Create a learning instance in Document Automation

Begin processing documents by creating a learning instance to extract data from invoices, utility bills, or receipts. A learning instance is a structure that holds information such as document type, language, and the fields to be extracted.

Prerequisites

To create a learning instance, you must be a Learning instance creator user. See Document Automation users.
To create a learning instance that uses a Google Document AI model, you must create a credential and locker to hold your security token. See Create credential and Create locker.
Note: When configuring the locker, you must add the aari-document-processor custom role as a locker consumer.
The default OCR is ABBYY FineReader Engine. Alternatively, Cloud Control Room users can create a learning instance to process documents in Google Vision OCR.

Watch this video for the complete end-to-end process of creating a learning instance:

Procedure

From the Control Room home page, navigate to Manage > Learning Instances > Create Learning Instance.
Enter a name and description for the learning instance.
Document Automation does not allow duplicate learning instance names, so the name you provide must be unique.
Select the document type: Invoice , User-defined, Utility Bill, or Receipt
Use the user-defined document type to process documents that are visually similar to invoices, such as purchase orders and sales orders, which contain key-value pairs and a table structure. In this document type, you create and configure all of the form and table fields.
Select the language.
Document Automation supports English, Dutch, French, German, Italian, Portuguese (Brazilian), and Spanish.
Note: Extraction for French, German, and Spanish language documents is currently in preview. Extraction results will improve in future releases.
If you selected Invoice: Select the provider.
If you selected the English language in step 4, Automation Anywhere (Pre-trained) is auto-selected.
Optional: You can use the Send validation feedback option to send feedback to the system to improve extraction results. For more information, see Improving extraction accuracy through validation
Optional: Select the OCR provider. By default, Document Automation processes documents in ABBYY FineReader Engine.
Users with a Cloud Control Room can select to process documents in Google Vision OCR.
Click Next.

We recommend that you open a sample document side by side with the Control Room window as you configure the form and table fields.

Note:

A form field is a type of field that occurs only one time in a document.
A table field is a type of field that reoccurs throughout a document, typically in the form of a table.

Configure the form and table fields for extraction.
Document Automation offers a standard set of form and table fields, many of which are not initially visible. You can search for a field by field name, field label, or data type.

To see the full list of fields, click Show unused fields. See the following video for a demonstration: <span>Click the Show unused fields icon to the top right of the window</span>

Click a field to open the fields editor. You can edit most attributes of a field. You cannot edit the name and default aliases. Document Automation assigns default aliases, which are hardcoded keywords, to standard fields to help with extraction. You cannot modify or delete default aliases, but you can add aliases in the Custom aliases field. See the following video for a demonstration of creating a custom alias: <span>Animation of editing a field</span>
Note: Custom fields are not available for learning instances that use the Google Document AI provider.

Mouse over the menu icon to the right of a field to access the up/down arrows. Use the arrows to rearrange the order of the fields for a more efficient manual validation. The order of the fields does not impact extraction.

To learn more about the other field attributes, review the table in the next step.

To add a field, click Add a field and complete the following fields:

Option	Description
Field name	Enter a field name that begins with an alphabetical character (A-Z or a-z). In standard fields, the field name is hardcoded and cannot be changed.
Field label	Enter a user-friendly name to help validators. For example, you can rename Organization tax number to a localized name, such as VAT number. The field label does not affect extraction.
Confidence	Set a threshold to reduce potential false positives. At the time of processing, the Document Automation engine assigns a score to each field in a document to indicate the certainty that the data was correctly extracted. If the document contains fields with a score that is lower than the confidence threshold, the document is sent to the validation queue. If you enter a high confidence threshold, more documents will be sent to the validation queue. If you enter a low confidence threshold, fewer documents will be sent to the validation queue. Supports values from zero to 100.
Data type	Choose from Address, Text, Number, and Date. If the data in the field does not match the data type, the document is sent to the validation queue. Document Automation supports date and number format variations. If you are configuring a learning instance with a user-defined document type, the form fields include the address data type, which extracts the entire structure of an address.
Format Date/Number	Set a standard appearance to convert extracted dates and numbers to a specific format. This ensures consistency and accuracy in your databases and other systems of record. For example, if you select to standardize dates to `MM/DD/YYYY` and a processed document contains a date that appears as `12 Feb 2023`, the learning instance will reformat the date to `02/12/2023`. If you select to standardize numbers to the English (United States) locale and a processed document contains a number that appears as `100,00`, the learning instance will reformat the number to `100.00`.
Required	Select one of the following: Required: Field cannot be empty. Optional: Field can be empty or not exist in the document.
Default aliases	No action is necessary for this field. Document Automation assigns default aliases, which are hardcoded keywords, to standard fields to help with extraction.
Custom aliases	Additional keywords to help Document Automation locate the field. For example, add country or region-specific names for fields such as VAT number as an alias to an Organization tax number custom field. Note: Custom aliases must be unique. They cannot duplicate the default alias of another field. Exception: Form fields can have duplicate custom aliases as the table fields and vice versa.
Validation rules	Depending on data type, create rules using patterns, formulas, lists, and statements such as starts or ends with. Formula validation \| Pattern validation

Click Create.

When a new learning instance is created, the Control Room creates a folder with the same name as the learning instance in the Automation > Document Workspace folder. The folder contains two bots (extraction and download), a process, and a form. Example screenshot of learning instance assets

Example screenshot of learning instance assets

Process: Manages the process using if/else scenarios through which Document Automation extracts data from uploaded documents, assigns documents to users for validation, and downloads the extracted data. To learn more, see About the Automation Co-Pilot process in Document Automation
Extraction bot: Extracts data from defined fields in the uploaded documents.
Download bot: Downloads the extracted data to a specific folder on the device or shared network.
Form: Defines the input parameters that are sent to the process. Input parameters include the learning instance name, uploaded file, and output file path.

Copy a learning instance in Document Automation

You can copy an existing learning instance (Public and Private) to reuse the learning instance fields, parser, document type, and language. When you copy a public learning instance, the copied learning instance remains in the private mode. After copying a learning instance, you can edit the existing parameter (except parser and heuristic feedback) as required.

To copy a learning instance,

Go to Manage > Learning Instances.
A list of all the learning instances display.
Click the three dots icon next to the learning instance that you want to copy and click the Copy option.
In the Copy Learning Instance window, enter the name for the instance.
Note: When you copy a learning instance, the default name starts with the prefix Copy of <<instance name>>. You can change the name, if required.
Select the required copy options and click Copy.
You can select the existing form and table field attributes, validation rules of existing form and table fields, and heuristic feedback.
A success message indicates that the copy process is started and a Control Room notification is triggered to indicate success or failure of the copy process.

Consider the following points when you copy a learning instance:

You cannot copy the dashboard metrics for the learning instance.
You cannot copy the connected learning instance.
You cannot copy the Doc AI learning instance credentials.
If the underlying package of the learning instance is deleted before copying an instance, an error might occur during the copy process.
Each copied learning instance is separate one.
You can delete the original or copied learning instance when required.
An audit entry is logged.

Next steps

Upload documents to the learning instance, fix validation errors, and verify the extracted data: Process documents in Document Automation

Automation 360

Create a learning instance in Document Automation