How msft-innovation-hub-india/agentic-ai-service-sample works
Main Components
a) Client Layer:
• Bot Framework Emulator (development)
• Azure Bot Service (production)
b) Application Layer:
• Bot Application (Python-based)
• State Management System
• User Profile Management
• Conversation Handler
c) Integration Layer:
• Azure AI Agent Service
• GPT-4 Model Integration
• REST API Integration
• Azure Logic Apps Integration
d) External Services:
• Contoso Retail Fashion REST APIs
• Azure SQL Database (for shipment orders)
┌─ Client Tier ─────┐ ┌─ Application Tier ──┐ ┌─ Integration Tier ─┐ ┌─ External Services ─┐
• Bot Framework │━━━>│ Bot Application │━━━>│ Azure AI Agent │━━━>│ Contoso Retail │
• Emulator/ │ │ (app.py) │ │ Service │ │ Fashion REST APIs │
• Azure Bot Service │ │ │ │ │ │ │
• State Management │ │ GPT-4 Model │ │ Azure Logic Apps │
• (state_management_ │ │ │ │ (Shipment │
• bot.py) │ │ │ │ Processing) │
• User Profile │ │ │ │ Azure SQL Database │
• Management │ │ │ │ (Consignments) │
• ┘ └──────────────────────┘ └────────────────────┘ └──────────────────────┘
Data Flow Representation
Following key flows:
Important Details
a) Configuration Components:
• Environment variables
• Swagger definition integration
• Logic App triggers and actions
b) Data Models:
• Conversation data structure
• User profile schema
• Consignment table schema
Technical Stack Callouts
• Python as primary language
• Microsoft Bot Framework
• Azure AI Agent Service
• GPT-4o-mini model
• Azure Logic Apps
• Azure SQL Database
Deployment View
• Local development setup
• Azure deployment options (App Service, Container Apps, Function Apps)
Remember to:
• Use consistent naming conventions
• Version information where relevant
• Provide zoom-in views for complex components