AI Doc Extraction skills

Please post if you are successful with tables or checkboxes in the reconcile as while labeling appian does not allow

Example Gender Male , Female are checkboxes and marital status etc on STANDARD Govt forms also.

how to handle , even in reconcile i saw it does not see what is in checkbox and not caputures the X mark also even i kept the rectangle around the whole 

x Male

I kept the rectangle including x , I hoped it will extract value as 'x Male' so i can do post processing , but gets only Male

I did 2 variables EMP_Gender_Male , EMP_Gender_Female and labelled/tagged the whole area of each field. 

the X mark does not get extracted

you can also achieve this having boolean function if male is selected then it will receive (true) male if  not then female (false) then you can mapped that data in Gender field using script task

you can also achieve this having boolean function if male is selected then it will receive (true) male if  not then female (false) then you can mapped that data in Gender field using script task

Hi rajthakor , I'm glad to hear you were able to extract these fields! Rayudu Addagarla for more details on how checkbox extraction works, please keep in mind the following:

1. In order to extract a checkbox in a document to a field in your AI Skill, the AI Skill field must be of type "Checkbox". You will not be able to extract checkboxes in the document to fields of type "Text" in your AI Skill.
2. You should have one AI Skill checkbox field for every checkbox you extract from your document

Also, the Reconcile task is needed to allow an end user to establish initial mappings between document checkbox fields and the fields in your AI Skill, but once you have reconciled a few docs you should no longer need it. Instead, the checkboxes will be automatically extracted and included in the output of the Extract from Document smart service. These mappings will also be deployed alongside the AI Skill object from one environment to the next. This docs page has more information about how Document Extraction works.

Please let me know if this information resolves the extraction challenges you're facing, or if you have additional questions!

 Louis Prensky  , I guess you are the Sr Product Manager for IDP especially,

I think a detailed video for checkbox and table extraction from Appian will be very helpful 

I am not sure we can edit a published model or have to redo all again
ex: If developer marked a checkbox as text vs checkbox and published, can he go back and edit the model mark those fields as checkbox and submit for publishing. i did not see that option for editing a already published mode.

Rayudu Addagarla Thank you for the feedback! You cannot edit a published model once it has been trained; however, you can easily create a new model and use the dataset of the previous model as a starting point. To do so, click "Create Model", then select the already trained model in the modal that appears. Once you do so, you can change the relevant fields from Text to type Checkbox and train and publish the new model. This WILL NOT require you to label any additional training data, and should be a quick process.

Please let me know if you have any additional questions.