Document Understanding activities

With DocumentUnderstanding.Activities, you can manage documents using a unified approach, by storing every information from the Document Understanding^TM process within a Document Data object. Moreover, DocumentUnderstanding.Activities is integrated with Modern projects, enabling reusability.

Known limitations

We are aware of the current limitations that the DocumentUnderstanding.Activities package has, as we plan to resolve soon. The following features are not yet available:

• Support for splitting documents.
• Business rules.
• Training models.
• Support for models from tenants other than where the automation is deployed.
• Support for Automation Suite.

The sections below describe each phase of the document understanding process using Document Understanding activities.

1. Processing Documents

Processing documents involves preparing the PDF files for extraction. With the Document Understanding activities, you can:

• Extract text, images, specific pages, or merge multiple PDFs.
• Change the password of encrypted PDF documents

To process PDF files with Document Understanding activities, use the following activities:

Activity	Description
Set PDF Password	Changes the password of a specified PDF file.
Merge PDFs	Joins a collection of file objects.
Get PDF Page Count	Provides the total number of pages in a PDF file.
Extract PDF Text	Extracts the text from a PDF document.
Extract PDF Images	The activity extracts all the images it finds in the PDF file.
Extract PDF Page Range	Extracts a specified range of pages from a PDF document.

2. Extracting data

Use the Extract Document Data activity to:

• Extract data from an input file saved as a Document Data object.
• Store the extraction results into the same Document Data object.

Document Data is a resource that serves both as an input and output variable, within your Document Understanding workflows. The Document Data object holds all the necessary information about a single document. If you classify a document, the object includes the Document Type. If you extract data, the object contains the corresponding extracted fields. Irrespective of the activity, Document Data consistently contains the document's text and DOM (Document Object Model).

Provide the file as input only the first time you use Extract Document Data. The output, known as Document Data, should be reused throughout the workflow to prevent re-digitizing the same file, which costs 1 AI Unit per page.

Visit Document Data for more details.

3. Classifying data

Use the Classify Document activity to:

• Choose from various classification models.
• Output the classified data into a Document Data object.

4. Validating data

The validation step of the document process means sending the processed documents for validation with members of your team within Action Center. You can also configure the process of validating documents in Action Center using the following activities:

Activity	Description
Create Validation Task	Creates a validation action to suspend the workflow until it is completed.
Wait for Validation Task and Resume	Pauses the action until validation is complete and then resumes it automatically.
Create Validation Task and Wait	Creates an action in Action Center for visualizing and modifying extraction results, and pauses the workflow until the action completes.
Create Classification Validation Task	Creates an action to verify classified Document Data without waiting for its completion.
Create Classification Validation Task and Wait	Creates an action to verify classified data and waits for its completion before resuming the workflow.
Wait for Classification Validation Task and Resume	Waits for a Classification Validation action to complete before resuming the workflow.