Questions About Document Understanding in Action Center

Help Document Understanding
1

Hi all,

I have some questions regarding the Document Understanding feature in UiPath’s Action Center, specifically related to tasks such as classification, extraction, and validation. I’m hoping to get some clarification on the following points:

  1. Handling Undefined Form Types in Classification:
    During the classification stage in the Action Center, if a form type doesn’t match any of the predefined categories in my list, is there a way to classify it as “undefined” or handle it later for further review or processing?
  2. Reordering Extracted Fields in the Action Center:
    When working in the extraction stage of the Action Center, I’ve noticed that the extracted fields are displayed in a fixed order. Is there a way to rearrange these fields to make them more logical or easier to work with?
  3. Viewing All Assigned Tasks:
    Is there a way to view all tasks in the Action Center and see who they are assigned to? This would help with monitoring the task distribution.
  4. Identifying Task Types (Validation vs. Extraction):
    Is there a way to quickly distinguish between validation tasks and extraction tasks from the task list without having to open each task individually?
  5. Splitting PDF Forms in the Validation Stage:
    In the validation stage, I need to split a PDF form that contains multiple types of forms into individual kits. Each kit will contain one form type, but there is an image on the last page that should be attached to every kit. Is there a way to do this in the Action Center, or would it require additional steps outside the Action Center?

Thanks in advance for your help!

2

@yuvalB

there is no undefined option

can be done by tweaking the taxonomy json mnaually

you have a pending tab from where you can verify these

you can add tags to differentiate

this is not possible in action center

cheers

1 Like
3

Hey Anil_G,

Thanks so much for your helpful responses! I really appreciate the insights you’ve shared.

I’m working on a solution to handle Action Center efficiently for large amounts of tasks. I’d like to streamline as much as possible directly from the JSON (through API requests), instead of doing things manually in Action Center.

I’ve got a few follow-up questions I’m hoping you could might help with:

  1. Displaying Assigned User in Pending Tasks:
    In the Pending tab, I can see tasks, but I need to click each one to see who’s assigned. Is there a way to show the assigned user directly in the list?
  2. Automating Task Tagging via API:
    You mentioned adding tags to tasks. How can I do that programmatically using the Swagger API? I’d like to automatically add tags like “Extractor” or “Classifier” based on the task type.
  3. Efficiently Assigning Multiple Tasks to Users:
    I’ve got 300 tasks that I need to assign to 3 employees (100 each). What’s the best way to handle this in bulk?
  4. Editing Document Type Names in Document Understanding:
    Is there a way to edit the names of the document types I’ve defined in Document Understanding? I haven’t found an option for this and would like to modify them if possible.

Thanks again for all your help! Appreciate it.

1 Like
4

@yuvalB

I dont think you can as of now

ou dont need a separate api,options in create form task directly …you can use all or any one also

The image is a screenshot of documentation providing detailed information and notes on various fields such as Labels, Orchestrator Folder Path, Reference, and TaskCatalog, along with important permissions and error information. (Captioned by AI)
The image is a screenshot of documentation providing detailed information and notes on various fields such as Labels, Orchestrator Folder Path, Reference, and TaskCatalog, along with important permissions and error information. (Captioned by AI)1150×607 66.9 KB

you have assign task activity if you want to do it automatically via process

in taxonomy you can give the document types…there you should be able to…if you are askign on AI center you might need to create new again

cheers

1 Like