From PDF Invoice to Structured JSON: The Data Mapping Problem Customs Needs to Solve

Every day, billions of dollars in goods cross borders on the back of unstructured, free-form commercial invoices.

Customs systems need structured data.

The invoice doesn't provide it.

That gap is where delays, errors, and compliance failures live.

I vibe-coded a concept prototype to make that problem visible.

iComply Pro from NaviGate (both fictional) takes a commercial invoice, runs it through AI-assisted OCR extraction, maps the data to structured JSON, flags missing or ambiguous fields, and suggests HS code classifications -- all in a single flow.

The data is illustrative.

The point is the architecture.

What I want you to see:
- The jump from a paper-style invoice to a fully nested JSON object with shipper, receiver, line items, constituent materials, weights, and classification codes is not magic.

It is a data mapping problem. A solvable one. But only if we agree on what the output schema should look like.

That is the deeper question the Global Customs & Trade Forum is working on.

Links to sample invoices to test the iComply app with are included below. Upload one, walk through the analysis, and export the JSON.

Then ask yourself: what would it take to do this at scale, across 200 jurisdictions, with validated data sources behind it?

That conversation is worth having.

Try the demo here: https://lnkd.in/eXrr-_HD

Sample Invoices:

VN to US (Electronics) - https://lnkd.in/eWWde7WE

US to EU (Pharmaceuticals) - https://lnkd.in/epjnrzq4

EU to CN (Machinery) - https://lnkd.in/e2_K-e3T

CA to MY (Aerospace) - https://lnkd.in/ecYT7wEd

Or load one of your own and see how it does!