TURI by Stoopa.AI — Vessel Operations

Smarter Berths, Safer Ports: How TURI's Automated Berth Allocation Rewrites the Schedule

Half of all maritime incidents happen inside port and terminal boundaries — and a huge share of that risk traces back to one thing: how berths get scheduled. TURI's AI-powered berth allocation, with full manual override, turns the most political spreadsheet on the port into a continuously optimized, auditable plan.

By the Stoopa.AI team · 5 June 2026 · 8 min read

The berth plan is the most expensive spreadsheet in shipping

Walk into the operations room of almost any port in the world and you will find the same scene: a senior planner, a wall of monitors, an Excel sheet that has been edited so many times its formulas have stopped working, and a phone that does not stop ringing. This is the berth plan. It decides which ship goes where, when, for how long, with which cranes, and at what cost. And it is almost always wrong by the time the next vessel transmits an updated ETA.

The cost of getting this wrong is not theoretical. It shows up in three places that every port leader already feels:

~50%
Share of maritime incidents in 2022 that occurred inside port and terminal boundaries — at berth, anchorage, or in harbour transit [1]
+10%
Year-over-year rise in global shipping casualties in 2024 (3,310 vs. 2,963) [2]
41.8%
Share of accidents in container terminals attributable to traffic incidents — the single largest category [3]

Berth allocation sits upstream of all three numbers. A poorly sequenced berth plan creates anchorage queues, which lengthen voyage times, which push more vessel hours into the port, which crowds the yard, which forces emergency truck moves, which is exactly when the gate-area accidents happen. A late berth assignment forces a tug crew to maneuver under time pressure. A sub-optimal crane split forces gangs to rush and inflates near-misses. Maritime cyber incidents — up 103% in 2025 — increasingly target the scheduling systems themselves [4]. The berth plan is not just a productivity artifact. It is a safety document.

Why the berth allocation problem is genuinely hard

The Berth Allocation Problem (BAP) is one of the most studied optimization problems in operations research [5]. It is hard for a reason: it is multi-objective, dynamic, and political. The planner is trying to minimize vessel waiting time and minimize total service time, while balancing crane availability, draft restrictions, tide windows, labor rosters, customer SLAs, demurrage exposure, and the fact that the vessel that just emailed an ETA update is the one paying the highest rate this quarter. Every constraint moves in real time.

Doing this by hand has predictable consequences. Academic and industry analyses of port-call optimization consistently find double-digit waste: up to 20% fuel savings per voyage are available simply by aligning vessel arrivals with realistic berth availability rather than forcing ships to "rush to wait" at anchor [6]. That fuel waste is the same waste that shows up as emissions in the host city's ESG report — the berth plan is also a sustainability document.

The berth plan is the single most leveraged decision a port makes every day. It is also the one most ports still make in a spreadsheet.

The hard truth: the human planner is not the bottleneck because they are not skilled. They are the bottleneck because the problem they are being asked to solve in real time has a search space larger than chess, and the data they need is scattered across the TOS, the AIS feed, the customer email thread, the harbour master's WhatsApp, and a tide table taped to the wall.

What automated berth allocation actually does

Modern automated berth allocation is not a replacement for the planner. It is a tireless, data-fluent assistant that produces a recommended plan, explains it, and lets the planner override anything they want. The published evidence on what well-designed automation delivers is striking — and consistent across very different ports.

Port of Hong Kong
−28% vessel waiting time
A two-stage "predict-then-optimize" framework — first predicting actual vessel arrival times, then optimizing the berth plan against those predictions — cut vessel waiting time by 28% versus the port's actual operational plan [7].
Busan, South Korea
+79% ship punctuality, ~$7.3M additional annual revenue
An AI-driven scheduling environment — projecting vessel arrivals days in advance, calculating optimal fuel use, and flagging safety incidents before they occur — was projected to deliver a 79% improvement in ship punctuality and roughly USD 7.3M in additional direct annual revenue at a single hub [8].
Yangtze River Delta (10 ports, 49 areas, 2,180 berths)
−20% wait time, +15% berth utilization
An intelligent berth recommendation framework using spatiotemporal knowledge graphs and AI agents reduced ship waiting times at anchorages by ~20% and increased berth utilization by ~15% across a network of 2,180 berths [9].

These are not lab numbers. They are operational results from working ports. And they tell a consistent story: the upside from getting berth allocation right is so large that the question is no longer whether to automate the decision, but how to automate it without losing the planner's hard-won judgment.

