Skip to content

OCR & Document Import

SynkriaOps’s OCR (optical character recognition) engine analyses your PDF, JPEG or PNG files and automatically extracts key accounting fields: net amount, VAT, gross amount, invoice date, due date, supplier name and invoice number. For a clearly printed supplier invoice, extraction takes under 10 seconds and eliminates manual re-entry — the main source of data-entry errors.

Supported formats are: supplier invoices (PDF ≤ 50 MB, JPEG, PNG — native or scanned), native bank statement PDFs (commercial banks) and Mobile Money statements in PDF or image format (Orange Money CM, MTN MoMo). For Mobile Money statements exported as CSV from the operator app, use the direct import in Treasury → Import a Mobile Money statement instead — it is faster and more reliable than OCR on a PDF.

The OCR workflow runs in three stages: File uploadAutomatic extraction (10–30 seconds depending on quality) → Human review (30 seconds for a legible invoice) → Accounting validation in the Purchases or Treasury module. OCR proposes data — validation always remains an explicit human action.

To use OCR in SynkriaOps, you need:

  • An open fiscal year covering the period of the documents
  • The supplier created in your contacts directory (or create it on the fly from the OCR validation form)
  • Files in PDF, JPEG or PNG format, maximum size 50 MB per file

The quality of the scan or photo directly determines extraction accuracy. Here is what affects the result:

Document typeExpected accuracyRecommendation
Native PDF invoice (received by email)98 %+Direct import, quick review
Scanned invoice at 300 dpi minimum93–97 %Check amounts and date
Smartphone photo in good light85–93 %Check all fields before validating
Native Mobile Money PDF statement92–96 %Check transaction fees
Poor-quality scanned statement< 80 %Prefer operator CSV export

If this is your first import, start with Import an invoice via OCR — the guide walks you step by step from upload to accounting validation, with examples of common corrections.