Inovai / Use cases
Accountants still match invoices by hand.
AI compares the invoice with the purchase order and accountants handle only deviations.
Problem
OCR extracts the fields but won’t tell you whether they match the purchase order.
What we’ll do
Extraction, line-by-line matching and approval rules.
Result
Routine invoices go through with one confirmation.
How it works
Four steps; a person decides on the exception.
deviation
Document
Step 2ExtractionHeader and line items: prices, quantities, VAT, due date.OrderfoundItems12 of 12Item 7+4%
Step 3Matching and rulesItem by item against the purchase order, delivery note and limits.1exception
Step 4Approval and exportAccountants confirm invoices without deviations in one go; the data goes to ERP.Verified by a humanWhat changes
Accountants handle deviationsRoutine invoices go through with one confirmation.
Fewer data errorsMatching to the purchase order catches typos and duplicates.
Traceable checksEach invoice shows what was checked and why it passed.
When it makes sense
It pays off when
- Hundreds of invoices a month
- There is a purchase order or delivery note to compare against
- Approval rules can be described
What to watch out for
- Scans with handwritten edits need a wider tolerance
- The workflow flags invoices without a purchase order but doesn’t approve them itself
- Discounts, bonuses and returns need their own rules