Manually extract fields using Copilot Auto Extract

Use the procedure in this topic to define Copilot fields manually using Copilot Auto Extract.

If you require only a few fields on a document, use this procedure to manually add fields. If however, there is a lot of data on a document, consider using Copilot proposed fields instead.

Before you begin

Before manually adding extraction fields using Copilot Auto Extract, verify the following steps are complete.

Add at least one configurable document type.
Upload several training and test documents.
Enable field extraction.
See tips in Guidelines for descriptions.

Procedure

If the field extraction screen is not already displayed, select Fields .

The first document type in your solution is displayed in two parts. The document types and their fields appear on the left and the document image for the currently selected document is on the right.
Select a document and then select  Pause extraction.

By default, each time you add a new field, all fields are sent for extraction. This is not ideal when using Copilot Auto Extract because each call to the AI provider costs money.

Pausing extraction enables you to add all of your fields before performing extraction.
Optional. Select  Hide suggested fields.

This option is useful if the list of suggested fields interferes with readability.

You can select  Show suggested fields at any time to view or add suggested fields.
Select  Add Copilot Auto Extract fields.

The Add Copilot Auto Extract fields window is displayed.

You can move or resize this window so you can better view the document image.
In the Add Copilot Auto Extract fields window select Add field.

An editable row is added to the list of Copilot fields.

For the new row, edit the following settings.

Setting	Description
	If this check box is selected, the corresponding field is extracted during production. Clear the field to remove it from the extraction results.
Name	Enter a distinctive and useful name that is displayed if any user actions are configured. See Field name guidelines for more information.
Description	Add a description for the content you are extracting. For more information about useful descriptions, see Guidelines for descriptions. For example, for a field called "Name," add a description such as "The name of the person submitting the form."
Type	The type of data expected by the field. This can be Text, Date, or Number. (Default: Text)
In table	Indicates if a field may be located in a table. (Default: No) If a field you want to extract is likely located in a table, set this setting to Yes. This setting provides location guidance only and does not restrict extraction if a value is found inside or outside a table.

Repeat the previous step to add additional copilot fields.
Optional. If you no longer require a field, select  Delete to remove the field.

The field is removed from the list.
Select Save.

Your new Copilot Auto Extract fields are added to the Fields tab for the selected document type. When a value is found, it is highlighted on the document, and the training circle indicates the field confidence. If the training circle icon is orange, add this field to additional documents.

The highlight might not match the actual content if the information is inferred.
Repeat steps 2 to 9 to manually add Copilot Auto Extract fields to other document types in your solution.

Next steps

Optional. Add suggested fields or lasso fields.
Optional. Edit field properties for your extraction fields.
Configure your import channels.

More info