Document Automation v.38 release

Review what's new, and the fixes and limitations in Document Automation for the v.38 release.

What's new

High-scale クラウド抽出サービス

The クラウド抽出サービス feature allows for large-scale document processing with enhanced speed and cross-platform flexibility, eliminating the need for additional infrastructure or Bot Runner. It offers benefits such as scalability and faster response times compared to traditional タスク Bot extraction methods.

The クラウド抽出サービス feature provides the following benefits:
  • Process requests faster in a typical front-office operations
  • Process large volumes of documents quickly
  • Automatically scale based on document volumes, ensuring consistent performance
  • Eliminate the need to invest in setting up and managing own infrastructure

クラウド抽出サービス

適応型検索クエリ

Automatically adjust to layout variability across thousands of formats and dynamically apply the correct query at runtime. Users can test and define specific search queries for documents with similar layouts within a single learning instance, enhancing data extraction accuracy for varied document formats.

The 適応型検索クエリ feature provides the following benefits:
  • Test and validate search queries at the cluster level.
  • Define different search queries for the same field across multiple clusters within a single learning instance.
  • Promote validated cluster-specific queries to production for robust and consistent extraction performance.
  • Minimize manual query tuning and improve extraction accuracy across varied document formats from within a single, unified learning instance.

適応型検索クエリ

Azure AI Document Intelligence version 4.0 support in 標準フォーム (Service Cloud Case ID: 02166473)

You can now create custom models in Document Automation using Azure AI Document Intelligence version 4.0. The new capabilities include improved custom models, improved accuracy in data extraction, and signature detection.

The following capabilities are included in the Control Room:

  • The Models page now displays the version number used to create each Azure AI Document Intelligence model.
  • Support for signature detection is introduced, enabling identification of handwritten and digital signatures in documents. This improves automation accuracy in workflows that require signature validation.

標準フォームの学習インスタンスを作成する | Document Automationの検証ルール

Support for new languages in Automation Anywhere user-defined provider

You can now process documents in the following languages for the User-defined document types when using the Automation Anywhere (user-defined) provider and ABBYY OCR or Google Vision OCR provider:

  • Czech
  • Danish
  • Finnish
  • Norwegian
  • Slovak
  • Swedish

Document Automationでサポートされている言語

Fixes

When you open the process of a learning instance from the Learning Instances page, the process will be opened in the 次世代 Process Composer.

Previously, the process opened in the legacy Process Composer.
The test mode option will be enabled even when a large number of learning instance versions exist. You will now be able to view the Process and View status test mode options.

Previously, the Process and View status options were not displayed as there was some delay in the loading of versions.

Service Cloud Case ID: 02273741
After moving learning instances with optional form fields that include a formula from IQ Bot to Document Automation, optional or disabled form fields now correctly skip formula rule evaluation during document processing.

Previously, documents using such form fields were sent to Automation Co-Pilot validator in certain scenarios.

Service Cloud Case ID: 02279788, 02278791
When you extract table data from multi-page documents containing blank pages in between, the extraction now works as expected.

Previously, the extraction stopped when blank pages were detected, and table data from subsequent pages was not captured.

Service Cloud Case ID: 02274630
When you are using a proxy server and using 生成 AI provider for data extraction, the extraction now completes successfully without any errors.

Previously, document extraction failed in such a scenario.

Service Cloud Case ID: 02218893, 02214721, 02200518, 02232155
When you import a learning instance into a Control Room where the same learning instance already exists, all aliases from the exported learning instance are now correctly retained.

Previously, the imported learning instance missed some aliases in such a scenario.

Service Cloud Case ID: 02227635
When you remove data from table cells in the Automation Co-Pilot validator and use the Auto fill option subsequently, the auto-fill process will automatically update the manually emptied cells.

Previously, the manually emptied table cells were left empty.
When you have documents pending validation and update to the v.38 release, you will no longer see an error after validating the first document in the queue and clicking the Submit button.

Previously, an error was displayed in such a scenario.
After updating to the v.38 release, documents that were previously processed without the need for validation will continue to be processed without validation.

Previously, documents were sent for validation in certain scenarios.
When a user is assigned the Conversational Automation Co-Pilot User license, the user will now be able to process documents in public learning instances in Document Automation.
The Auto fill option in the Automation Co-Pilot validator now works as expected.

Previously, this option did not work as expected in certain scenarios.
You can now upload and process documents without errors for learning instances.

Previously, an error was displayed in certain scenarios.

Service Cloud Case ID: 02135184, 02135889, 02154359
You can now process the same documents subsequently without errors.

Previously, an error was displayed in the validator in certain scenarios.

Service Cloud Case ID: 02187660
You can now create 標準フォーム models without seeing any performance issues.

Previously, users were experiencing performance issues in certain scenarios.

Service Cloud Case ID: 02219164
After training documents using 標準フォーム Neural model, the Run analysis option will extract appropriate information from test documents.

Previously, data was not extracted for certain fields.

Service Cloud Case ID: 02185363, 02185363

The validation feedback field in the version history in test mode will only display fields that were validated by users.

Previously, fields extracted using 生成 AI were displayed in the validation feedback field.
When you have the Test mode enabled and validation feedback disabled for a learning instance, the cluster ID is now displayed correctly for a document processed in test mode.

