- Registered User Login
- Allows users to log in for personalised interactions.
- Guest Login
- Enables non-registered users to interact with limited functionalities.
- AI Assistant Chat Interface
- Provides a conversational interface for interacting with the assistant.
- Streaming AI Assistant Responses
- Delivers real-time response streaming for smoother interactions.
- AI Backend API Service
- Central API layer for managing requests and orchestrating responses.
- Routing Requests to the AI Agents
- Routes user queries to the appropriate AI agents for processing.
- Session Management Service
- Manages user sessions for maintaining context in multi-turn conversations.
- Sending Requests to Responsible AI ML Ops Monitoring Platform
- Ensures all requests are logged and monitored for compliance and ethical standards.
- Customer Support Supervisor
- Oversees and escalates unresolved issues to fallback mechanisms or human support.
- Order Processing Agent
- Handles order status retrieval, tracking, and return eligibility queries.
- Order Database
- Stores and retrieves order-related information in real time.
- Product Supervisor
- Delegates product-related tasks to specialized agents.
- Product Attribute Agent
- Retrieves product specifications, pricing, and availability.
- Product Comparison & Recommendation Agent
- Provides personalised product comparisons and recommendations.
- Product Statistics Agent
- Offers statistical insights such as average ratings and popular items.
- Product Database
- Central repository of product information for real-time access.
- Content Refinement Service
- Refines and structures responses for user-friendly presentation.
- Fallback Agent
- Redirects unresolved queries to human support or alternative agents.
- Exit Handler
- Ensures smooth termination of user sessions with proper summary or follow-ups.
- Responsible AI Monitoring Service
- Monitors AI interactions to ensure compliance with ethical guidelines.
- Fine-Tuned Dataset
- Provides optimized datasets for training and improving the AI assistant.
- Responsible AI Evaluation Pipelines
- Validates AI responses for accuracy, bias, and ethical compliance.
- Data Sanitization Pipeline
- Cleans and preprocesses input data for model training and operations.
- LLM Fine-Tuning Pipeline
- Fine-tunes language models to improve performance and adapt to new data.
- Responsible AI Dashboards
- Visualizes AI performance metrics, bias reports, and operational insights.
- Responsible AI Alerts
- Sends notifications for potential issues such as ethical violations or system failures.