❄️ Snowflake Intelligence Data Generator

A powerful Streamlit application that generates tailored demo data infrastructure for Snowflake's Cortex Analyst and Cortex Search services. Create realistic, AI-powered demo environments in minutes with custom data that matches your customer's industry and use cases.

🎯 What This App Does

The SI Data Generator creates complete demo environments that showcase Snowflake's AI capabilities:

• 🤖 AI-Generated Demo Ideas: Uses Cortex LLM to create 3 tailored demo scenarios based on customer information
• 📊 Realistic Structured Data: Generates business-relevant tables with proper relationships and constraints
• 🔍 Searchable Unstructured Data: Creates text chunks optimized for semantic search
• 🔗 Semantic Views: Builds AI-ready views with relationships for Cortex Analyst
• 🔎 Cortex Search Services: Configures semantic search services for document retrieval

🚀 Key Features

🎨 Intelligent Demo Generation

• Customer-Specific: Tailors demos based on company URL, team members, and use cases
• Industry-Aware: Generates relevant scenarios for e-commerce, healthcare, financial services, etc.
• AI-Powered: Uses Snowflake Cortex LLM to create realistic business contexts

📈 Complete Data Infrastructure

• Structured Tables: 2 tables with PRIMARY KEY constraints and realistic business data
• Unstructured Content: Searchable text chunks with metadata
• Joinable Architecture: All tables connect via ENTITY_ID for comprehensive analytics
• Semantic Views: AI-ready views with synonyms, relationships, and example queries

🔧 Production-Ready Setup

• Automated Schema Creation: Creates organized database schemas
• Proper Constraints: PRIMARY KEY constraints for optimal join performance
• Cortex Integration: Ready-to-use Cortex Search services
• Comprehensive Documentation: Generated demo guides and example queries

🏗️ Architecture

SI_DEMOS (Database)
├── APPLICATIONS (Schema)
│   ├── SI_DATA_GENERATOR_REPO (Git Repository)
│   └── SI_DATA_GENERATOR_APP (Streamlit App)
└── [COMPANY]_DEMO_[DATE] (Generated Demo Schemas)
    ├── TABLE_1 (Structured data with ENTITY_ID PK)
    ├── TABLE_2 (Structured data with ENTITY_ID PK)
    ├── CONTENT_CHUNKS (Unstructured searchable text)
    ├── [COMPANY]_SEMANTIC_VIEW (AI-ready view)
    └── [TABLE]_SEARCH_SERVICE (Cortex Search)

📋 Prerequisites

• Snowflake Account with ACCOUNTADMIN privileges
• Cortex Access enabled in your Snowflake account
• Streamlit feature enabled
• Git Integration capabilities

🛠️ Quick Setup

1. Run the Setup Script

-- Execute the complete setup script
-- File: Setup.sql
-- This creates the entire environment including:
-- - Database and schemas
-- - Roles and permissions  
-- - Git integration
-- - Streamlit application

2. Access the Application

• Navigate to Data > Streamlit Apps in Snowsight
• Open SI_DATA_GENERATOR_APP from SI_DEMOS.APPLICATIONS
• Start generating demos!

🎮 How to Use

Step 1: Customer Information

Enter:

• Company URL: Customer's website (e.g., https://acme.com)
• Team Members: Who you're presenting to (e.g., "CTO, Data Team")
• Use Cases: Specific requirements (optional)
• Record Count: Number of sample records per table

Step 2: AI Demo Generation

• Click "Generate Demo Ideas"
• App uses Cortex LLM to create 3 tailored scenarios
• Each demo includes:
  • Compelling title and description
  • Industry focus and business value
  • 2 structured tables + 1 unstructured table
  • Purpose and use case explanations

Step 3: Demo Selection

• Review the 3 AI-generated demo ideas
• Each shows:
  • Structured Table 1: Primary business data
  • Structured Table 2: Supporting/related data
  • Unstructured Table: Searchable content
• Select the best fit for your customer

Step 4: Infrastructure Creation

• Choose schema name (auto-generated)
• Enable optional features:
  • Semantic View: AI-ready analytics view
  • Cortex Search Service: Semantic search capability
• Click "Create Demo Infrastructure"

Step 5: Demo Ready!

The app creates:

• ✅ Database Schema with organized structure
• ✅ Structured Tables with realistic data and PRIMARY KEY constraints
• ✅ Unstructured Table with searchable text chunks
• ✅ Semantic View with relationships and example queries
• ✅ Cortex Search Service for document retrieval
• ✅ Complete Documentation with demo flow and example questions

🎯 Example Demo Scenarios

E-commerce Analytics

• Tables: Sales transactions, customer profiles
• Search: Product reviews and feedback
• Queries: "What are our top-performing customer segments by revenue?"

