Click to watch our presentation video for the O'Reilly AI Katas Challenge
Click to watch our AI Agent Demo
This project is designed to deliver ShopWise Solution's AI Support Assistant which serves as the first line of support on our online e-commerce store, handling customer inquiries on products, orders, returns and refunds while seamlessly integrating with our company’s database. This AI-driven solution is designed to streamline customer interactions through advanced Natural Language Understanding (NLU), managing product inquiries, order tracking, and return eligibility via personalised, multi-turn conversations. With real-time access to product and order information, the AI assistant enables efficient comparisons, tailored recommendations and instant product availability checks.
This expansion into AI-powered support reflects ShopWise Solution's commitment to delivering an exceptional customer experience and reliable service across its diverse catalogue, which includes electronics, apparel, home goods, and more.
- 📖 Overview
- 📝 Requirements
- 🚀 Features and Roadmap
- 🔷 Architecture and Design
- 📊 Architecture Decision Records (ADRs)
- 💻 Tech Stack
- 🧪 Test Cases
- 🛠️ Deployment
The ShopWise AI Assistant is an innovative AI solution that provides the first line of support on our e-commerce platform, streamlining customer interactions with personalised AI assistance for product inquiries, order management, and returns through seamless real-time data integration. it enables:
- ✅ Personalised Recommendations
- 🗣️ Multi-Turn Conversations with contextual understanding
- 📊 Real-Time Product Comparisons and availability checks
- 🚚 Seamless Order Management (tracking, returns, and shipping)
Hi! We are the "ConnectedAI Team" – a group of innovative minds from Woolworths Group, Australia.
- Abhinav Singh - Senior ML Engineer
- Kenichi Nakanishi - Lead Data Scientist
- Meena Kasi - Solutions Architect
- Xun Zhou - ML Engineer Manager
- Yogesh Pathade - Tech Lead
📄 View Functional Requirements
Defines core features like Natural Language Understanding, product comparisons, and order management.
📄 View Non-Functional Requirements
Outlines performance benchmarks, accuracy, and system usability requirements.
📄 View Assumptions and Constraints
Details the foundational assumptions and system limitations.
📄 View Features
Lists all the features of the Web Based AI Assistant System.
📄 View Product Roadmap
Details the MVP release and future enhancements.
Provides an end-to-end architectural view of the AI Assistant System.
Below is the legend used for our C4 Architecture Diagrams
Provides an overview of how users and external systems interact with the Core System.
Illustrates the containers that make up the core AI Assistant.
Highlights the Responsible AI Monitoring and MLOps components.
Shows the components involved in order processing.
Highlights the components handling product specifications, availability, and attributes.
Details the logic and components used to compare and recommend products.
Provides an overview of the components managing product statistics and catalog insights.
Explore the rationale behind critical architectural decisions in the project.
- 📄 ADR Template
- 📄 NoSQL Database Storage
- 📄 Pre-Trained Models
- 📄 Custom API Gateway
- 📄 In-Memory Session Management
- 📄 LangGraph Framework
- 📄 Hosting Decisions
- 📄 Evaluation Framework
- 📄 Authentication & Authorization
- 📄 GitHub Actions CI/CD
- 📄 Caching Strategy
- 📄 Containerization
- 📄 WebSocket Implementation
📄 View Test Cases and Demos
Detailed test cases for functional and non-functional validation, including:
- Context Understanding
- Product Comparisons
- Order Status and Shipping Queries
Langfuse and Ragas facilitates observability, metrics, evaluations, prompt management for the AI agent.
📄 View Deployment Instructions
Step-by-step guide for running the system on your local machine, including backend and frontend setup.
📄 View Data Preprocessing and Loading Instructions
Step-by-step guide for data preprocessing and loading.
📄 View Cloud Hosting Instructions
Step-by-step guide for hosting the system in the cloud, including backend and frontend setup.