What problem does it solve?
A vessel that sails at full speed to arrive as scheduled can still end up waiting, because berth, pilot or nautical service capacity is not ready the moment it arrives. Robin de Puij, Ocean Network Express's Head of Operations in Rotterdam, describes the cost this way: vessels that need to wait and end up anchoring cost a lot of money, and sailing too fast to arrive early results in both that waiting time and unnecessarily high fuel consumption. At Shell's Europoort Terminal in Rotterdam the problem showed up differently: a baseline measurement found vessels waited an average of 210 minutes after finishing cargo operations before they could depart, because the terminal and the shipping agent each planned from their own view of the schedule. More broadly, every party involved in a port call, the shipping line, the terminal, the pilots, tugs and bunkering companies, has traditionally planned from its own view of the schedule, so no one party can see, or fix, the delay building up across the whole call.
How does it work?
- Build one shared timeline per port call. Public vessel tracking data is combined with data shared directly by the shipping line, the terminal and nautical service providers into a single timeline for that vessel's visit.
- Predict arrival and service times. PortXchange, the platform behind both deployments on this page, describes its Synchronizer product's arrival predictions as built on a tree based machine learning model trained on vessel position and route data.
- Recommend a just in time speed. The vessel gets a recommended arrival window so it can sail at a slower, more fuel efficient speed instead of rushing to arrive early and then anchoring.
- Coordinate the terminal and services around the same timeline. Berth planners, pilots, tugs and bunkering companies work from the same shared timeline so capacity is ready when the vessel actually arrives.
- Measure waiting and turnaround time. Anchoring time, berth turnaround time and schedule adherence are tracked per port call against the period before the platform was used.
- Audience
- Employee facing
- Autonomy
- Assist
- Adoption
- Early adopters
- Channels
- Internal tools, API and system to system
What is it worth?
Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Cycle time reduction | Too few to pool | Not pooled: up to 20% | 0plus 1 up to | 1 organization |
Value drivers: Lower cost to serve, Speed and cycle time.
Indicative value
A shipping line with 350 port calls a year at a major container port
USD 50,000 to USD 1.4 million
Waiting and anchoring cost avoided across a year of port calls per year
How this is calculated
Formula: portCalls * costPerHourWaiting * hoursSavedPerCall. The low scenario uses every low input, the high scenario every high input.
| Input | Low | High | Basis |
|---|---|---|---|
| Port calls per year at the optimized port portCalls, port calls per year | 200 | 500 | Editorial assumption, set around the roughly 350 calls a year Ocean Network Express reports at the Port of Rotterdam. |
| Cost of a vessel waiting or anchoring per hour, including charter and fuel costPerHourWaiting, USD per hour | 1,000 | 4,000 | Editorial assumption for a mid sized container or tanker vessel; replace with your own charter and fuel cost. |
| Departure idle time avoided per port call hoursSavedPerCall, hours per port call | 0.25 | 0.7 | Based on the only baseline this page has: a measurement of vessels waiting an average of 210 minutes after cargo operations at Shell's Europoort Terminal before Synchronizer, and Shell's own reported reduction of up to 20% in departure waiting time. 20% of 210 minutes is about 42 minutes, 0.7 hours, so that is the high end; the low end allows for a smaller reduction at other terminals or vessel types, and for the fact that the only quantified result on this page is from a liquid bulk terminal while the reference organization here is a container line. |
What it leaves out: Waiting and anchoring cost only. It leaves out the platform and integration cost, the value of fuel saved by sailing at a more efficient speed, and any benefit to the terminal or port itself; neither deployment on this page reports a single dollar figure that this calculation could be checked against.
Who already uses it?
2 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.
Ocean Network Express (ONE)
Singapore · Logistics and transportation · 2019
Ocean Network Express, one of the world's largest container shipping lines with around 350 calls a year at the Port of Rotterdam, ran a two month trial of the PortXchange platform (formerly known as Pronto, a development of the Port of Rotterdam Authority) in 2019 to monitor, analyse and optimize the approach and handling of its vessels there, then continued using it afterwards. Its Rotterdam operations team, led by Robin de Puij, worked alongside ONE's London vessel scheduling team on the platform's shared timeline to plan just in time arrival and to track the time between the end of terminal operations and a vessel's departure, aiming to reduce both anchoring time and the time spent in port after cargo work finishes.
No outcome disclosed.
Shell
United Kingdom · Energy and utilities · 2018
Shell's shipping and maritime business piloted PortXchange Synchronizer, the platform that began as Pronto, built by the Port of Rotterdam Authority and spun out as the separate company PortXchange in 2019, to close the gap between the end of cargo operations and departure at its Europoort Terminal in Rotterdam, where a baseline measurement found vessels waiting an average of 210 minutes before leaving. Shell Europoort Terminal, the Loodswezen pilots and the Vopak agency shared their planning data on the platform, which PortXchange describes as using a machine learning model to predict vessel arrival times and which sends notifications and warnings of planning conflicts before they occur. PortXchange reports that since the pilot, Shell has been using the Synchronizer platform across different terminals, geographies and business units internally.
