COD Manager — Full Order Management Platform
A custom-built web application to centralize order management, logistics tracking, financial analytics, and automation for a multi-country COD e-commerce operation.
The Problem
The business had no centralized order management system. Orders were scattered across multiple Shopify stores with no unified view. Key issues included:
- No interface to manage, modify, or cancel orders across all stores
- No address validation — parcels shipped to wrong addresses, increasing return rates
- No real-time logistics tracking — impossible to know where parcels were
- No accurate profitability data per product — decisions made blindly
- No automated customer communication (emails, SMS, calls)
- Facebook Ads data disconnected from actual sales performance
The Solution
We built COD Manager, a complete web application that replaces all scattered tools with a single centralized platform. Every aspect of the business is now managed from one dashboard.
Results
Key Modules
Financial Dashboard
Complete financial analytics: revenue, COGS, shipping costs, ad spend, ROAS, margins, profit per product. FIFO calculation across 16 SKUs. Margin & Profit Simulator.
Orders Management
Centralized view of all orders across 4 Shopify stores (SK, CZ, HR, GR). Search, filter, modify, cancel. 8,452+ orders with real-time status updates.
Address Verification
Google Maps geolocation integration to detect address errors before shipping. Validation scoring (95-100%) flags risky orders for manual review or call center follow-up.
Logistics Integration
Full logistics provider API integration: real-time parcel status (warehouse, shipped, delivered), delivery times, shipping costs, return rates. AWB tracking per order.
Facebook Ads Integration
Marketing campaigns connected to real sales data. Track ad spend (€259K total), cost per order, ROAS per product, and campaign performance against actual revenue.
Customer Service Automation
Automated emails and SMS for order tracking. Dedicated call center module to contact customers with address issues before parcel shipment.
Automation Pipeline
The COD Manager runs an 11-step infinite automation loop that processes orders end-to-end without manual intervention:
Step 1 — Import Shopify
Sync orders from all 4 stores (last N days configurable)
Step 2 — Sync Logistics
Update AWB statuses from logistics provider
Step 3 — Cancel Duplicates
Intelligent duplicate detection + call center cancellations
Step 4 — Check Addresses & Phones
Google Maps validation + invalid phone number detection
Step 5 — Export
Export valid orders with available stock to logistics
Step 6 — Update Fulfillment
Sync delivery status back to Shopify stores
Step 7 — Financial + FIFO
Daily financial sync at midnight across all orders
Steps 8-9 — Email & SMS
Automated customer notifications for flagged orders
The Platform in Action
Tech Stack
COD Automation System — Email/SMS
Multichannel automation system (Email + SMS) for a Cash on Delivery e-commerce company operating across 12 Eastern European countries.
The Challenge
A Cash on Delivery (COD) e-commerce company operating across 12 Eastern European countries was managing 4 Shopify stores and generating €1.2M in revenue over 6 months.
Their parcel acceptance rate was stuck at 65% — a critical problem in COD, where every refused parcel means a total loss (shipping + return costs). Without any notification system, customers weren't informed about their order status and would forget or refuse delivery.
The Solution
We designed and deployed a complete automation system using n8n that connects all 4 Shopify stores to an internal order management platform, Mailgun (emails) and SMS.to (SMS) to send personalized multichannel notifications at every stage of the delivery journey.
Results
Key Features
Automatic Order Capture
When an order is placed on any of the 4 Shopify stores, a webhook triggers the workflow. The system automatically captures name, email, phone, delivery address, and ordered products.
Personalized Email + SMS Notifications
7 stages covered automatically: Confirmation, Preparation, Shipment, Imminent Delivery, Successful Delivery, Failed Delivery, At Agency/Relay. Fully customizable templates.
Multi-Carrier Tracking
Personalized tracking link generated automatically based on the carrier and destination country. 10+ carriers supported across 12 Eastern European countries.
Control Panel
Manual/Auto toggles for each notification type and workflow. The client manages everything without any technical intervention.
Statistics Dashboard
Real-time metrics: emails/SMS sent, open rates, delivery rates, customer satisfaction. Filter by shop, period, and type. CSV export.
Complete Error Handling
Every email and SMS branch includes error handling with logs and failure notifications, ensuring system reliability around the clock.
The System in Action
Multi-Country Coverage
The system supports 10+ carriers across 12 countries, with automatic generation of personalized tracking links for each shipment.
| Country | Carrier |
|---|---|
| 🇧🇬 Bulgaria | Econt |
| 🇭🇷 Croatia | DPD Croatia |
| 🇸🇮 Slovenia | DPD Slovenia |
| 🇬🇷 Greece | ELTA |
| 🇭🇺 Hungary | GLS Hungary |
| 🇱🇹 Lithuania | Venipak |
| 🇱🇻 Latvia | Venipak |
| 🇪🇪 Estonia | Venipak |
| 🇵🇱 Poland | Inpost |
| 🇷🇴 Romania | Cargus |
| 🇸🇰 Slovakia | SPS Slovakia |
| 🇨🇿 Czech Republic | PPL |
Tech Stack
AI Chatbot — Intelligent Website Assistant
A custom AI chatbot deployed across all pages of smrtflow.io, providing instant bilingual support to visitors using Claude AI and n8n workflow automation.
The Problem
Most website visitors have questions but don’t want to fill out a contact form and wait for a reply. Key challenges included:
- Visitors leave without engaging — no instant way to get answers
- Contact forms have low conversion rates (typically 1-3%)
- Third-party chatbot SaaS solutions cost $50-300/month
- Bilingual audience (French + English) requires support in both languages
- Generic chatbots give scripted, unhelpful responses
The Solution
We built a custom AI chatbot using n8n workflows and Claude API (Anthropic). The chatbot understands context, speaks both French and English, and knows everything about SmartFlow’s services, pricing, and expertise.
The entire system runs on a 5-node n8n workflow that processes messages in real-time, maintains conversation history per session, and responds with relevant, accurate information.
Results
Workflow Architecture
The chatbot runs on a 5-node n8n workflow that handles every message from reception to response:
1. Chat Webhook
Receives incoming messages via HTTP POST from the website’s JavaScript widget
2. Prepare Request
Formats the message with system prompt, conversation history, and SmartFlow context
3. Call Claude API
Sends the request to Anthropic’s Claude Haiku model for fast, intelligent responses
4. Save Response
Stores conversation history in n8n static data for session continuity
5. Respond to Webhook
Returns the AI response to the website widget in real-time
Workflow Overview
Key Features
Claude AI Powered
Uses Anthropic’s Claude Haiku for fast, accurate, and natural responses. Understands context and nuance.
Bilingual FR/EN
Automatically detects the visitor’s language and responds in French or English accordingly.
~$5/month Total Cost
n8n self-hosted (free) + Claude API usage (~$5/month). No expensive SaaS subscription needed.
Session Memory
Remembers conversation context within a session using n8n static data storage. No database required.
Tech Stack
Project Timeline
Morning
n8n workflow design: webhook setup, Claude API integration, system prompt engineering
Afternoon
Frontend widget: vanilla JS chatbot UI, responsive design, dark theme matching site
Evening
Deployment across all 34 pages, bilingual testing, session memory validation
Have a similar project? Let's talk
We design custom automation solutions for your business. Free consultation.
Contact Us