Woo AI Advisor
A WordPress plugin that deploys an AI-powered product advisor for WooCommerce stores.
Project Overview
The Woo AI Advisor is a WordPress plugin that enables e-commerce websites to deploy an AI-powered product advisor capable of answering customer questions and recommending products in real time. The system connects a WordPress frontend with a Cloudflare Worker acting as an AI gateway, allowing product data, documentation, and website content to be used as contextual knowledge for the AI. The goal of the project was to create an intelligent shopping assistant that improves product discovery and helps customers navigate complex catalogs using natural language. This project demonstrates my ability to design AI-powered customer experience tools, integrate LLM systems with e-commerce platforms, and build scalable serverless architectures.
My Role
I designed and built the full solution including the WordPress plugin architecture, AI orchestration layer using Cloudflare Workers, product data ingestion pipeline, prompt design and guardrails for LLM responses, and the recommendation logic used to match user questions with relevant products. The project required combining AI orchestration, data structuring, and frontend UX design to deliver a practical AI advisor for e-commerce.
The Problem Context
Many e-commerce stores struggle to help customers navigate large and complex product catalogs. Traditional navigation systems such as search bars and filters often fail when customers ask natural language questions about their needs.
- Customers ask natural language questions that traditional search cannot interpret.
- Large product catalogs make discovery difficult.
- Generic AI chatbots lack structured product knowledge.
- Lack of guardrails often results in unreliable recommendations.
The objective was to build an AI advisor capable of understanding user intent while grounding responses in structured product data and website knowledge.
Key Features & Workflow
AI Product Advisor
Customers can ask natural language questions and receive contextual product recommendations.
Structured Product Context
Product attributes and metadata guide the AI when generating recommendations.
Knowledge Injection
The AI can reference WooCommerce products, WordPress pages, and external knowledge files.
Serverless AI Gateway
Cloudflare Workers manage AI orchestration, security, and request handling.
Technical Architecture
How the system components fit together.
WordPress Frontend Plugin
Provides the chat interface embedded in WooCommerce stores and handles user interaction.
Cloudflare Worker Gateway
Acts as the AI orchestration layer handling authentication, prompt injection, and LLM requests.
Knowledge Sources
Includes WooCommerce product catalogs, WordPress pages, and uploaded knowledge datasets.
Design Decisions
A major design decision was to separate the AI processing layer from the WordPress plugin using a serverless gateway.
- User asks a question through the chat interface.
- The request is sent to the Cloudflare Worker AI gateway.
- Product and knowledge data are injected into the prompt.
- The LLM generates a contextual answer and product recommendations.
This architecture ensures responses remain grounded in store-specific knowledge while maintaining scalability and security.
Future Roadmap
- Vector embeddings for semantic product search.
- Advanced product ranking algorithms.
- CRM and analytics integration.
- Personalized recommendations based on user behavior.
Technology Stack
Core technologies used to build this project.
Capabilities Demonstrated
Key areas of expertise highlighted by this project.