How TURI does automated berth allocation

TURI's Vessel Operations module is built around a single conviction: the best berth plan is the one a senior planner would make if they had infinite time, perfect data, and no phone. TURI is engineered to be that planner's twin — fast, never tired, and always auditable.

1. Continuous, data-grounded recommendations

TURI ingests the live signal a planner would want but cannot watch all at once: AIS-derived ETA predictions, vessel particulars, declared cargo profiles, draft and tide windows, crane availability and split options, gang rosters, yard congestion levels from the Gate & Yard module, customer SLAs, and the financial profile of each call from the Billing Integration module. Every time any of these inputs changes — an ETA slips, a crane goes amber, a tide window narrows — TURI re-solves the berth allocation problem and surfaces the new recommended plan, with the delta from the previous plan called out.

2. AI optimization, not just visualization

Most "berth planning" tools on the market are timeline visualizers — a prettier Gantt chart on top of the same human guesswork. TURI is different. The recommended berth plan is the output of an actual optimization, balancing minimization of total vessel waiting time, maximization of berth and crane utilization, adherence to draft and tide constraints, and respect for customer-tier service commitments. Predictive Analytics ("what if" scenarios) lets ops leaders ask questions like "what if the 06:00 ETA slips three hours?" and see the propagation across the next 72 hours of berth occupancy.

3. Manual override is a first-class citizen

Here is what most automated berth allocation systems get wrong: they treat the human as an exception handler — someone who is allowed to "veto" the AI on a bad day. TURI takes the opposite view. The planner is the operator of record. They can accept the recommended plan with one click, modify any individual berth assignment by drag-and-drop, override TURI completely and create a manual allocation, or pin specific assignments and ask TURI to re-optimize around them.

Every override is captured: who made it, when, why (free-text or selected reason), and what TURI predicts the impact will be on waiting time, utilization, and revenue. Over time, this becomes a goldmine — TURI learns from the planner's judgment, not just the historical data, and the recommendations get better at reflecting how this port actually runs.

Why the override matters. Berth allocation is full of context the data cannot see — a VIP customer call, a known issue with a tug crew, a labor dispute that has not made it into the system yet. TURI is designed to assist the planner, not to second-guess them. The audit trail on every manual change is what turns "the AI gave us a plan" into "we have a defensible plan we can put in front of regulators, customers, and our own board."

4. Safety baked into the same plan

Because TURI's berth recommendations are computed against the same live data feeding Anomaly Detection and the Control Center, safety constraints are not an afterthought — they are part of the optimization. Tide-window violations, crane-clearance conflicts, simultaneous bunkering and hot-work restrictions, and over-allocation of pilots or tugs all get surfaced before they make it into a published plan. Given that 50% of maritime incidents happen inside port boundaries [1], moving safety upstream into the scheduling layer is one of the highest-leverage interventions a port can make.

Auto vs. manual: how the two modes work together

The most common question we get from operations leaders evaluating TURI is: "if we turn on automated berth allocation, do we lose control?" The honest answer is no — TURI is explicitly designed so that "automated" and "manual" are not a binary. They are two ends of a dial the port chooses where to set, per shift, per terminal, per vessel class.

Capability
TURI Automated Mode
TURI Manual / Override Mode
Berth assignment
Recommended by TURI's optimizer
Set by planner; TURI advises
Re-plan trigger
Continuous, on every data change
On planner request
Constraint handling
Hard + soft constraints enforced
Hard constraints still enforced; soft can be waived with reason
Audit trail
Every recommendation versioned
Every override logged with author and rationale
Outcome learning
Plan vs. actuals scored automatically
Override outcomes feed back into the model

In practice, most TURI ports settle into a working pattern within the first quarter: TURI runs the routine planning continuously, the planner spends their time on the 10–15% of decisions that are genuinely judgment calls, and the override audit trail becomes the single source of truth when finance, customers, or regulators ask why a particular assignment was made.

The numbers a port should expect

Combining the published academic evidence with TURI's own deployment characteristics, the order-of-magnitude impact a port can plan for in the first 12 months of running TURI's automated berth allocation looks like this:

−15 to −28%
Vessel waiting time reduction (range from published AI berth-allocation case studies) [7][9]
+15%
Berth utilization uplift, observed across 2,180 berths in published research [9]
~20%
Fuel saving available per voyage from realistic, optimized port calls [6]
14 days
TURI standard go-live — about 75% faster than typical port-software implementations

The fastest payback rarely comes from the headline efficiency number. It comes from the second-order effects: fewer demurrage disputes (because TURI's plan is auditable), fewer near-misses at the gate (because crane and yard congestion is balanced upstream), lower pilot and tug overtime (because the schedule respects their rostering), and a real reduction in unplanned overtime for the planner themselves — who can finally take a Saturday off.

