — ChatGPT costs $20/month. The cost is the re-derivation tax: 15 to 90 minutes of rebuilding logic that already existed.
Build → Refine → Save → Reuse
The behavior change is four steps:
- Build. Open the Foundry. Paste your product, audience, and objection into the variable fields. Run.
- Refine. Use the built-in actions (More Logic, Simplify, Make Reusable) to sharpen the prompt itself, not just the output.
- Save. Click Save. Name the run something like "Homepage Hero — Q3 launch (audience: fractional CMOs)." Future you is grateful.
- Reuse. Next quarter, open the saved run. Edit the audience variable. Click run. Done in 90 seconds.
That's the whole loop. The first run takes time; the rerun takes a variable edit.
The 5 reusable AI prompt workflows
Three are live in the Foundry today — you can build and save them in the next 5 minutes. Two are concept workflows on the roadmap; save the structure now, and the runnable recipe will follow.
1. Homepage Headline Angle Finder
Live in Foundry-
{{product}}— 50 words, the boring honest version -
{{audience}}— 30 words, specific -
{{main_objection}}— the one your sales team hears every week — optional
Role: You are a senior conversion strategist.
Variables: {{product}}, {{audience}}, {{main_objection}}.
Task: Generate 8 homepage hero headlines for {{product}} targeting {{audience}}. Rank the top 2.
Return format:
1. Diagnosis (one sentence)
2. Strategic angle (one phrase + why)
3. 8 headlines (max 12 words each)
4. Best option + why it wins
5. Alternative + why it's the backup
6. What not to say (3 vague phrases to reject)
Ranking/decision logic: Force ranking on clarity for {{audience}} and how directly each headline neutralizes {{main_objection}}.
What not to say: Avoid "unlock your potential", "revolutionary", or any phrase with no measurable claim. 2. Best Buyer Finder
Live in Foundry-
{{audience}}— the cohort name / segment -
{{goal}}— what you're trying to learn — e.g., "find the strongest objection" -
source material— paste 5–15 transcripts or notes inline
Role: You are a senior customer-research analyst.
Variables: {{audience}}, {{goal}}, {{source_material}}.
Task: Synthesize {{source_material}} into a structured ICP profile for {{audience}}.
Return format:
1. Recurring objection (verbatim, with frequency)
2. Trigger event that puts them in market
3. Verbatim language they use
4. Buying-process bottleneck
5. The single insight most worth acting on
Ranking/decision logic: Rank insights by how repeatable they are across the source material. Reject any insight grounded in fewer than 2 sources.
What not to say: No generic personas. No language not present in the source material. 3. Ad Angle Tester
Live in Foundry-
{{product}} -
{{audience}} -
{{channel}}— Meta, LinkedIn, X, Google, YouTube, etc. -
{{goal}}— sign-up, demo, trial, paid-conversion
Role: You are a direct-response advertising strategist.
Variables: {{product}}, {{audience}}, {{channel}}, {{goal}}.
Task: Generate 6 ranked ad angles for {{product}} on {{channel}} targeting {{audience}} with goal {{goal}}.
Return format (per angle):
1. Angle name (one phrase)
2. Hook line (under 12 words)
3. Testable hypothesis behind the angle
4. Buyer pain it speaks to
5. Why this beats the next-best angle
Ranking/decision logic: Rank by expected click-through times expected conversion for {{audience}} on {{channel}}.
What not to say: No "build awareness", no "engagement", no slogans without a payoff. 4. 30-Day Launch Plan
Coming soon-
{{product}} -
{{audience}} -
{{offer}} -
{{goal}} -
{{channel}}— the primary launch channel mix
Role: You are a senior go-to-market strategist.
Variables: {{product}}, {{audience}}, {{offer}}, {{goal}}, {{channel}}.
Task: Build a 30-day launch plan for {{product}} targeting {{audience}} with goal {{goal}}.
Return format:
1. Positioning diagnosis
2. Top 3 channels ranked by speed-to-impact
3. Week-by-week execution (4 weeks)
4. What not to do
5. Metrics to track
Ranking/decision logic: Reject low-leverage tactics with one-line reasoning. Force ranking on speed-to-impact, not reach.
What not to say: No "build awareness", no "post consistently", no calendar without a goal per week. Concept workflow — not live in Foundry yet. Save the structure now; the runnable recipe is on the roadmap. In the meantime, the closest live recipe is the Homepage Headline Angle Finder — apply this workflow's structure to the live recipe to get partial value today.
5. Weekly Campaign Review
Coming soon-
{{product}} -
{{channel}}— or channel set -
{{goal}}— the campaign objective -
{{metrics}}— this week's numbers, pasted in
Role: You are a senior growth analyst.
Variables: {{product}}, {{channel}}, {{goal}}, {{metrics}}.
Task: Synthesize {{metrics}} (spend, CTR, CVR, CAC, conversions) into a weekly campaign review for {{product}} on {{channel}} with goal {{goal}}.
Return format:
1. What is winning and why
2. What is losing and why
3. The single biggest leverage point
4. Three specific tests for next week with the hypothesis behind each
5. What to stop doing
Ranking/decision logic: Rank tests by expected lift / effort. Reject any test without a clear hypothesis.
What not to say: No "keep iterating", no "continue monitoring". Every recommendation must be a decision. Concept workflow — not live in Foundry yet. Save the structure now; the runnable recipe is on the roadmap. In the meantime, the closest live recipe is the Ad Angle Tester — apply this workflow's structure to the live recipe to get partial value today.
How to build a reusable prompt library for marketing
- Start with one recurring job: homepage copy, ad angles, buyer research, launch planning, or campaign review.
- Turn fixed details into variables: use
{{product}},{{audience}},{{offer}},{{goal}}, and{{channel}}. - Add logic: force the AI to rank, reject, compare, and explain tradeoffs.
- Save the prompt template: name it by workflow and campaign, not by random chat title.
- Reuse it next week: change the variables, keep the structure.
Before you save a prompt, sanity-check the structure: Grade a bad prompt before saving it →
How to use Promptmkr with the live workflows
Each live workflow takes about 3 minutes to set up the first time:
- Click any of the workflow buttons above.
- The Foundry loads with the workflow pre-selected.
- Paste your product, audience, and (where applicable) the main objection.
- Click Run.
- Click Save. Name the run.
Next quarter — or next week, if it's the Weekly Campaign Review — open the saved run, change one variable (audience, channel, or goal), and click Run again. The output is yours in seconds; the prompt structure stays intact.
This is the entire pitch: build it once, save it, reuse it. That's it.