- Cycle time reduction: up to 20%, pilot phase, reported 18 April 2018
"We have reduced the waiting time up to 20% for departing ships."
Claimed by: organization
How do you implement it?
A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.
Data you need
- Vessel positions and estimated arrival times from an automatic identification system feed
- Terminal, berth and nautical service schedules shared by the parties involved in each call
- Historical port call data by vessel type and terminal
- Weather and tidal data affecting the vessel's approach
Systems to integrate
- Port community system or a dedicated port call data platform
- Terminal operating system
- Vessel traffic and automatic identification system data feed
- Shipping line voyage planning system
Complexity: Medium
The prediction models are a well understood forecasting problem once vessel tracking data is available; the harder part is getting the terminal, pilots, tugs, bunkering companies and the shipping line to all share their own timing data into the same platform, since the forecast only improves as more of them do.
- 1
Start with one shipping line and one terminal relationship
PortXchange's account of Ocean Network Express's adoption in Rotterdam describes its Rotterdam operations team working alongside ONE's London scheduling team on the platform, while the company's Singapore head office followed the work with interest; neither source says ONE rolled the platform out across its wider network.
- 2
Get vessel tracking and terminal timing into one timeline first
A shared, continuously updated timeline for arrival, pilotage, berthing and departure is the foundation; the just in time speed recommendation only becomes trustworthy once that timeline is reliable.
- 3
Bring nautical service providers into the same platform
Bunkering, towage and waste collection times shape when a vessel can actually leave a berth; a timeline that only covers the terminal misses a real source of departure delay.
- 4
Recommend a speed, do not set one
Give the vessel's master and the shipping line's planners a recommended arrival window and let them decide the actual sailing speed and route, since safety and weather judgment stay with the ship.
- 5
Measure at the level of the whole port, not one vessel
Track anchoring time, berth turnaround and schedule adherence across all participating vessels, since optimizing one ship's arrival can simply move a delay onto another one if the wider port picture is ignored.
Guardrails
- Nautical safety rules and pilotage requirements always take precedence over a recommended arrival window
- A vessel's master and the port authority confirm the final approach plan; the platform only recommends it
- Data shared by one party in the timeline is used only for coordinating that port call, under agreed data sharing terms
KPIs to instrument
- Anchoring or waiting time per port call, before and after joining the platform
- Berth turnaround time and schedule adherence
- Fuel burned or emissions per port call
Human in the loop
Vessel planners, port agents and the vessel's master review the recommended arrival window and service schedule, and pilots and terminal planners confirm the final berth and service plan before the vessel changes speed or course.
Common failure modes
- A timeline only one party keeps updated
- The forecast is only as good as the data shared into it; a terminal or service provider that does not update its own timing breaks the plan for every other party relying on it.
- Optimizing one vessel at the cost of the queue
- Bringing one vessel in earlier can crowd a berth or a pilot slot another vessel needs; plan and measure at the level of the whole port, not one ship at a time.
What are the risks and rules?
EU AI Act
Minimal risk
Annex III point 2 lists critical infrastructure sectors that carry specific obligations: digital infrastructure, road traffic, and the supply of water, gas, heating or electricity. Port call and vessel arrival scheduling is not one of the sectors listed there, and the system does not decide on a natural person's access to an essential service, so it carries no specific EU AI Act obligation beyond the Article 4 AI literacy duty that applies to all providers and deployers. A port that later applies the same kind of model to safety critical vessel traffic management should reassess the tier for that specific use.
Rules that apply
Controls to put in place
- Data sharing agreements that state which timeline data each party may see and use, and for how long
- Annual review of forecast accuracy against actual port call outcomes, shared with the participating parties
Frequently asked questions
- Does the AI control the vessel's speed?
- No. Design it so the platform surfaces predicted timing and a recommended arrival window, and lets the vessel's master, the shipping line's own planners and the port's nautical service providers decide the actual speed, route and berth plan. Safety and pilotage judgment should always stay with the people responsible for the vessel, not the software.
- What results have shipping companies reported?
- Shell reported reducing the waiting time for its departing vessels by up to 20% after a pilot at its Europoort Terminal, where vessels had waited an average of 210 minutes after cargo operations before departing. Ocean Network Express, which adopted the same underlying platform for its Rotterdam port calls, reports early improvements in waiting time without giving a checked figure, so treat that second result as directionally positive rather than quantified.
- Does this only work for one port?
- No, but the quantified result on this page is from Rotterdam. PortXchange reports that since the pilot, Shell has been using the Synchronizer platform across different terminals, geographies and business units internally, though no number has been published for that wider use. The underlying platform began as Pronto, built by the Port of Rotterdam Authority, which spun it out into the separate company PortXchange in 2019 specifically to offer it to ports worldwide, naming Shell as one of its first partners for that expansion.
How to cite this page
Blits.ai AI Use Case Library, "AI port call optimization for vessel arrival, berth and turnaround planning", last verified 29 September 2026, https://www.blits.ai/ai-use-cases/port-call-and-vessel-operations-optimization. Licensed under CC BY 4.0. Method: how we verify use cases.
Changelog
- 29 September 2026: First published