How receipt recognition works

What happens behind the scenes after you upload a receipt, and what the statuses mean.

After the upload, BizCortex stores the file safely and starts the processing in the background. You can keep working while it runs.

Steps

1. Store
The file is stored in the document storage of your company. The status is "Uploaded".
2. Read the text
The status changes to "Parsing". A text recognition service turns the photo or PDF into text.
3. Extract the data
An AI model reads the text and extracts merchant, address, tax number, date and time, net, tax, tip, gross, payment method, card digits and every line item with category. If text recognition fails for a photo, the AI reads the image directly instead.
4. Save
The data is saved and the status becomes "Parsed".
5. Update the knowledge graph
Finally, the merchant, products, categories, payment method and location are added to the knowledge graph of your company.

Uploaded: The file is stored and waiting for processing.
Parsing: Text recognition and data extraction are running.
Parsed: The data was read successfully.
Failed: Processing failed. Open the receipt to see the error. Try the Reparse button, or upload a better photo.

For PDF files there is no image fallback. If text recognition fails for a PDF, the receipt fails.
The receipt content is sent to the AI service configured for your BizCortex installation.
Automatic reading can make mistakes. Always check amounts and tax before you book them.

Uploading receipts
Upload photos or PDFs of receipts and let BizCortex read merchant, date, amounts and items.
Knowledge graph explorer
Explore the network of merchants, products, categories and payment methods found in your receipts.