Copilot Auto Extract fields

This field extraction method uses a large language model alongside natural language instructions to extract data. You provide a simple description of what data to extract. There is no need for additional configuration, scripts, or training. Recognition is performed only on the first ten pages of a document.

You can add multiple auto-extract fields on a single document. You can also combine auto extract fields with suggested fields and lassoed fields.

The Copilot Auto Extract method is disabled for predefined document types, and test documents.

Tips:

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. Therefore, Pause extraction until you add all your fields.
Sometimes, the suggested field highlighting can interfere with your ability to view the auto extract field highlighting and cause confusion. You can disable the suggested field highlighting by selecting  Hide suggested fields.
If you want to add suggested fields, add them before hiding the highlighted fields.

You can also extract inferred information that is not printed on the document. See Inferred or derived fields for more information.

Currently, this functionality works best for Western languages. The results for other languages, such as Chinese or right-to-left languages, are not guaranteed.