QDocs
Verified against source (6)
Attachment/Attachment.Search.xsltAttachment/SplitPDF.xsltattachment/Attachment.csExtensions/AttachmentExtensions.csqbo4.Document.AsposePDF/AsposePdfSplitterOptions.csExtensions/AttachmentExtensions.cs
Most of QDocs runs behind the scenes: a workflow splits, classifies, and reads documents as they arrive, and you see the results on the record. This page covers what you’ll see, the document options you can run yourself, and how to work items sent to you for review.
Options appear only where your administrator has turned them on and your role has permission to use them. Menu labels can vary by installation.
What you’ll see on a record
Section titled “What you’ll see on a record”| Where | What QDocs put there |
|---|---|
| Documents panel | One document per document in a package, each filed under its document template. A document split from a package is linked to the original package, so you can always open the source. |
| Document template | The kind of document QDocs decided it is, such as Repair Estimate or Lease. It drives who can see the document, which workflow picks it up, and which worklist it lands on. |
| Description | The document’s text, read by OCR, so you can search for it. |
| Labels | Tags such as a language or a review flag. |
| Scores | The values read from the document, one item per field, each with a confidence. See Read a score. |
Run QDocs options on documents
Section titled “Run QDocs options on documents”Select one or more documents in the Documents panel, then choose from the menu:
- Split PDF breaks a PDF that holds several documents into separate documents. In the dialog you can set where the new files are saved and which queue runs the job. QDocs uses the PDF’s bookmarks when it has them; your administrator sets up how other PDFs are split.
- OCR reads the text of the selected documents into their descriptions.
- Classify works out which document template each file is and sets it. See AI in QBO for end users.
- Transcribe turns an audio recording (mp3) into text on the document.
These options run in the background. Refresh the panel to see the results.
Check extracted values
Section titled “Check extracted values”Each extracted value has a confidence between 0 and 1: how sure the model was that it found the right value. Before you rely on a value:
- Open the score on the document and look at the items with the lowest confidence first.
- Compare each one with the document. An item with no value, or a status such as Multiple Matches or Parsing error, always needs a look. See Confidence and item status.
- Correct or flag what’s wrong, as your process requires, and complete the review task if one was assigned to you.
Work the review queue
Section titled “Work the review queue”Documents and values that need a person are routed to a worklist, usually by a rule on the document type or the confidence of its values. You work them like any other item:
- Choose Next to get the most important item waiting for you. See Get your next item.
- Check the document and its values, correct anything that’s wrong, and update the item’s status to send it on. See Work the item.