Financial Services

• Tables: Transaction monitoring, compliance events
• Search: Regulatory documents and policies
• Queries: "Show me high-risk transactions and their compliance status"

Healthcare Analytics

• Tables: Patient outcomes, treatment protocols
• Search: Clinical notes and research documentation
• Queries: "Which treatment protocols have the best patient outcomes?"

🔍 Demo Flow Example

1. Structured Analytics (Cortex Analyst)

Question: "What are the top 5 performing entities and their key metrics?"
→ Cortex Analyst queries structured tables
→ Joins data using ENTITY_ID relationships
→ Returns analytical insights with visualizations

2. AI Reasoning Follow-up

Question: "What could be the reasons for these performance differences?"
→ Agent uses AI reasoning (not data querying)
→ Provides business insights and hypotheses
→ Suggests potential factors and correlations

3. Knowledge Retrieval (Cortex Search)

Question: "Find relevant best practices for improving these metrics"
→ Cortex Search queries unstructured content
→ Returns contextual information from text data
→ Combines with previous analysis for complete insights

📊 Generated Data Features

Structured Tables

• Realistic Business Data: Industry-relevant columns and values
• PRIMARY KEY Constraints: ENTITY_ID for optimal joins
• 70% Data Overlap: Meaningful relationships between tables
• Proper Data Types: NUMBER, STRING, DATE, TIMESTAMP, BOOLEAN
• Business Context: Relevant to customer's industry

Unstructured Content

• Chunked Text: Optimized for semantic search
• Rich Metadata: Document types, sources, timestamps
• Searchable Attributes: CHUNK_ID, DOCUMENT_ID, DOCUMENT_TYPE
• Realistic Content: Business-relevant text samples

Semantic Views

• AI-Ready Relationships: Properly defined joins
• Business Synonyms: Multiple ways to reference data
• Example Queries: Pre-built questions for demos
• Cortex Analyst Extension: Enhanced AI capabilities

🔧 Customization Options

Demo Templates

Modify fallback demo ideas in the application:

def get_fallback_demo_ideas(company_name, team_members, use_cases):
    # Add your custom industry templates
    # Healthcare, Finance, Retail, Manufacturing, etc.

Data Generation

• Record Counts: 20 to 10,000 records per table
• Industry Focus: Automatic detection from company URL
• Use Case Tailoring: Custom scenarios based on requirements

Compute Resources

-- Scale warehouse based on usage
ALTER WAREHOUSE SI_DEMO_WH SET WAREHOUSE_SIZE = 'X-SMALL';

📈 Business Value

For Sales Teams

• Rapid Demo Setup: Create tailored demos in minutes
• Customer-Specific: Relevant to prospect's industry
• Complete Infrastructure: Ready-to-use AI capabilities
• Professional Presentation: Polished, realistic data

For Technical Teams

• Best Practices: Proper constraints and relationships
• AI Integration: Cortex Analyst and Search examples
• Scalable Architecture: Easy to extend and modify
• Documentation: Complete setup and usage guides

For Customers

• Real-World Scenarios: Relevant to their business
• AI Capabilities: See both analytics and search in action
• Immediate Value: Understand capabilities through familiar data
• Future-Ready: Architecture supports growth and expansion

🧹 Cleanup & Management

Remove Demo Data

-- List all demo schemas
SHOW SCHEMAS IN DATABASE SI_DEMOS LIKE '%_DEMO_%';

-- Remove specific demo
DROP SCHEMA IF EXISTS SI_DEMOS.ACME_DEMO_20250115;

Monitor Usage

• Compute Costs: Monitor warehouse usage
• Storage: Track table sizes in demo schemas
• Cortex: Review AI function call costs

🆘 Troubleshooting

Common Issues

"Insufficient privileges" Error

USE ROLE ACCOUNTADMIN;
-- Re-run setup script

"Cortex function not accessible" Error

GRANT USAGE ON FUNCTION SNOWFLAKE.CORTEX.COMPLETE(STRING, STRING) 
TO ROLE ACCOUNTADMIN;

"Streamlit app not loading"

• Verify git repository is synced
• Check file path in app definition
• Ensure all permissions are granted

📚 Documentation

• NATIVE_APP_PROVIDER_GUIDE.md: Complete guide for Native App providers (installation, testing, releases)
• Setup.sql: Automated environment creation for standalone deployment
• Generated Demo Guides: Custom documentation for each demo

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🎉 Success Stories

"This tool has revolutionized our demo process. We can now create customer-specific demos in minutes instead of hours, and the AI-generated scenarios are incredibly realistic and relevant." - Sales Engineering Team

"The semantic views and Cortex Search integration make it easy to showcase the full power of Snowflake's AI capabilities. Our customers immediately understand the value." - Solutions Architect

---

Ready to create amazing Snowflake demos? 🚀

Native App Provider Guide | Standalone Setup Script | Report Issues