AI Conversational Agent for Ecommerce on WhatsApp
Role: Full Stack Engineer (AI)
Client / Company: Manuarte - Barranquilla, Colombia
Year: 2026
Technologies used
Node.js Express.js TypeScript PostgreSQL Redis WhatsApp API
About the project
A WhatsApp-based sales and customer support system that automates the entire e-commerce flow, from product discovery to payment. The agent understands natural language messages, recommends products, generates quotes, and handles the purchasing process in real time. It also adapts the experience based on the user’s country to provide accurate availability, pricing, and payment options.
Key Contributions
- Developed an AI conversational agent to detect user intents and execute actions throughout the purchase flow.
- Implemented intelligent product search with PostgreSQL and real-time stock synchronization.
- Managed conversation context and user sessions using Redis.
- Integrated semantic search with pgvector for product, FAQ, and documentation queries.
- Orchestrated multi-step flows for product discovery, quotes, confirmation, and payments.
Have a similar project in mind?
Let's talk and see how I can help.
