Loamist is attending Sibos 2026 in Miami from 28 September to 1 October. Pete Christensen and Nick Pachnev will be on site for all four days.

What we are bringing

We are bringing a live study to Miami: thousands of live L/C transactions, with findings measured under a repeatable methodology. Not a curated demo set. Real credits, real document sets, across all major trade corridors, examined the same way every time so the results can be compared and challenged.

What Validator is

Validator is a documentary credit examination engine. It checks export letters of credit for workability against UCP 600, eUCP and ISBP 821E, flags the terms that cannot be satisfied as drafted, and then checks the document set against the credit. Every finding carries a decision trace back to the article or practice it relies on, so a checker can accept or reject the result rather than take it on trust.

For a bank audience that last point matters most. An automated examination result is only useful to a trade operations team if it can be verified faster than it can be redone. A trace to UCP 600 article 14 or an ISBP paragraph, with the document and field named, is what makes that possible.

What we will show

  • How the engine performed against live credits and global shipping corridors, with the methodology.
  • A cost and accuracy comparison of a fine-tuned examination engine against general-purpose large language models on the same presentations.
  • How a major shipping disruption changed workability, presentation and payment cycles in the dataset, before and after.

How the study was run

Every credit in the dataset went through the same three passes. First, the L/C text itself was read for workability: shipment and expiry dates against the corridor's realistic transit and production times, the document list against what the named parties can actually issue, and the clauses against UCP 600 and ISBP 821E for internal contradictions. Second, each document set was examined against the credit, field by field and across documents, with every finding tied to the article or practice it relies on. Third, the results were compared with what happened at the bank, so the accuracy figures are measured against real outcomes rather than against our own opinion of what a checker should have found.

Because the method is identical for every credit, the numbers can be sliced by corridor, by issuing bank, by document type and by month, and any single finding can be pulled up and argued about. That is the part we most want to do in person.

Who should book

  • Trade operations and document-checking leads who want to see how an automated examination result can be verified faster than it can be redone.
  • Trade product and innovation teams weighing a fine-tuned examination engine against a general-purpose model, and who want the cost and accuracy numbers side by side.
  • Corporate trade-finance and treasury teams whose exporters carry the refusal cost, and who want to know which clauses break first when a corridor is disrupted.

Book a time

If you are attending and want to see how the engine performed against live credits, book a session with Pete or Nick at loamist.com/sibos. Sessions are 30 minutes and we are happy to run Validator on a credit you bring.

Meet us in Miami

Sibos 2026 runs 28 September to 1 October at the Miami Beach Convention Center. Pete and Nick are booking 30-minute sessions to walk through the live study and run Validator on a credit of your choosing.

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