Extract data from documents
Create a bot to extract fields from the documents using the learning instance you created.
Create a learning instance before you create a bot.
Create a bot, and then use the action to process your documents and extract data . Users can upload multi-page PDF documents for data extraction.
Note: IQ Bot uses the ABBYY FineReader Engine for OCR.
On the left pane, click Automation.
A list of available bots and forms is displayed.
- Click Create a bot.
- In the Create Task Bot window, enter a name and click Create & edit.
- The bot designer window appears.
- In the IQ Bot to display the available actions. column, search for
- Drag the action into your workflow.
- In the Input File Path field, use Browse to navigate to the file you want to process.
- From the Learning Instance that will process the document drop-down list, select your learning instance.
Optional: In the Save response code to variable
(optional) field, select prompt-assignment –
String from the drop-down list.
The action response code is saved in the provided variable.
Optional: Drag the Message box
action to display results in the message box. Define the message
you want to see displayed in the message box.
- Click the Insert a value (F2) icon in the Enter the message to display field.
- In the Insert a value dialog box, select prompt-assignment - String from the drop-down list and click Yes, insert.
- Click Save and then click Run to run the bot.
After document processing is complete, IQ Bot sends successfully extracted data to the output folder, specified during learning instance creation, and the Success subfolder for successfully processed documents; while failed documents are sent to a different folder for validation.
Note: The output files are in .json format.
The extracted files are sent to the output folder you previously configured as a prerequisite step:
- Go to the Output\Success folder on your machine to view the successfully extracted data.
- Go to the Output\Data folder and the Output\Meta folder on your machine to view information about the failed documents requiring validation.
- Go to the Output\Unprocessed folder on your machine to view the documents not extracted by the Document Extraction action.
Based on the extraction results thedashboard displays the updated STP (straight-through processing), total number of uploaded documents, and the number of documents pending review.