40+ ChatGPT & Claude prompts for Marketing

**Role:** Positioning consultant trained in April Dunford's "Obviously Awesome" method. You've repositioned 30+ B2B SaaS companies and learned which questions surface real positioning vs which produce taglines.

**Context:** Company: [name]. Current positioning (if any): [what we say today]. Best customer: [name + why they bought]. Best customer's competitive alternative: [what they considered or were using before].

**Task:** Run the 5-step positioning exercise and produce a positioning statement.

1. Competitive alternatives: what do customers compare us to? Not just direct competitors — also "doing it manually" or "hiring a person."
2. Unique attributes: what do we have that the alternatives don't? Be specific — features, integrations, expertise, pricing structure.
3. Value (and proof): what value do those attributes deliver? Quantify where possible. Cite a customer story.
4. Target market characteristics: who cares about that value the MOST? Be specific — company size, role, trigger event, tech stack.
5. Market category frame: what market are we IN, in the customer's mind? If we frame it wrong, we lose to the wrong alternatives.

**Constraints:**
- Be honest about what we DON'T have vs alternatives
- Target market must be narrow enough that we can name the persona
- Market category frame is a choice — show 2 options and pick one
- Final positioning statement ≤60 words

**Output format:** 5 numbered sections + final positioning statement + 1-sentence "what we will say in cold emails."

**Role:** Conversion copywriter who has shipped 200+ landing pages and learned which sentences earn the scroll vs which ones lose it.

**Context:** Product: [name + category]. Audience: [persona + the job-to-be-done]. Current page H1: [paste the existing headline]. Conversion rate today: [%]. One specific competitor's homepage we're losing against: [URL].

**Task:** Rewrite the landing page above the fold. Goal: earn the scroll. No buyer needs a "leading provider" of anything.

1. Headline: 8-12 words. Names the job-to-be-done in plain language. Not "AI-powered platform" but "Close vendor security questionnaires in 24 hours instead of 4 weeks."
2. Sub-headline: 1 sentence. Removes the obvious objection.
3. Three benefit bullets. Each one has a proof point — a number, a customer name, or a screenshot reference. Not "save time" but "Acme cut SOC 2 response time from 18 days to 2."
4. CTA: action verb + specific noun. Not "Get started" but "See your first audit ready in 5 minutes."
5. Risk reversal: 1 line. "No card required for the first audit" beats "free trial."

**Constraints:**
- No "AI-powered" / "innovative" / "best-in-class"
- Every claim has a proof point
- Headline can't be a question
- Avoid "the only platform" superlatives unless verifiable

**Output format:** Markdown · 5 sections (Headline, Sub, 3 bullets, CTA, Risk reversal) · ≤180 words.

**Role:** Senior paid social copywriter with 5+ years scaling Meta ads past $1M/mo at multiple DTC brands.

**Context:** Product: [DESCRIBE]. ICP: [persona]. Current best-performing ad: [paste]. Budget: [$/day]. Objective: conversions.

**Task:** Produce 5 distinctly-angled ad variants:
1. Pain-led (open with the customer's frustration in their words)
2. Outcome-led (open with the specific result, no setup)
3. Proof-led (open with a number or named customer)
4. Contrarian (open by challenging the obvious advice)
5. Question (open with a question that earns the next sentence)

For each: primary text, headline, CTA. Plus: the A/B test plan with sample size, primary metric (CPA), kill criterion.

**Constraints:** Each variant truly different angle (not paraphrase) · proof points cited · no AI-marketing-speak · creative spec specified separately if needed.

**Output format:** 5 ad blocks + test plan + creative notes.