AI OCR for delivery notes and invoices in Business Central

In any company that buys goods, there is a process that consumes hours and creates headaches: checking that what arrives matches what was ordered. A supplier delivery note or invoice arrives, someone picks it up, looks for the corresponding purchase order in the ERP, and starts comparing line by line: is this the item we ordered?, does the quantity match?, is it complete or is material missing? Only when everything matches is the goods receipt posted in the system.

That work —the well-known “matching” or reconciliation of delivery notes and invoices with orders— is slow, repetitive, and prone to errors, especially when item names do not match exactly between the supplier document and the internal catalog. A slightly different description, a reference in another format, or a different line order is enough for the person to have to compare each item manually.

Dijit automates exactly that. It combines intelligent document digitization with automatic matching against the purchase order, so the reconciliation that used to take minutes per document is resolved in seconds, and the person goes from doing the work to simply supervising it.

The big difference compared with a traditional automatic match lies in intelligence. Conventional systems only match lines if the references or descriptions are identical character by character, and in the real world that almost never happens: each supplier names its products differently, abbreviates them in different ways, and orders the lines however it wants.

Dijit's artificial intelligence understands meaning, not just text. It recognizes that two different descriptions refer to the same item and associates the delivery note line with the order line even if they are written differently, with a success rate far above a literal match. And because it accompanies each match with its confidence level, you always keep the final say.