Use the IQ Bot Validator
Use the Validator to manually validate documents from which IQ Bot could not extract data.
- Data is not readable, such as blurry text or dark background.
- Data is missing from a required field.
- Document failed an automated processing rule.
- Fields did not meet the set optical character recognition (OCR) confidence threshold.Nota: To learn more about how field-level OCR confidence can be used to improve the quality of STP output, see Mejorar la calidad de los resultados mediante la confianza del OCR.
A user with the AAE_IQ Bot Validator role opens the Validator window, where they correct the flagged fields.
Validator interface overview
The Validator interface opens with the first document from the validation queue. The data requiring validation distinguished with red text or outlined in red boxes.
- If no files are available in the validation queue.
- If a file does not exist in the validation queue.
- If another user is working on the remaining file in the validation queue.
- 1. Provide the values for fields IQ Bot could not extract
- Click the empty text field, then select the blue-bounded box in the document that contains the data to extract. After you select a box, the data is highlighted in yellow and a green outline appears, which you can resize or reposition to include all the data that you want to extract.
- 2. Fields awaiting validation are outlined in red
- In the example above, the table fields Item total and Quantity require manual entry.
- 3. Add or delete a row
- To insert or delete a row, hover over the ellipses to show the icons for adding or deleting table rows.
- 4. Mark document as invalid
- Removes the document from the validation queue and saves a copy of the document to the Invalid folder. Click here to mark a document as invalid, then select from the following reasons:
- Fields missing
- Tables missing
- Wrong values
- 5. Guardar documento actual
- After you have validated all the fields, click here so IQ Bot can finish processing this document.
- 6. Skip to next file
- Skip a file without correcting the errors in the current document.
Linked tables in the Validator
Some pages in a document might contain more than one table. When there is a relationship between the tables, such as in the example page below, where a parent table contains patient information and child tables contain additional information related to that patient, these are called linked tables.
In the example page below, the red box outlines the parent table and each of the blue boxes outline a child table.
To successfully validate a page with linked tables, add the fields from the child tables to same table-repeated section as the parent table.
If shared fields between the linked tables do not match values, the Validator does not let you complete validation and save your changes. This functionality helps avoid manual validation errors that can occur when there is a mismatch of field values between the parent and child tables. The Validator shows an error message and prevents the save.
Using the Validator
Open the Validator user interface from the LEARNING INSTANCES tab. If a learning instance contains documents that require validation, there will be a document icon with a red exclamation mark in the Actions column. Click the document icon to launch the Validator.