MarineDesk AI: Bringing AI Automation to the Maritime Industry

The maritime industry moves the world's goods, yet many of its everyday workflows still depend on email chains, spreadsheets and disconnected systems.

After more than 15 years working in marine spare parts procurement and international supplier coordination, I've experienced these challenges firsthand.

The problem we're trying to solve

A single vessel's spare parts requirement can involve multiple suppliers, technical specifications, quotations, approvals, purchase orders and shipping documents.

When information is scattered across different systems, teams lose time searching for historical quotations, comparing supplier offers and coordinating deliveries.

Introducing MarineDesk AI

MarineDesk AI is our effort to build an intelligent enterprise platform connecting six maritime divisions:

• Procurement and supplier collaboration
• Operations and chartering
• Crew management
• Technical management
• Dry dock management
• Marine superintendent workflows

Our goal is to reduce repetitive administration and help maritime professionals make faster, better-informed decisions.

Our technology approach

We're exploring how modern web applications, structured operational data and AI-assisted workflows can work together.

Our development stack includes Next.js, TypeScript, Supabase and Vercel.

Some of the areas we're working on include RFQ information extraction, supplier quotation comparison, historical procurement intelligence and connected vessel workflows.

The challenge isn't simply adding AI to a dashboard. It's making AI useful within real maritime business processes while maintaining human review and control.

Explore the project

We've published an interactive showcase to introduce our direction and the maritime workflows we're designing.

Interactive showcase: https://marinedesk-ai.github.io/

GitHub: https://github.com/marinedesk-ai

We're particularly interested in feedback from developers building enterprise AI applications, maritime technology professionals and people experienced in workflow automation.

What would you prioritize when designing AI software for an industry that still depends heavily on documents, email and manual approvals?