bezgoFresh
The Operations Engine
bezgoFresh ran its entire delivery operation over WhatsApp chat — orders, menus, billing and delivery assignments all lived in one stream. I built a single bot to carry that work, so the team could focus on serving customers instead of holding the operation together.
The business worked — but only because people spent hours every day holding it together. Every order passed through multiple hands, and each hand-off created mistakes: wrong quantities, missed items, duplicate orders. I built a 'conversation-to-operation' engine that turns one WhatsApp message into a complete order-to-delivery pipeline, so the system carries more of the work.
The Challenge
The Bottleneck
Orders arrived through WhatsApp. Menus changed every morning. Vendors updated prices manually. Operations teams copied orders into spreadsheets and created invoices by hand. Delivery assignments happened over phone calls. It worked — but only because people spent hours every day holding it together, and every new order created more manual work.
The Solution
The Reality
Instead of asking people to work harder, I made the system carry more of the work. The bot parses orders from real-time WhatsApp messages, regenerates menu cards from live pricing, issues invoices and payment links automatically, and syncs every order to fleet and operations dashboards — turning WhatsApp chaos into a streamlined business system.
"I didn't just build a chatbot, engineered an operations engine that turns scattered WhatsApp messages into a complete, auditable order-to-delivery pipeline."
Under the Hood
Technical Execution
A breakdown of the core systems engineered to deliver reliability and speed.
Intelligent Order Parsing
- Built a hybrid location extractor that pulls GPS coordinates from Google Maps links with Haversine fallback.
- Implemented Fuse.js fuzzy matching to resolve shop and item names from free-text messages.
- Achieved accurate single-shop and multi-shop order parsing with live price lookups.
Automated Billing & Menus
- Engineered Canvas-generated menu cards with live pricing from Google Sheets.
- Integrated Zoho Books to generate invoices and payment links in seconds.
- Built multi-shop billing logic with delivery charges, discounts, and tax exemption handling.
Resilient Operations
- Managed WhatsApp sessions with QR-based pairing and automated session recovery.
- Containerized the bot with Docker and deployed to Railway with health checks and restart policies.
- Synced every order to fleet and butcher dashboards in real time.
Want to build something like this??
Reach out to me →The Outcome.
Went from manual WhatsApp juggling to a fully automated order-to-delivery workflow — menus regenerated on demand, invoices issued in seconds, and every order logged to Google Sheets in real time.
- 24/7 Order Processing
- Automated Invoicing
- Unified Operations
What I learned
This project taught me that chaos is the best product spec.
Automation only earns trust one order at a time — the system works when the humans stop noticing it.
Every messy WhatsApp thread was a requirement I hadn't written down yet.