Shipping Fleet AI

Fleet Ops AI is a multi-agent decision-support prototype for maritime fleet-management teams, comprising an Engineering Superintendent, Marine Superintendent and Fleet Manager. When a vessel reports a problem, technical findings, operational constraints and commercial information must be brought together through calls, emails, meetings and repeated updates. Fleet Ops AI coordinates that process: Engineering assesses the technical evidence, Marine evaluates voyage and commercial implications, and the Fleet Manager reviews their findings, requests follow-ups and considers priorities across the fleet. The demonstration uses Aster Tide Marine, a fictional company, and MT Aster Horizon, a fictional oil tanker. Agents consult selected OEM manual and Safety Management System excerpts and produce a concise decision brief, source-backed guidance and notification drafts. Built during the hackathon, the application includes case submission, agent handoffs, source-quotation checks, structured outputs, bounded correction attempts and benchmark scenarios. It aims to reduce coordination effort and support more transparent, informed decisions. Time savings and decision-quality improvements remain to be independently evaluated. No vessel commands or external messages are sent automatically.