.png)
Maersk
Logistics · Denmark
5-10% fuel saved per voyage
Maersk's NavAssist delivers real-time route advice to captains on 250+ ships while AI agents handle 5,000+ office hours weekly
NavAssist sits on the ship's bridge, giving captains real-time route and speed recommendations based on weather, ocean currents, and cargo. In parallel, autonomous AI agents handle booking, claims, and document processing in shore-side operations.
Friction
Fuel is the largest operational cost and the biggest CO2 source in container shipping. Captains made route and speed decisions based on manual calculations and local weather forecasts. On shore, repetitive backoffice tasks consumed thousands of staff hours weekly.
Breakthrough
NavAssist combines weather models, ocean current data, vessel characteristics, and historical voyage data via ML to generate real-time route and speed recommendations. Backoffice agents built on LLMs with retrieval over Maersk's document library and integrations to booking and claims systems.
Impact
5–10% fuel savings per voyage, measured across 250+ vessels in pilot. 5,000+ hours per week saved in backoffice processes by autonomous AI agents in Q1 2025. At full fleet scale, potentially hundreds of thousands of tonnes of CO2 reduction per year. Maersk's broader AI roadmap targets €250M+ in annual savings.
Unlock the full analysis with breakthrough, impact, what made it smart and its technical approach below!
Problem
A container ship burning heavy fuel oil for two weeks between ports is one of the most fuel-intensive operations in global trade. Even a 5% efficiency improvement at Maersk's fleet scale translates into enormous cost and emissions reductions. Yet routing decisions were still largely manual.
What made it smart
NavAssist is not a planning tool run before departure. It is a real-time bridge-side advisor that updates recommendations continuously as weather and ocean conditions change. The captain retains full authority; the AI provides the data and the recommendation.
Technical approach
NavAssist integrates weather models, ocean current data, and vessel-specific characteristics via ML to generate real-time route and speed recommendations displayed on the bridge. Backoffice AI agents run on LLMs with retrieval over Maersk document repositories and live integration to operational systems.
Strategic lesson
In asset-heavy operations, the highest-ROI AI is the one deployed at the point of the most consequential recurring decision. In Maersk's case, a ship captain's route choice, made hundreds of times per voyage.
Reflection question
In your operations, which recurring decisions have the largest downstream cost or efficiency impact, and are those decisions currently made with the best available data and recommendations?