Important! Read this.

We have launched a new documentation website. If you're using Enterpise Edition 10 (or higher) or Express/Business Edition 2.4.1 (or higher), visit this page.

Skip to end of metadata
Go to start of metadata

We have moved to a new documentation platform. This section is no longer supported. For the up-to-date information, see Account payable.

Account Payable is a sample business process that harvests a set of invoices from the Invoice Plane application, saves the invoices to an S3 emulated bucket, converts the invoices to HTML format, and submits them as HTML files to WorkSpace to be processed as a manual task by a Worker.

Specify Business Process Input Data

To view and change if needed the input data, do as follows.

  1. Go to the Data tab containing a link to the application to get the invoices from.
  2. Click Next to proceed to business process Design to check the workflow and business process steps.

Run Business Process

To start the business process, go to the Run tab.

  1. Click Run This Process to initiate business process execution.
  2. If the business process succeeds to launch, the following message window emerges.
  3. Click View Results to go to the Results tab and check how the business process is executed according to its workflow, or View Business Processes list to return to the overview of business processes.


The business process contains a manual task that must be completed in WorkSpace. The business process cannot be completed as long as the manual task is not processed.

Business Process Workflow

Go to the Design tab to see the entire business process workflow and check each particular step (Task or Rule).


Tasks can incorporate variables, which are used to transfer the data from one task to another, or identify an external source (for example, predefined S3 bucket) for the task input data. Such variables are declared externally, for example, in configuration files, and not included to the input data for the business process.

In such case, you will see the warning message as below. The warning message is redundant, so you can skip it.

The workflow of the current business process consists of the following steps.

  1. Start (initial step)
  2. Get invoice links >>
  3. Split the acquired data to separate records >>
  4. Save the invoices to the S3 bucket >>
  5. Convert the invoice images to HTML using the OCR plugin >>
  6. Send the invoices in HTML format to WorkSpace to be processed manually >>

  7. End (final step)

Double click a task to see its structure and properties.

Let's look at the business process workflow in detail.

Step 1 – Workflow Start

Workflow is initiated and started.

Step 2 is a bot task.

Input data:

  • 1 record – link to the application with the invoices, defined in Data (in column applicationlink)
Input Data - link to application with invoices
<required name="applicationlink"/>
	<var-def name="startPage">

Bot Task:

  1. Launch the web application (Invoice Plane) via the link defined in the Input data and log in.
  2. Harvest the links to invoices in the image format.
  3. Create the links to invoices (Output data).

    Create output file with links to invoices
        <export include-original-data="false">
     		<multi-column list="${invoiceLinks}" split-results="true">
    			<put-to-column name="invoicelink"/>

Output data: 

  • 5 links to invoices created by the Export plugin

Step 3 – Split Data

Step 3 is a rule to split the data acquired in Step 2 to process each document separately.

Input data: 

  • 5 links to invoices created by the Export plugin at the previous step

Output data: 

  • 5 records – each record contains a link to the invoice created by the Export plugin at the previous step

Step 4 – Save Invoices to S3

Step 4 is a bot task used to save the invoices (image files) from the links collected at Step 2 to the S3 file storage bucket. The links to the invoices in the local S3 File Storage are provided as the input data to the next step.

Input data: 

  • 5 records – links to invoices created by the Export plugin at Step 2

Bot Task: 

  1. Get an invoice from the link in Input data.
  2. Put the invoice in the local S3 File Storage.
  3. Create and put the link to the invoice in the S3 bucket to Output data
  4. Repeat the procedure for all the links in Input data.

Output data: 

  • 5 records – links to invoices in the local S3 File Storage

Step 5 – Transform invoice image to HTML page (OCR)

Step 5 is a bot task to recognize the image files (invoices) located in the local S3 File Storage (via s3_invoice_link) and convert them into HTML format via the OCR plugin.

When the invoices are recognized, the Workers will not enter the information from image by hand, but select the information with the mouse and mark it with the respective tags. In this way, the tasks can be accomplished much faster and chances of human errors are significantly minimized.

