› Workflow › Case definitions › Create a case fragment

Create a case fragment

Create a case fragment to use within a case.

You can also generate a case fragment using Generative AI.

Procedure
  1. To create a case fragment, do either of the following;

    • Open a case map. On the case modeling bar, select Fragments and then select New.

      The New case fragment Design view opens in a new tab.

    • Navigate to Workflow > Business processes, and select New.

      A process map Design view opens in a new browser tab. In the properties panel, on the Type list, select Case fragment. (Default: Business process)

  2. Add activities, decisions, and end nodes to the Start node. See Design a case.

    When you configure an activity with the business rule, create new job, or custom service, and if you double click on the activity, the configured rule, process, or custom service are opened in a new tab.

  3. Configure the properties as needed. See Process properties.
  4. Save the case fragment.
  5. Release the map.
    The new case fragment is associated with the case. On the list of case definitions, expand the case to view the list of fragments associated with it. You can also view the case definitions listed on the Business processes list page.

Generate a case fragment using Generative AI

Generate a Case fragment using the Generative AI functionality.

Procedure
  1. Open a case map.
  2. On the case modeling bar, select Fragments and select New.
    The New case fragment Design view opens in a new tab.
  3. Select the Chat with Copilot icon in the bottom right corner.

    The floating Chat with Copilot window appears with the following:

    • A welcome message: "Hello <username>, what type of process would you like to create?", and a predefined request that enables automatic business process creation.

    • The Learn more link. Select the link to know more about Tungsten Automation's commitment to responsible AI and compliance.

  4. In the Chat box:
    1. Do one of the following:

      • Use the default request available in the Chat window.

      • Use your own request as a plain text, such as "Validate claim".

      • Use the attachment icon to upload images and files and then enter your request such as "Create a process using the image".

        • When you upload an image, a thumbnail is displayed, and the image content is passed as input to the provider. The image types of JPEG, PNG, JPG, GIF, and WebP are supported. If an unsupported type is used, an error appears.

        • When you upload a file, the text from the file is passed as input to the AI provider. The file types of .xml, .csv, .txt are supported.

        • When you request the AI provider to create a process based on the image, not all models support the use of images. Currently, the supported model is "gpt-4-vision-preview."

        • When you request the AI provider to create a process based on the uploaded file, the vision model is not supported.

    2. Select the Return icon or press Enter.

      You can cancel your request by selecting the stop button . This stops the AI provider from generating the response.

    The Copilot returns the following response:

    • A process (read-only) with a process name. You can save and release the process once the process is generated.

    • The read-only Properties panel.

    • A message in the Chat window: "A process has been generated. Are you ready to generate the process or would you like to make any changes or request changes via the chat such as add an additional task?"

    • The following options:

      • Accept process: The system saves the AI-generated process (read-only) with a process name, and adds it to the process designer. You can modify the process further if needed. The newly generated process opens in the TotalAgility Process designer and is locked. The process properties are enabled. You can use this process like any other process.

        An audit log entry is recorded that the process is generated using generative AI.

      • Clear & Start over: Discards the current AI-generated process definition and resets the conversation, allowing the user to provide new requirements and generate a different process from scratch.

        If you have started to create a process but not saved or released it, you can still use Generative AI but the process will be replaced. Once you save or release a process, the Chat with Copilot icon disappears.

  5. In the Chat box:
    1. Do one of the following:

      • Use the default request available in the Chat window.

      • Use your own request as a plain text, such as "Validate claim".

      • Use the attachment icon to upload images and files and then enter your request such as "Create a process using the image".

        • When you upload an image, a thumbnail is displayed, and the image content is passed as input to the provider. The image types of JPEG, PNG, JPG, GIF, and WebP are supported. If an unsupported type is used, an error appears.

        • When you upload a file, the text from the file is passed as input to the AI provider. The file types of .xml, .csv, .txt are supported.

        • When you request the AI provider to create a process based on the image, not all models support the use of images. Currently, the supported model is "gpt-4-vision-preview."

        • When you request the AI provider to create a process based on the uploaded file, the vision model is not supported.

    2. Select the Return icon or press Enter.

      You can cancel your request by selecting the stop button . This stops the AI provider from generating the response.

    The Copilot returns the following response:

    • A process (read-only) with a process name. You can save and release the process once the process is generated.

    • The read-only Properties panel.

    • A message in the Chat window: "A process has been generated. Are you ready to generate the process or would you like to make any changes or request changes via the chat such as add an additional task?"

    • The following options:

      • Accept process: The system saves the AI-generated process (read-only) with a process name, and adds it to the process designer. You can modify the process further if needed. The newly generated process opens in the TotalAgility Process designer and is locked. The process properties are enabled. You can use this process like any other process.

        An audit log entry is recorded that the process is generated using generative AI.

      • Clear & Start over: Discards the current AI-generated process definition and resets the conversation, allowing the user to provide new requirements and generate a different process from scratch.

        If you have started to create a process but not saved or released it, you can still use Generative AI but the process will be replaced. Once you save or release a process, the Chat with Copilot icon disappears.

Work with case fragments

You can open, modify, copy, delete, view versions, and unlock a case fragment. The procedure to perform these operations is like Cases. To open a case fragment for a case, expand the case. See Case definitions for more information.