Claude-optimized prompt structure shaped for chatgpt.
# Task
Create a case study outline from customer notes. Use ONLY the provided facts.
# Inputs- Customer context: {customer_context}- Problem: {problem}- Solution: {solution}- Results: {results}- Proof points: {proof_points}- Audience: {audience}- Tone: {tone}# Anti-hallucination rules1. Every claim must trace to the provided inputs. No invented metrics or quotes.
2. If a result lacks a specific number, use the qualitative description as-is. Don't round up or embellish.
3. Mark any gaps with [NEEDS DATA] so the writer knows what to fill.
# Narrative arc (follow this structure)1. **Headline** (benefit-led, under 80 chars, includes the key metric if available)
2. **Executive summary** (3-4 sentences: who, what problem, what result)
3. **The challenge** (Situation → Complication):
- Who is the customer and what's their context?
- What specific problem were they facing? What was the business impact?
4. **The solution** (Resolution):
- What did they choose and why?
- How was it implemented? (timeline, scope)
5. **The results** (Proof):
- Quantified outcomes (use exact numbers from inputs)
- Qualitative outcomes (team feedback, workflow changes)
- Format key metrics as callout boxes: [CALLOUT: metric]
6. **Pull quotes**: Select 2-3 candidate quotes from the inputs (or suggest where to source them). Each should support a different section.
7. **CTA**: What should the reader do next?
# SEO- Suggest a URL slug, meta title (55-65 chars), and meta description (150-160 chars).
# Output
Return the complete outline with all sections, callout markers, and SEO suggestions.
claudegrounded_factsexamples
claude variant
Claude-optimized prompt structure shaped for claude.
<context><customer_context>{customer_context}</customer_context><problem>{problem}</problem><solution>{solution}</solution><results>{results}</results><proof_points>{proof_points}</proof_points><audience>{audience}</audience><tone>{tone}</tone></context><task>Create a case study outline using ONLY the provided facts.</task><instructions>
No invented metrics or quotes. Use qualitative descriptions as-is. Mark gaps [NEEDS DATA].
Arc: Headline (benefit-led, <80 chars) > Executive summary > Challenge (situation → complication) > Solution (choice + implementation) > Results (quantified + qualitative, with [CALLOUT: metric]) > Pull quotes (2-3 candidates) > CTA.
SEO: URL slug, meta title (55-65 chars), meta description (150-160 chars).
</instructions><output>
Return the complete outline with callout markers and SEO suggestions.
</output>
geminigrounded_factsexamples
gemini variant
Claude-optimized prompt structure shaped for gemini.
Customer context: {customer_context}
Problem: {problem}
Solution: {solution}
Results: {results}
Proof points: {proof_points}
Audience: {audience}
Tone: {tone}
Rules: only provided facts, no invented metrics, mark gaps [NEEDS DATA].
Arc: Headline (<80 chars) > Summary > Challenge (situation → complication) > Solution > Results (with [CALLOUT: metric]) > Pull quotes (2-3) > CTA.
SEO: slug, meta title (55-65 chars), meta description (150-160 chars).
Based on the entire content above, return the complete case study outline.
Related templates
More templates in the same working category.
LinkedIn Post
Write a LinkedIn post with a scroll-stopping hook, structured insight, proof, and engagement-optimized formatting.