Loamist is bringing a real, live study of thousands of export L/Cs to Sibos 2026 in Miami. Pete Christensen and Nick Pachnev will be at the Miami Beach Convention Center from 28 September to 1 October.

The study

We ran an AI engine against thousands of real letters of credit transactions spanning all major trade corridors. We checked actual L/Cs for workability against UCP 600 and ISBP 821E, and verified documentation against each credit's terms. Every result is reproducible: the same engine, the same rules, the same method on every credit, so the findings can be interrogated rather than admired.

The Hormuz before-and-after

The study also includes a before-and-after dataset analysis of the Strait of Hormuz disruption and its impact on L/C workability, presentation and payment cycles. When a corridor is disrupted, credits that were workable last month stop being workable: shipment dates slip past latest shipment, routings change the ports on the transport document, and presentation periods that assumed a direct sailing no longer hold. The data shows how quickly that propagates into refusals and delayed payment, and which clauses broke first.

Fine-tuned engine versus general LLMs

If you want to see a cost and accuracy comparison of a fine-tuned AI engine against general-purpose large language models on the same presentations, and how the Hormuz disruption affected each, book a session. We will show the numbers, the method and the misses on both sides.

What a session covers

Thirty minutes, structured around whatever you bring. A typical session runs in three parts. We open with the headline findings from the study: refusal rates by corridor, the clauses that most often make a credit unworkable, and how long a discrepant presentation actually delays payment in the data. Then we go to the Hormuz before-and-after and show which terms failed first when transit times moved. We close by running Validator live on a credit, yours if you bring one, and walking through the decision trace behind each finding so you can judge the reasoning rather than the verdict.

What to bring

  • A letter of credit, ideally one that gave your team trouble. A redacted MT700 is enough for a workability read.
  • A document set, if you want to see a full pre-check rather than a workability check alone.
  • A question about a corridor. If your exporters or clients ship through a route that has been disrupted this year, we can look at how credits on comparable routes behaved in the dataset.

If you would rather not share documents, we run the same walkthrough on credits from the study.

Why a live study rather than a demo

Every vendor at Sibos will show a demo, and every demo works. A study on live credits is different in kind: the credits were not chosen to flatter the engine, the outcomes are the banks' real decisions, and the misses are in the data alongside the hits. We think that is the only honest way to talk about accuracy in document examination, and it is the conversation we want to have with trade operations people who know how the work actually goes wrong.

Book a meeting

Book your meeting at loamist.com/sibos. If you would rather just find us, we are on the floor all four days.

Book your Sibos session

Sessions with Pete or Nick run 30 minutes at the Miami Beach Convention Center. We will walk through the study, the Hormuz before-and-after, and the cost and accuracy comparison, and run Validator on a credit you bring.

Run a free workability check on a recent L/C, or book a demo and we will walk through a live decision trace.