Previously, incorrect cluster ID was displayed for the processed document in such a scenario.

Limitations

When you edit and check in a process linked to a public learning instance that is using クラウド抽出サービス, the extraction Bot might not be visible in dependencies, and document processing might fail.
Workaround: Use one of the following options:
  • Update to the v.39 release
  • Use the default process or copy the learning instance to create a new process and then publish the learning instance before processing documents
When you create, edit, or copy a learning instance, it takes a bit longer than usual to perform the operation.
When a user is assigned the “Conversational Automation Co-Pilot User” license instead of the “Automation Co-Pilot Business User” license, the user will not be able to process documents in public learning instances in Document Automation.

Workaround: Assign the “Automation Co-Pilot Business User” license.
For a learning instance, if you keep the default table prompt used in the learning instance and add a custom prompt for any table field in the 適応型検索クエリ, the table prompt in the 適応型検索クエリ is reset to empty.

Workaround: You must enter the same default table prompt (used in the learning instance) with some minor changes in the table prompt section of the 適応型検索クエリ.
When you clear table prompts in 適応型検索クエリ, the prompts might not be cleared, and the UI will display the previously used table prompt (custom or default prompt).
The Cloud extraction option works only when the File storage location is set to Document Workspace Storage in the Start event of the process. By default, the File storage location is set to Document Workspace Storage when you create a learning instance.
When you import a learning instance that was exported in v.33 or earlier releases, you might not see the option to select the Cloud extraction option on the Import Learning Instances page.

Workaround: Perform one of the following workarounds:

When you add an invalid regex pattern in the field and document rules for a learning instance and process documents, data extraction will fail.

Workaround: Validate the regex pattern before adding it to the field and document rules and then process the documents.

When using the 適応型検索クエリ feature in クラウド-Sandbox, if you add a custom query for a field, process a document using the custom query, and then open the search queries option to view the custom query, an error will be displayed.
When you have documents pending validation and update to the v.38 release, an error is displayed after validating the first document in the queue and clicking the Submit button.

Workaround: Perform one of the following workarounds:
When you update to v.38 release, if you have an existing 標準フォーム project, create a subsequent project, add the same set of documents to train as the previous project, create a model, and train the model, an error is displayed. The error is displayed only when you create a subsequent project and train with the same set of documents as the previous project.

Workaround: Reload the Projects page before creating a new project. If you create a project without refreshing the page and an error occurs, create a new project after reloading the Projects page and restart the training. Existing learning instances, projects, and models are not affected.
Limitations from previous releases
When you try to check in a process without having the required permissions, the audit log entry for this event will include Source as Control Room instead of Document Workspace and Item name as Unknown instead of <learning-instance-name>.
You will not be able to use a custom table name in Japanese, Korean, or Chinese when creating a custom table.
When you restore a learning instance to any previous version from version history and reprocess a document, the document upload count is increased.
When you process documents with file names greater than or equal to 75 characters in test mode, you might not see improvements in the reprocessing time for such documents.
If you have disabled the OCR provider in the administrator settings and if you are using a language other than English for your Control Room, you will see an error to enable the OCR provider settings in English in the following scenarios:
  • When you create a learning instance that is using an OCR provider
  • When you change the OCR provider for an existing learning instance
When you use the ドキュメント分類子 アクション (Classify, Classify documents, and Train Classifier) and Extract Data アクション in the ドキュメント抽出 パッケージ together in a Bot, the Bot will fail to execute.

Workaround: Ensure that you create separate Bot when using any of the アクション from the ドキュメント分類子 パッケージ and the Extract Data アクション of the ドキュメント抽出 パッケージ. If you need to execute these Bot in a sequence, include these Bot in an Automation Co-Pilot process.
If you copy a learning instance that is using a third-party parser configured in Document Automation and process documents using the copied learning instance, data extraction will fail.
When a user processes a document on a custom process learning instance, the validate document count is not updated post extraction. Also, if the user submits the document, the validate document count is updated with a negative value.
A user with the Automation Co-Pilot administrator permissions is unable to view the Document Automation tasks that are assigned or requested and are in pending or complete status.
When you are using the IQ Bot Pre-processor パッケージ アクション and if the Output folder path contains Japanese characters, you will see an error when processing documents.

Workaround: Create an output folder in a folder path that does not contain any Japanese characters and provide the path in the output folder path field.

Completed deprecations

Support for Google Custom Document Extractor (CDE) is deprecated

With this deprecation, users can continue to use their existing learning instances that are using Google CDE. However, users will not be able to create new learning instances with Google Document AI (User-defined) as the provider. We recommend that such customers use 標準フォーム with Azure AI Document Intelligence. See Document Automationを使用して標準フォームにカスタム モデルを作成する.

非推奨化が完了した機能

Updates to the interface

Document Automation
You can now select the クラウド抽出サービス feature when creating or copying a learning instances.

Interface showing options to create or copy a learning instance with クラウド抽出サービス

High-scale Cloud Extraction Service

You can now use the 適応型検索クエリ feature to test and define specific search queries for documents with similar layouts within a single learning instance, enhancing data extraction accuracy for varied document formats:

Adaptive Search Query option

Adaptive Search Queries