Understand document types and extraction

Overview

• Current document types are Supplier invoice, Expense receipt, Supplier credit note, and Unknown.
• Classification uses email subject/filename/body signals together with structured document extraction results.
• MiniTill can extract fields such as supplier/merchant, document number, date, currency, subtotal/GST/total, and invoice line details where the source supports them.
• Incoming email is checked for spam/virus verdicts before ingestion.
• No supported attachment, unknown document type, or extraction errors are surfaced as document warnings instead of being silently ignored.

When to use this

• Use this guide when a document was classified incorrectly, fields are missing, or staff are unsure whether the extracted data can be trusted without checking the original.

Step-by-step

1. Open the document.
2. Compare the detected type and extracted fields with the original attachment.
3. Review any warnings.
4. Correct document-specific fields in the review workflow where supported.
5. If extraction failed before the review workflow initialized, use Retry extraction.

Common mistakes

• Do not treat OCR/extraction as bookkeeping approval.
• Do not ignore an Unknown type warning and then assume invoice automation will run.
• Do not apply a supplier invoice while pack-size or price warnings are still unexplained.

Troubleshooting

• If a clear invoice is detected as Unknown, inspect the attachment/subject and retry extraction if the original processing failed.
• If extraction repeatedly fails on a valid format, leave it for review rather than inventing missing values.