Story Narrative Generation

Generate a complete demo narrative script that anchors product demonstration inside a real customer story. The output is a structured demo flow where every feature is shown through the lens of how a specific customer used it to solve a problem the prospect shares.

Prerequisites

• Customer story record (company, challenge, solution approach, results, key quotes)
• Prospect context (company, industry, size, pain points from discovery, stakeholder roles)
• Product feature catalog (features with descriptions)
• Anthropic API key

Steps

1. Assemble story and prospect inputs

Build the input object:

{
  "customer_story": {
    "company": "Acme Corp",
    "industry": "B2B SaaS",
    "headcount": 120,
    "challenge": "Manual onboarding taking 3 weeks per customer, causing 20% churn in first 90 days",
    "solution_approach": "Automated onboarding sequences with milestone tracking",
    "results": {
      "primary_metric": "Onboarding time reduced from 3 weeks to 4 days",
      "secondary_metrics": ["First-90-day churn dropped from 20% to 6%", "NPS increased 32 points"],
      "timeframe": "90 days"
    },
    "key_quotes": [
      "We went from losing a customer a week to not losing one in two months",
      "My team finally stopped firefighting and started building"
    ]
  },
  "prospect": {
    "company": "Beta Inc",
    "industry": "B2B SaaS",
    "headcount": 85,
    "pains": ["Slow onboarding frustrating new customers", "CS team overwhelmed with manual setup tasks"],
    "stakeholder_role": "VP Customer Success",
    "discovery_quotes": ["Our onboarding is basically a spreadsheet and prayer"]
  },
  "feature_catalog": [
    {"name": "Automated Workflows", "description": "..."},
    {"name": "Milestone Tracking", "description": "..."}
  ]
}

2. Generate the story-based demo narrative via Claude API

POST https://api.anthropic.com/v1/messages
x-api-key: {ANTHROPIC_API_KEY}
anthropic-version: 2023-06-01
Content-Type: application/json

{
  "model": "claude-sonnet-4-20250514",
  "max_tokens": 3000,
  "messages": [{
    "role": "user",
    "content": "Generate a demo narrative that tells a customer story while demonstrating the product. The prospect should see themselves in the customer's shoes.\n\nCustomer story:\n{customer_story_json}\n\nProspect context:\n{prospect_json}\n\nFeature catalog:\n{feature_catalog_json}\n\nRules:\n- Never say 'Let me show you this feature.' Always say 'Here's what [Customer] did when they faced [same challenge].'\n- Open with the customer's pain — the prospect must recognize their own problem.\n- Each feature shown must be anchored to a specific moment in the customer's journey.\n- Use the customer's exact quotes at emotional peaks.\n- Adapt the story details to mirror the prospect's specifics (company size, industry language, role context).\n- Close by connecting the customer's result to the prospect's stated pain.\n\nReturn this exact JSON:\n{\n  \"opening_hook\": \"1-2 sentences connecting the customer story to the prospect's pain. Reference the prospect's own words from discovery if available.\",\n  \"story_arc\": [\n    {\n      \"phase\": \"The Problem\",\n      \"narration\": \"What to say — telling the customer's challenge in words the prospect relates to\",\n      \"product_show\": \"What to demonstrate on screen during this narration\",\n      \"customer_quote\": \"Direct quote from customer, if relevant to this phase\",\n      \"duration_minutes\": 2\n    }\n  ],\n  \"emotional_peak\": \"The single moment in the story with highest emotional resonance — the 'aha' where the customer's pain turned into relief\",\n  \"closing_bridge\": \"1-2 sentences that connect the customer's outcome to the prospect's situation. End with a question, not a statement.\",\n  \"feature_to_story_map\": [\n    {\"feature\": \"feature name\", \"story_moment\": \"when in the story this feature appears naturally\"}\n  ],\n  \"adaptation_notes\": \"How the story was adapted to mirror the prospect's context\"\n}"
  }]
}

3. Generate via OpenAI API (alternative)

POST https://api.openai.com/v1/chat/completions
Authorization: Bearer {OPENAI_API_KEY}
Content-Type: application/json

{
  "model": "gpt-4o",
  "messages": [
    {"role": "system", "content": "You generate story-based demo narratives. Every demo feature must be shown through a customer story, never as a standalone feature tour."},
    {"role": "user", "content": "{SAME_USER_PROMPT_AS_ABOVE}"}
  ],
  "max_tokens": 3000,
  "temperature": 0.5
}

4. Validate the narrative

Check:

• story_arc has 3-5 phases (problem, turning point, solution, result at minimum)
• Every feature in feature_to_story_map corresponds to a phase in story_arc
• opening_hook references the prospect's specific pain, not a generic opener
• emotional_peak contains a concrete moment (quote, metric, or event), not an abstract statement
• closing_bridge ends with a question
• Total duration_minutes across phases is 15-25 minutes

If validation fails, re-prompt with specific corrections.

5. Store and attach to demo record

Store the narrative as an Attio note on the deal record, tagged with the customer story used and the prospect. Fire PostHog event:

{
  "event": "story_narrative_generated",
  "properties": {
    "deal_id": "...",
    "customer_story_company": "Acme Corp",
    "prospect_company": "Beta Inc",
    "story_phases": 4,
    "features_mapped": 3,
    "adaptation_depth": "high"
  }
}

Cost Estimates

• Anthropic Claude Sonnet: ~$0.04-0.10 per narrative
• At scale (50 narratives/quarter): under $5/quarter in API costs

Error Handling

• No matching customer story available: Fall back to a composite narrative built from multiple customer data points. Flag story_type: composite in output. Recommend creating a real case study for this segment.
• Insufficient discovery data: Generate narrative using industry-level pain assumptions. Flag personalization_depth: low.
• Feature catalog missing: Generate story arc only (no feature mapping). Flag for manual feature selection.

Tool Alternatives

| Tool | Method | Notes | |------|--------|-------| | Claude (Anthropic) | Messages API | Best narrative structure | | GPT-4 (OpenAI) | Chat Completions API | Good alternative | | Gemini (Google) | Generative AI API | Alternative LLM | | Manual | Story template | Fallback with fill-in-the-blank fields |