By automating document data extraction you can save precious time on manual data entry.

The documents come in different forms and there are several ways to extract information from the document. You can run several workflows on the same document depending on what you need to extract.

There are several types of documents that we will cover here:

Standard Documents

Bitskout platform constantly adds support of standard documents. As an example, let's set up the invoice extraction model.

  1. Go to A.I. models and press Add to add a new model.

  2. Specify the model name and click on the Data Extraction option.

  3. Scroll down and choose the Invoices option in the drop-down.

  4. The description screen will tell you what is possible to do:

  5. Scroll down to Standard Detected Fields and choose the fields you need to extract (or select all).

    Please note that the Line Items in the invoice are extracted as a CSV table. It means that they need further processing.

  6. Select the checkbox to automatically create the workflow and press Apply.

Now the model is created. You can now create an Output to write data to your tool's fields.

Structured Documents

The structured documents look like a form. Here are a couple of examples:

Shippers Declaration

IRS Form

Proprietary Maintenance Report

As you can see all those documents have a certain structure. It is called a form - there is a field name(Name, Email, etc) and the field value ("Bruce", "", etc).

To create a model we will use the Extract Forms from Documents feature:

  1. Go to A.I. models and press Add to add a new model.

  2. Specify the model name and click on the Data Extraction option.

  3. Scroll down to the list of options and choose Extract Forms from Documents.

  4. Once you've selected the Forms option, the system will display a description:

  5. Scroll down and add an example of a document:

  6. Once the document is loaded, you will see the button "Refresh list of fields" activated.

  7. Once pressed, you will see the list of detected fields and will be able to select the ones you want to extract.

    Another example with shipper's declaration:

    And with another report:

    Selecting those fields will allow you to extract information automatically from all documents that have the same structure.

  8. Once you've selected the required fields, scroll down and select the box "Automatically create workflow" and press Apply.

    Now your model is created and you can proceed to create an output to write the data to your tool.

Unstructured Documents

Unstructured documents have a free-text form. For instance, there are contracts where the data is written as a piece of text. Extracting information from such sources is also important in boosting productivity.

Extracting data from a contract to fields:

Did this answer your question?