The challenge
HDG has been a known name in fresh fruit and vegetable quality control for years. From grower to final destination, from citrus to tomatoes, for clients across Europe, Latin America and Africa. The inspections themselves are first-rate. The bigger opportunity sits where plenty of firms, from inspectors to traders, keep hitting the same wall: document processing.
It bites hardest in peak season. New shipments arrive and the packing lists climb just as fast. In the very weeks when the inspection team is flat out on the real work, hundreds of packing lists still need reading and entering into QC4U. Repetitive work, against the clock, asking a lot of people whose expertise is better spent elsewhere.
A client emails a packing list, almost always a PDF. Sometimes one per shipment, sometimes dozens at a time. Someone opens the email, reads the list, picks out the right fields, classifies the document and moves the structured data into QC4U so the inspection and the invoicing can carry on. Each list takes a few minutes. At 200,000 a year, that's a full-time job on its own. It's also the kind of work where mistakes creep in after a few hours: the wrong batch, a typo in the number of colli, a document filed under the wrong trip.
For HDG, doing this better was a given. The real question was how to set it up so the solution grows with the volume while the team stays the same size.
Our approach
- 01
Packing list arrives in the mailbox
The pipeline picks up emails on its own, works out which attachments are packing lists and which aren't, and lifts them out of the message.
- 02
Read and classify automatically
The relevant fields come out of each list. It makes no difference whether the list is from supplier A or B, or whether the layout shifts from one shipment to the next. Earlier tools fell over the moment a supplier changed its template; the model reads a packing list the way a person would, looking at what's on the page and recognising what each field means.
- 03
Structured data out
The raw PDF turns into a row of data: products, quantities, batches, origin, destination, references. Validation sits exactly where errors usually pile up.
- 04
Ready for inspection and invoicing
The data flows through to QC4U, where the inspection and the admin can carry on. Nobody sits there opening packing lists and re-typing fields any more. The team handles the exceptions and spends the rest of its time on the inspection work itself.
The result
Time. The hours that went on reading packing lists go back into the inspection work and other parts of the business. At this volume, that's a big chunk of a full-time role.
Errors. An automated pipeline doesn't make typos and doesn't lose documents, and anything it's unsure about gets flagged for review. That beats manual work, where a mistake can stay buried until someone trips over it by chance.
Scale. 200,000 packing lists is a lot, and the curve keeps climbing. Done by hand, that means finding, training and coordinating more people. Through a pipeline, it just means more documents over the same infrastructure.
The pattern behind the HDG partnership shows up with other clients too: it starts from a specific, everyday pain point, the solution runs quietly in the background, and the pipeline is built from day one for the volume that's actually there. Thanks to Bas Lok and the whole HDG team. Let's do this.