The Standards Moment Customs Has Been Waiting For

SaaS vendors are deploying AI across customs and trade. Most are hitting the same wall — not capability, not compute, not commercial appetite. The absence of agreed standards for what customs data is, what it means, and how it should be structured.

This isn't a new problem. Aviation solved it. IATA's data standards underpin every airline system that interoperates globally. Banking solved it. SWIFT and the Basel frameworks gave financial institutions a common language for transactions, risk, and reporting that AI can now reason across reliably.

In both cases, the standard came before the technology could scale. The technology didn't create the standard — the standard unlocked the technology.

Customs hasn't had that moment yet.

Every customs administration, every SaaS platform, every AI deployment is working from its own interpretation of what a trade declaration means, what a consignment is, how origin is asserted and evidenced. The models are capable. The schema beneath them is contested — or absent entirely.

A canon, in this context, is the agreed body of first principles that the field builds on. Not a regulation. Not a vendor specification. The governing layer that makes everything else interoperable.

The Global Customs Canon is that attempt for customs. Eight pillars across three layers and four domains of application. A TechSpec built on a formal Customs Ontology, a Trade Line Assertion schema, a Five-Layer Architecture, and a GraphRAG deployment model. The intellectual foundation that should precede the RFP, the architecture, and the model — not follow them.

The field is ready for this. The technology is waiting for it.

Read it, challenge it, build on it. Download the Canon and artifacts at gctforum.org/gcc.

First published on LinkedIn, May 2026.