How to roll out TURI's automated berth allocation

The teams that get the most out of TURI do not flip a switch labelled "automate everything" on day one. They run a deliberate, three-phase rollout that mirrors how the platform itself learns.

Phase 1 — Shadow mode (weeks 1–2)

TURI ingests the port's live data and produces a recommended berth plan in parallel with the planner's own plan. Nothing changes operationally. The team compares TURI's recommendations against the actual plan every shift and tunes any port-specific constraints. By the end of week two, TURI's recommendations and the planner's plan are typically converging on the routine calls and diverging on the genuinely contested ones — exactly the pattern you want.

Phase 2 — Recommend-and-approve (weeks 3–8)

TURI's plan becomes the default. The planner reviews and either accepts (one click) or overrides (with reason). This is where the override audit trail starts compounding into real intelligence: the port discovers which constraints were missing, which customer rules were undocumented, and which planner habits were tribal knowledge. The model learns; the planning gets faster; the meetings get shorter.

Phase 3 — Automated with exception escalation (week 9+)

For routine vessels and standard berths, TURI's plan is automatically published. For exceptional cases — VIP vessels, tide-tight calls, anomaly-flagged operations, financial outliers — TURI escalates to the planner with a recommended action and a one-click approval. The planner moves from "doing the plan" to "owning the plan." The port moves from reactive berth allocation to predictive berth allocation — exactly the shift TURI is built to deliver.

One thing automated berth allocation will not fix. If a port has unresolved upstream issues — chronic understaffing, broken billing, a yard that cannot handle the throughput a smarter berth plan unlocks — TURI's recommendations will surface those constraints sharply and quickly. That is a feature, not a bug. But it does mean that the conversations the platform triggers in the first 90 days are sometimes uncomfortable. The ports that get the biggest gains are the ones that lean into those conversations rather than around them.

From spreadsheet to system of record

The deepest change TURI's automated berth allocation drives is not the percentage points on waiting time. It is the change in what the berth plan is. It stops being a fragile, single-author artifact that lives on one planner's laptop and becomes a continuously optimized, fully auditable system of record — one that integrates safety, finance, customer commitments, and operational reality into a single decision the whole port can trust.

That is the gap between a port that is running well today and a port that will still be running well in five years, when volumes are higher, vessel sizes are larger, regulatory disclosure is stricter, and the next disruption — whether a typhoon, a cyber incident, or a sudden Red Sea reroute — arrives unannounced. TURI is built for that port.

See TURI's automated berth allocation on your port's data

AI-powered. Manual override always on. Built to integrate, not replace. Go live in 14 days. Average ROI payback in nine months.

Book a demo →
sales@stoopa.ai · www.stoopa.ai

Sources

  1. Rightship — Half of maritime incidents in 2022 occurred in ports and terminals.
  2. Allianz Commercial — Safety and Shipping Review 2025 (3,310 reported casualties in 2024 vs. 2,963 prior year).
  3. MDPI — Risk Assessment of Work Accidents in Container Terminals (traffic accidents at 41.8% of incident categories).
  4. SAFETY4SEA — Maritime cyber incidents jumped 103% in 2025.
  5. Wikipedia / academic overview — Berth allocation problem; Oxford Academic — Berth allocation and scheduling at marine container terminals: state-of-the-art review.
  6. AXSMarine — Port Call Optimization: How Smart Data Cuts Time and Fuel Waste (up to 20% fuel savings per voyage).
  7. ScienceDirect — Dynamic berth allocation based on vessel arrival time prediction (Port of Hong Kong, −28% waiting time).
  8. Sumo Analytics — From Reactive to Predictive Port Management (Busan +79% punctuality, ~$7.3M revenue).
  9. IEEE — Learning-driven berth allocation optimization (Yangtze River Delta: −20% waiting, +15% utilization, 2,180 berths).
  10. Product details (Vessel Operations module, Predictive Analytics, Anomaly Detection, Control Center, ROI metrics): TURI Digital Brochure — Stoopa.AI.