← Back to Portfolio

AI Cargo Matching System

An intelligent Gmail-integrated system that automatically matches available ship cargo slots with client cargo requests. Built as an MVP prototype using Next.js, AI, and MongoDB to help logistics brokers find optimal cargo-vessel matches in seconds.

AI AutomationLogisticsGmail IntegrationMVP Prototype

The Challenge

A logistics broker was drowning in email communications:

  • 200+ emails daily: Ship owners sending available cargo slots and routes
  • Client cargo requests: Multiple customers needing to ship various cargo types
  • Manual matching: Broker spending 4-6 hours daily cross-referencing emails
  • Complex criteria: Port routes, cargo type, volume, dates, vessel specifications
  • Lost opportunities: Unable to respond quickly enough to secure bookings
  • Inconsistent formats: Each ship owner uses different email templates

The broker needed a system that could automatically extract information from emails and intelligently match cargo requests with available vessel slots.

MVP Objectives

1

Gmail Integration

Automatically monitor inbox for cargo slots and requests

2

AI Extraction

Parse email content to extract cargo details

3

Smart Matching

AI-powered matching based on multiple criteria

4

Quick Interface

Simple dashboard to review and confirm matches

MVP Features

A focused prototype demonstrating core matching capabilities.

Gmail API Integration

Connects directly to Gmail inbox using OAuth authentication. Monitors specific folders for cargo availability emails and client requests in real-time.

OAuth 2.0 secure access
Real-time monitoring
Label-based filtering

AI Content Extraction

Uses GPT-4 to extract structured data from unstructured email content. Handles various email formats and identifies key cargo information automatically.

Port route extraction
Cargo type identification
Volume & date parsing

AI Matching Algorithm

Intelligent matching based on port routes, cargo type compatibility, volume capacity, date alignment, and vessel specifications using semantic similarity scoring.

Multi-criteria scoring
Semantic route matching
Confidence ratings

MongoDB Database

Stores cargo slots, client requests, and matching history. Flexible schema handles varied cargo specifications and enables fast querying for match candidates.

Flexible schema design
Fast search queries
Match history tracking

Broker Dashboard

Clean Next.js interface showing suggested matches with confidence scores. Brokers can review details, compare options, and take action on promising matches.

Match suggestions view
Side-by-side comparison
Quick action buttons

Intelligent Criteria

Matching considers port routes, cargo type compatibility, volume capacity, shipping dates, vessel suitability, and special cargo requirements automatically.

Route optimization
Cargo compatibility
Volume matching

Technology Stack

Frontend & Backend

  • Next.js 14 - Full-stack React framework
  • TypeScript - Type-safe development
  • MongoDB - Flexible NoSQL database
  • Tailwind CSS - Modern styling

AI & Integrations

  • GPT-4 - Email content extraction
  • Gmail API - Email integration
  • OAuth 2.0 - Secure authentication
  • Semantic Matching - Custom AI algorithms

Prototype Validation

MVP successfully demonstrated core concept and potential time savings.

85%
Time Saved
On manual email matching
92%
Extraction Accuracy
From email content
78%
Relevant Matches
AI matching precision
5-15s
Match Speed
vs 20-30 min manual

Prototype Learnings

Concept Validation

Successfully proved AI can extract structured data from unstructured logistics emails

Time Savings Demonstrated

Broker could review suggested matches in minutes instead of hours of manual work

Matching Algorithm Works

Multi-criteria matching produced relevant suggestions with good accuracy

Ready for Enhancement

MVP provides foundation for additional features like automated bidding and notifications

Need an AI-Powered Logistics Solution?

Whether you need email automation, intelligent matching systems, or custom logistics software, we can build MVPs and production-ready solutions tailored to your workflow.