Check what was extracted
Open a document and you see the original page next to the fields Merchaint read from it. This is where you confirm it got things right — and the one screen that decides whether you can trust everything after it.
- A walkthrough video, then why this is the cheapest place to catch a mistake
- Use cases, and which tab to open for what
- How to check a document, and the expected result
- Two things that are easy to miss, then troubleshooting and FAQ
01 · Video
Feature walkthrough
02 · Why
Why do we have this?
Everything after this point treats the extracted values as fact. A wrong total will fail matching weeks later for reasons that look nothing like a reading problem. Catching it here is much cheaper than chasing it down the line.
Why is it laid out this way?
The document and the data sit side by side because checking means comparing. With the page on screen you never have to remember what it said.
03 · Use cases
The first few documents from a new source are worth opening properly. Once you've seen three read correctly you can stop checking every one.
If a record isn't pairing with its purchase order, open the source document and check the invoice number and total were read right before you look anywhere else.
The header totals can be right while the line items are wrong. Expand Line Items and check the count matches the document.
Which tab do I want?
| Tab | Shows |
|---|---|
| Extracted data | The values read from this document, and its line items |
| Extraction Schema | The fields this document type looks for — names, types, and which are required |
| Related Documents | Other documents connected to this one |
04 · How
- Open the document from
Inbox— click the row, or useView Details. - Check the header fields against the page. Invoice number, dates, supplier and totals first.
- Expand
Line Itemsand check the count and the amounts. - If something's wrong, use
Reprocess— see Fix a document that was read wrong.
05 · Expected result
The header shows the status, file size, when it was processed and the document type. The document is on the left, and Fullscreen makes it bigger. The fields are on the right, with line items grouped underneath showing quantity, unit price, amount and description.
A correctly read invoice matches the page exactly, currency included.
06 · Troubleshooting / notes
| What you see | Why | What to do |
|---|---|---|
| "No extracted data" | Nothing could be read from this document | Reprocess it — see Fix a document that was read wrong |
| A banner saying it isn't mapped to a schema | It was read but never filed | Use Choose schema on the banner |
| A field is blank but the page shows a value | The value wasn't where the schema expected it | Reprocess. If it keeps happening for one supplier, tell your admin. |
| Fields this document doesn't have | The wrong schema was used | Reprocess with the right one |
| Line items missing or short | The table wasn't read cleanly — common on scans | Check the image quality, then reprocess and re-read |
| The page won't display | The preview couldn't render it | Download the original to check the file |
07 · FAQ
What are line items?
The individual lines of a document — one per product or service — each with its own quantity, price and amount. They're read separately from the header fields.
What's the timing line at the bottom?
How long each part of the reading took. Mostly useful for understanding why one document was slow.
Common questions
Yes — Reprocess, which also lets you pick a different schema.