| name | kai-case-study |
| description | Produce customer case studies from interviews or data — Problem, Solution, Results structure with perception engineering and quality gates. Use when "case study", "customer story", "testimonial", "success story", "client results", or any request to document a customer win. |
Kai Case Study Skill
Produce compelling customer case studies using a Problem, Solution, Results structure with perception engineering layers.
Phase 0: Load Product Context
Check if MARKETING.md exists in the project root (same directory as CLAUDE.md, README.md, package.json).
If it exists: Read it — skip product discovery questions. It has the product name, ICP, value prop, monetization, brand voice, current channels, and competitive landscape.
If it does NOT exist: Auto-explore the codebase to create it in the project root (next to CLAUDE.md). Do NOT ask the user what the product is. Read CLAUDE.md, README.md, PROJECT.md, package.json, landing pages, and any project files. Search for email/ad/analytics config. Then create MARKETING.md using the template from /kai-email-system. Present draft to user for confirmation.
Phase 1: Discovery
Read from MARKETING.md. Only ask about things not covered there:
- Customer info — Company name, industry, size, role of contact
- Source material — Interview transcript, survey responses, data points, screenshots
- The problem — What was broken before? Quantify the pain.
- The solution — What did we do? Be specific about the product/service.
- The results — Hard numbers. Revenue, time saved, conversion lift, cost reduction.
- Persona alignment — Which harness persona does this customer map to? Load from
knowledge/personas/_persona-index.md
- Permission — Does the customer approve named use? Or anonymized?
Phase 2: Plan
Structure the case study:
- Load content checklist:
knowledge/checklists/content-checklist.md
- Load perception engineering:
knowledge/frameworks/content-copywriting/perception-engineering.md
- Define the narrative arc:
- Before state — The specific pain, in the customer's words
- Turning point — Why they chose us (decision trigger)
- After state — Measurable transformation
- Key quote selection — Pull 2-3 direct quotes that carry emotion + specificity
- Proof points — List every number, metric, and data point available
- Distribution plan — Where will this live? (website, sales deck, email, social)
Phase 3: Produce
Write the case study:
- Headline — Lead with the result, not the company name. Example: "73% Faster Onboarding: How [Company] Rebuilt Their Workflow"
- Snapshot box — Company, industry, challenge, result (scannable summary)
- The Challenge — 2-3 paragraphs. Paint the before state. Use customer language.
- The Solution — 2-3 paragraphs. What we did, how it worked. Be concrete.
- The Results — Lead with the biggest number. Use a data table or callout boxes.
- Customer quote — Close with their strongest testimonial line.
- CTA — What should the reader do next?
Apply perception engineering layers:
- Perception layer: Re-index the old way as the problem (not just "less good")
- Context layer: Make the new approach feel inevitable
- Permission layer: Remove risk from taking action
Phase 4: Quality Gates
Run all gates before delivery:
- Four U's Score:
python scripts/quality_gates/four_us_score.py <file>
- Minimum: 12/16 (content threshold)
- Banned Word Check:
python scripts/quality_gates/banned_word_check.py <file>
- AI Slop Check — No filler phrases ("In conclusion", "It's worth noting that", etc.)
- Specificity check — Every claim has a number or named example. No vague praise.
Max 2 auto-retry cycles. After 2 failures, surface to human with specific failure reasons.
Phase 5: Output
Deliver the final case study package:
- Full case study (long-form, 800-1500 words)
- One-page summary (for sales team, 250 words max)
- Pull quotes (2-3 standalone quotes for social/email use)
- Headline variants (3 options for different channels)
- Four U's scorecard
- Gate pass/fail summary
Write output to workspace/ with filename pattern: case-study-[company]-YYYY-MM-DD.md
