A WhatsApp-powered AI chatbot that brings Zimbabwe's commodity market to every trader's phone. Users can buy and sell commodities, calculate transport and warehouse charges, manage their orders, and get instant answers from a built-in AI agent — all without leaving WhatsApp. Backed by a full admin dashboard for exchange operators.
// Repository: Private — available on request
How a WhatsApp chatbot opened Zimbabwe's commodity market to every trader with a smartphone.
Zimbabwe's commodity exchange was largely inaccessible to small and mid-scale traders. Participating required navigating web platforms or physical offices, creating a barrier for rural producers and buyers who primarily use mobile phones. Logistics costs were also opaque — traders had no quick way to estimate transport or warehouse fees before committing to a deal.
We built ZMX Chatbot — a WhatsApp-first trading assistant that lets any registered user buy and sell commodities, calculate transport and warehouse charges, and manage their orders entirely through WhatsApp chat. A built-in AI agent handles free-text enquiries and guides users through complex workflows. Exchange admins get a full Next.js dashboard for oversight and operations management.
Three integrated components working together to deliver a seamless commodity trading experience.
The core backend is a Python API that serves as the bridge between WhatsApp and the exchange system. It connects to Twilio's WhatsApp Business API to send and receive messages, processes user commands, runs the conversation state machine, and routes requests to the AI agent or the business logic layer. All inbound WhatsApp events hit this service first.
A Microsoft SQL Server database holds all persistent data — registered WhatsApp users, commodity listings, buy and sell orders, warehouse allocations, transport rate tables, and conversation state. The structured relational model supports complex order matching queries and ensures data integrity across all transaction types.
Exchange administrators interact with the system through a Next.js web dashboard. It displays all registered WhatsApp users, active and historical orders, commodity listings, and flagged conversations. Admins can approve listings, manage user access, update logistics rate tables, and monitor exchange activity in real time.
// Both the WhatsApp chatbot and the admin dashboard share the same MSSQL data layer
A walkthrough of the WhatsApp chatbot flow — buying, selling, logistics calculation, the AI enquiry agent, and the admin dashboard.
Demo video coming soon — walkthrough of the full WhatsApp chatbot experience
The admin dashboard and WhatsApp interface. Click any screenshot to view full size.
// ZMX Chatbot WhatsApp Profile
Every component chosen to deliver reliability and accessibility at Zimbabwe's scale.
Everything a commodity trader needs — delivered through a WhatsApp conversation.
Users can browse available commodity listings, place buy orders, and list commodities for sale — all through a guided WhatsApp conversation flow without any app download.
Instant transport and warehouse charge calculation based on commodity type, quantity, and destination. Traders get accurate cost estimates before committing to a deal.
Users can view, track, and manage their open and historical orders directly via WhatsApp. Real-time status updates keep traders informed throughout the fulfilment process.
A conversational AI agent handles free-text enquiries about commodities, pricing, exchange rules, and logistics. Users can ask questions in natural language and get intelligent, contextual responses.
Next.js dashboard giving exchange operators full visibility — registered WhatsApp users, order activity, commodity listings, and flagged conversations — with tools to manage and intervene where needed.
WhatsApp-based registration flow with identity verification. Only approved users can place orders, protecting the exchange from fraud and ensuring regulatory traceability.
Making Zimbabwe's commodity market accessible to every trader with a mobile phone.
The core insight behind ZMX Chatbot is that the best interface for many Zimbabwean traders is not a web app or a native mobile app — it's WhatsApp, which they already use daily. By meeting users where they are, we removed the onboarding friction entirely. The AI agent layer was added to handle the inevitable edge cases that a rigid menu-driven chatbot cannot — turning ambiguous questions into actionable exchange operations.
I'm available for fintech engineering roles and ambitious project collaborations.
Add a YouTube, Vimeo, or direct video URL to the data-video attribute on the button to display it here.