A German air cargo handling company ran 4 kinds of shipment paperwork through super.AI from 2024 to 2026, and its team reviewed the pages that didn't pass a check.
Every air cargo shipment that comes through a handler's warehouse brings paperwork with it. There's a proof of delivery when freight changes hands, a pouch of documents that travels with the air waybill, and manifests on top of that.
Someone has to read the air waybill number, the piece count and the weight off each document and make sure they match the shipment record. If a waybill number or a weight is off, somebody ends up chasing it later.
That's the job we took on.
Documents reached us 3 ways: an email inbox, SFTP file transfers and our API. However they arrived, each one went through the same steps.
First we classified the document so we knew what we were looking at. Then we pulled out the fields that matter in air cargo, like the air waybill number, the number of pieces and the weight.
Before any of that data moved on, we checked it against the handler's shipment records. If something didn't match, the document went to a person in one review screen.
It started with proofs of delivery and delivery notes. In late 2025 the handler added 3 more document types: pouch documents, import handover documents and check manifests.
The first pages ran in September 2024, and volume built up fast. By January 2025 we were processing more than 13,000 pages a month, and for the rest of 2025 it ran at 16,000 to 20,000 a month.
From February 2026, monthly volume roughly doubled. Every month from February to September 2026 came in above 30,000 pages, and the busiest, May 2026, hit 39,030.
By our billing records, that added up to 501,976 pages over the 2 years, each one classified, read and checked before anyone on the handler's team had to look at it.
If your team reads waybill numbers and weights off delivery paperwork and checks them by hand, that's the work this handler handed to us. Here's more on supply chain documentation automation.