Input data: 

  • 5 records – links to invoices in the local S3 File Storage created by the Export plugin at Step 4

Bot Task: 

  1. Get the invoice from the link in Input data.
  2. Process the invoice with OCR and convert it to HTML.
  3. Put the invoice in HTML format in Output data.
  4. Repeat the procedure for all the links in Input data.

Output data: 

  • 5 records – recognized invoices in the HTML format


Upon step completion, a manual task is created in WorkSpace to be fulfilled by a Worker as described in Step 6.

Step 6 – Extract invoice fields

Step 6 is a manual task for a Worker to extract the information from the invoices. 

The manual task is carried out by a Worker in WorkSpace (Open Workspace), where the Worker who accepted the task should properly mark the respective text in the invoice (answer the questions) to extract these data. The input data are HTML files created by the OCR plugin. Each file is processed separately.

You can find more information about Manual Tasks in Information Extraction Online User Guide.

  • Log in to WorkSpace using your credentials.

  • Go to the Tasks tab and click Miscellaneous Tasks in the Select Categories section.
  • The tasks list appears. Click View to open and accept tasks for processing.
  • Click Accept to start processing the task. Select the Automatically accept the next task option to proceed to the next task automatically when the current one is complete.
  • The input data are HTML files created by the OCR plugin. Each file is processed separately. Use your mouse to select a field with the needed data and click on the respective answer, or apply the shortcut shown on the answer label.

    The following text fields should be extracted:

    • Invoice date – date of invoice issue
    • Invoice number – invoice registration number
    • Bill to – company name the invoice is issued to
    • Billing address – company address
    • Invoice amount – total sum of the invoice
    • Invoice currency – the currency 

    Repeat for all the fields and click Submit when ready.

  • On submitting, the next task opens. The tasks start automatically, if you have selected Automatically accept the next task, otherwise click Accept to process the next task. 
  • If you want to cancel processing the task, click Skip or Return, if the task has already been accepted.
  • Complete all the tasks and return to Control Tower to see the business process progress.

As all the invoices have been processed, the manual task is complete. Then the workflow proceeds to the final step.


To finish the business process, the manual task must be completed in WorkSpace.

Step 7 – Workflow End

When all the business process tasks are complete and submitted, the business process finishes.

Check Business Process Execution Results

During business process execution, you can check the execution results for the entire business process and for each particular task.

  • Business process execution results on the workflow: 
    • An icon in the task upper right corner indicates if the task is automatic (bot) or manual. Below each step, you can see the step progress in percents together with the ratio between incoming and outgoing records.
    • Steps 1, 2 and 3 complete. 1 link to the application with invoices processed, 5 links to invoices collected and provided to save invoices in the S3 file storage.
    • Steps 4 complete. 5 links to invoices processed, 5 invoices created in the S3 file storage and ready for OCR processing.
  • The Results page looks as follows.

    Click the task progress bar to switch to the Records view where all records, both in Input and Output data, are displayed in a table.

    Use the Choose Task dropdown menu to switch between tasks or view the final results.
  • When the bot tasks are complete, the status is as follows.

    • Use the Choose Task dropdown menu to switch to the manual task.

      Open WorkSpace via the link.
  • During business process execution or upon its completion, you can check the execution results for the entire business process and each particular task. The overall progress is displayed in Submission Statistics in the Results tab.
  • Click the progress bar on a task to view the records created by each step. To see the entire set of records select Final Results in the Choose Task dropdown menu.

  1. invoicelink – records with links to invoices collected at Step 2 and split at Step 3
  2. s3_invoice_link – links to invoices in the S3 file storage from Step 4
  3. records created by OCR at Step 4:

    1. content_url – links to invoices to be recognized by OCR

    2. content – content of invoices recognized by OCR

    3. html_content  – invoices converted to HTML

  4. html_content_tagged – invoices in HTML with tagged fields for information extraction created in the manual task at Step 6

  5. extracted information after the manual task completed at Step 6:

    1. invoice_date – invoice date

    2. invoice_number – invoice number

    3. bill_to – company name

    4. billing_address – company address

    5. invoice_amount – invoice total sum

    6. invoice_currency – invoice currency


If a step fails, or there are any other issues that prevent the business process from completion, refer to Crash Course in Business Processes.

  • No labels