use-case-generator
Generate complete use case implementations for specific industries and scenarios. Use when creating new demo pages, implementing industry-specific mascot integrations, or exploring new application areas.
Installation
Follow the repository instructions for this skill and your agent. Review the files and dependencies before adding it to your workflow.
View source ↗GuidesSkill instructions
Use Case Generator
You are an expert in creating industry-specific use case implementations for the emotive-mascot platform.
When to Use This Skill
- Creating new demo pages for specific industries
- Implementing mascot for client use cases
- Exploring new application areas
- Building proof-of-concept implementations
- Generating industry-specific emotion mappings
Available Use Case Templates
We have 4 fully implemented use cases you can reference:
1. Retail Checkout
Location: site/src/app/use-cases/retail/
Features:
- Item scanning simulation
- Payment processing flow
- AI checkout assistant (Claude integration)
- Error handling and recovery
- Success celebrations
Emotions: calm → anticipation (scanning) → joy (success) → concern (error)
2. Smart Home Dashboard
Location: site/src/app/use-cases/smart-home/
Features:
- Device control interface
- Voice command simulation
- Status monitoring
- AI home assistant
- Multi-device management
Emotions: calm → focus (listening) → anticipation (processing) → joy (success)
3. Healthcare Patient Intake
Location: site/src/app/use-cases/healthcare/
Features:
- Form validation
- Sensitive data handling
- Privacy reassurance
- AI patient assistant
- Empathetic interactions
Emotions: calm → empathy (concerns) → reassurance (explaining) → gratitude (complete)
4. Education Learning Platform
Location: site/src/app/use-cases/education/
Features:
- Quiz interface
- Progress tracking
- AI tutor assistant
- Hint system
- Celebration for correct answers
Emotions: calm → contemplation (thinking) → pride (correct) → encouragement (incorrect)
Generating New Use Cases
Step 1: Define the Scenario
// Define your use case
const useCase = {
industry: 'banking', // or retail, healthcare, education, etc.
context: 'loan-application',
userGoal: 'Complete loan application form',
painPoints: [
'Complex forms',
'Unclear requirements',
'Anxiety about approval',
],
mascotRole: 'Supportive guide and reassurance provider',
};
Step 2: Map Emotions to User Journey
const emotionJourney = {
onboarding: 'calm', // User arrives
formStart: 'encouragement', // User begins
complexQuestion: 'empathy', // User confused
validation: 'concern', // Error occurred
corrected: 'encouragement', // User fixed error
submission: 'anticipation', // Processing
approval: 'celebration', // Success!
rejection: 'empathy', // Support needed
};
Step 3: Create Page Structure
// site/src/app/use-cases/[industry]/page.tsx
'use client'
import { useState, useEffect, useRef } from 'react'
import PremiumAIAssistant from '@/components/PremiumAIAssistant'
export default function IndustryUseCasePage() {
const [mascot, setMascot] = useState(null)
const [showAI, setShowAI] = useState(false)
const [userProgress, setUserProgress] = useState('start')
// Initialize mascot
useEffect(() => {
const initMascot = async () => {
const { EmotiveMascot } = await import('@joshtol/emotive-engine')
const m = new EmotiveMascot({
canvas: canvasRef.current,
initialEmotion: 'calm',
enableGazeTracking: true
})
await m.initialize()
setMascot(m)
}
initMascot()
}, [])
// Emotion handlers
const handleUserAction = async (action: string) => {
const emotion = emotionJourney[action]
await mascot?.transitionTo(emotion, { duration: 1000 })
}
return (
<div>
{/* Canvas */}
<canvas ref={canvasRef} />
{/* Your UI */}
<YourIndustrySpecificUI onAction={handleUserAction} />
{/* AI Assistant */}
{showAI && (
<PremiumAIAssistant
title="Your Assistant"
context={useCase.context}
onLLMResponse={(emotion) => mascot?.transitionTo(emotion)}
onClose={() => setShowAI(false)}
/>
)}
</div>
)
}
Industry-Specific Patterns
Financial Services
- Key emotions: calm, reassurance, empathy, trust
- Triggers: Form complexity, data privacy concerns, approval anxiety
- Mascot role: Confidence builder, guide
E-Commerce
- Key emotions: joy, excitement, anticipation, gratitude
- Triggers: Product discovery, cart interactions, checkout success
- Mascot role: Shopping companion, celebration enhancer
Healthcare
- Key emotions: empathy, calm, reassurance, gratitude
- Triggers: Sensitive questions, privacy concerns, anxiety
- Mascot role: Empathetic supporter, privacy guardian
Education
- Key emotions: encouragement, pride, contemplation, joy
- Triggers: Correct/incorrect answers, hints requested, progress
- Mascot role: Patient tutor, achievement celebrator
Customer Support
- Key emotions: focus, empathy, concern, gratitude
- Triggers: Issue reporting, troubleshooting, resolution
- Mascot role: Problem solver, empathetic listener
Quick Start Template
Use this template to quickly scaffold a new use case:
# Create new use case directory
mkdir -p site/src/app/use-cases/your-industry
# Copy from existing template
cp -r site/src/app/use-cases/retail/* site/src/app/use-cases/your-industry/
# Customize for your industry
# - Update emotion mappings
# - Modify UI components
# - Adjust AI assistant prompts
# - Add industry-specific interactions
Resources
Examples
Read the usage examples in the source repository.
View source ↗Source history
- Indexed on
- 2026-01-20
- Index updated
- 2026-01-26
- Branch
- main
Index and sync dates describe this directory. View the source repository for code history.
GitHub ↗chat Comments (0)
Sign in to join the discussion and leave a comment.
No comments yet. Be the first to share your thoughts!