US Tanker Dashboards: Show What Could Change the Preferred Proposal
Put the deciding assumption beside the preferred proposal
A tanker dashboard should help a US Gulf desk understand why a proposal currently leads and what could change that conclusion. Displaying a favored vessel with a high confidence score can be misleading when the score combines well-established identity information with uncertain commercial availability. The useful output is a conditional recommendation whose important assumptions are visible.
Begin with the decision the user must make. A broker deciding whom to contact next needs different information from a charterer deciding whether to authorize negotiation on revised terms. The dashboard should make that distinction clear. A candidate can be the best next enquiry without being the preferred vessel for commitment.
NIST's explainable AI principles include explanations meaningful to their audience and recognition of knowledge limits. Those ideas support showing the relevant uncertainty alongside a recommendation. NIST explainable AI principles.
Replace a blended score with specific evidence
Imagine a fictional Gulf product cargo with three vessel proposals. One has an attractive price but an unresolved loading window. Another has a higher price and a confirmed commercial response for the requested dates. A third has suitable-looking published particulars but no current offer. A single ranking can obscure why those candidates occupy different stages of the conversation.
Show the freight basis, date condition, response status, and unresolved review separately. Then state the decision implication in ordinary language. The first proposal may warrant urgent clarification because it could become preferred if the dates work. The second may remain the strongest developed commercial alternative. The third may simply justify an enquiry.
Avoid decorative precision. A score of ninety-two does not help unless the reader understands what it measures and how it was validated. It is particularly unhelpful when it encourages users to interpret missing information as a small numerical penalty rather than a condition that could make the proposal unusable.
Use a reversal test
A powerful dashboard question is: what change would reverse the current ranking? Consider entirely hypothetical proposals A and B. A costs $1,000,000 on the stated base case. B costs $960,000 but would incur an additional $25,000 per day under a specifically modeled delay scenario. Assume A's modeled cost remains unchanged, and exclude all other differences for this illustration.
B begins with a $40,000 advantage. After one modeled day, its comparable cost is $985,000 and it still leads by $15,000. After two days, it costs $1,010,000 and A leads by $10,000. The break-even point is 1.6 days. These figures are invented decision inputs, not current freight or delay rates.
The dashboard should not convert this sensitivity into a forecast that two days of delay will occur. It should show that the decision is sensitive to a relatively small change in the assumed period. The next question is whether the team has enough evidence to accept that exposure or should preserve the less sensitive alternative.
Keep forecasts distinct from authoritative status
Weather is one example of information that can be useful yet easily misinterpreted in a commercial display. A forecast layer can inform scenario planning, but it should not be labeled as a terminal closure or an official port instruction. The dashboard needs to state what the source actually describes and when it was obtained.
NOAA explains that the hurricane forecast cone concerns the storm center's track and does not contain all potential impacts. That limits what a simplified map can imply. NOAA forecast guidance.
For the hypothetical Gulf cargo, a planner may decide to test a longer waiting scenario after reviewing official weather information. That is an analyst's commercial scenario, distinct from an official forecast and from any operational direction. Labeling all three separately helps the reader understand why the budget changed without implying an authority has made a decision it has not made.
Make missing information actionable
An unresolved field should point to the question that matters. If the owner has not confirmed a date, say which date and why it affects the comparison. If the cargo quantity is provisional, show the quantities evaluated. If a technical review remains outstanding, describe the commercial dependency without inventing a technical conclusion.
This is more useful than a large general warning panel. Readers can act on a specific uncertainty: ask for a revised offer, clarify an option, or preserve another proposal until the first is resolved. The dashboard's purpose is to direct attention toward the information with the greatest decision value.
Also show which facts are stable across alternatives. If both proposals use the same firm cargo quantity and destination, those inputs do not need to dominate the screen. Highlight the differences that explain the ranking while keeping the complete basis accessible for review.
Evaluate whether the interface improves decisions
Test the dashboard with scenarios in which the cheapest proposal is conditionally attractive, the most complete proposal is more expensive, and the favored alternative loses its availability. Ask users to explain the current recommendation and the event that would change it. Their answers reveal whether the design communicates uncertainty or merely displays it.
Measure correction behavior as well. Can a reviewer change an assumption and see which comparisons must be revisited? Can they distinguish a newly received fact from an analyst's estimate? A visually polished interface that conceals these distinctions can make an incorrect conclusion more persuasive.
Zeaclub's maritime AI ambition can benefit from designing around these practical questions. A useful product direction is to make commercial judgment easier to inspect, not to hide it behind a score. Bring Zeaclub a real decision pattern, with confidential details appropriately protected, and identify the assumption that usually changes the outcome. That provides a concrete basis for discussing a dashboard that helps a tanker desk decide what to investigate next.
Sources
- NIST, Four Principles of Explainable Artificial Intelligence. Explanation audience and knowledge limits; checked October 23, 2026.
- NOAA, Understand Forecast Information. Cone scope only, without current weather or port status; checked October 23, 2026.
Related articles
- 0422: US Tanker Weather Budgets: Model Disruption Without Predicting a Closure
- 0878: US Gulf Hurricane Scenarios: Price Flexibility Without Predicting a Closure
- 0998: US Freight Scenario Dashboard: Show What Can Reverse a Choice


