Part ofAI automation
Document and invoice flows
Automated document flows read incoming invoices, orders, delivery notes and contracts, extract the fields that matter and pass them into the business system. What separates this from classic OCR is that a language model can interpret a document whose layout it has never seen, which is what makes the flow useful when you have a hundred suppliers using a hundred different templates.
How extraction works
The document first goes through text extraction, either straight out of the PDF or via OCR if it is a scanned image. The model then receives that text along with a strict description of the fields required, and answers in a fixed format that can be validated. Anything that fails validation goes to manual review rather than into the system.
Checks before posting
Line items are summed and compared against the stated total, company numbers are looked up against the supplier register, and invoices without a matching order are flagged. These checks are ordinary code rather than model judgement, because arithmetic has to be right every time and not almost every time.
What happens when it is unsure
Every extracted field carries a confidence level. Below the threshold the document lands in a review queue with the interpreted value prefilled, so a person confirms rather than types. The threshold is tuned after a few weeks in production, once it is visible where the model actually gets things wrong.
Common questions about this service
Does it work on scanned paper invoices?
Yes, though with lower accuracy than on digital PDFs. Scanned documents go through OCR first, and the quality of that read sets the ceiling on what can be extracted. A skewed or faint scan shows up immediately in the accuracy measurement, and those suppliers are usually better handled by asking for e-invoices instead.
How do we know nothing is posted incorrectly?
Arithmetic and lookups run in ordinary code that gives the same answer every time, not in the model. Documents that fail the checks are not posted; they go to a review queue. On top of that, every extracted field is logged with its source text, so an incorrect value can be traced back to the line it came from.
More in the same area
n8n development
Workflows built in n8n, self-hosted or in the cloud, with code where code is warranted.
Document and invoice flows
Invoices, orders and contracts read, checked and posted without manual entry.
Customer support and ticketing
Incoming tickets sorted, prioritised and answered with the draft already written.
Reporting and data flows
Data pulled, assembled and commented on automatically, on a schedule.
