Once a document is uploaded in a queue, It is entered as a new row in the Received status. It has various properties such as when was it uploaded, by which user, time since upload, status, etc. 

To view or edit the data extracted from a document, the document name in the queue can be clicked on. This directs you to the “Edit Document” page.

In the edit document pageview, you can see the document and the data extracted. The sidebar shows the fields, and by clicking on filled fields, you can view where data was extracted from.

Data extracted can also be edited. You can 

• Click on a different piece of text 
• Drag a box across multiple pieces of text to select them 
• Just directly type in the pop-up 
• Type in the sidebar directly

Once you are happy with the result, you can click “Save” to accept that text. The moment you click on save, the AI tool understands the change and gets trained on it. This way, you can train the AI tool on any new document or format type. This elimiantes the need of templates, rules or coding for data extraction.

Other actions which can be performed on a single document are shown when the user clicks on the three dots under the “Actions” header. To perform actions on more than one document, select them by clicking on the square box to the left of the document/s. The actions include:

• Edit file: Used to edit the document and its data
• Download the original file: If the user wishes to download the original file, they can use this tab
• Move to completed: Once a file is edited and is ready to extract its data, it has to be moved to the completed queue
• Move to another queue: If a document has been wrongly placed, it can be shifted to the desired queue
• Scan again: This scans the document again. If the user wants to omit preprocessing, they can click on the latter option “Scan again without preprocessing”.
• Remove: The user can delete the unwanted documents with a click of this button.

‍

Tables

When the documents are uploaded into a queue, Staple AI will capture tables in the document. 

The captured tables can be audited and data captured from the tables can be edited.

‍

Step 1: For that, either click on the “Table 1 title” in the sidebar or directly click on the blue border lines present on the table. If the blue boundaries are not present on your table, then we need to add a table. We’ll see how to do that later on this page. 

‍

How to capture tables, if the table is not auto-deducted? 

Step 1.1: The bottom of the sidebar containing the fields has a button called a smart table. Clicking on the smart tables button allows you to draw a table on the document. 

‍

Step 1.2: Immediately, you will notice a table with only one row and one column. Staple AI is a flexible tool which lets you adjust the table size in all directions, and add more rows and columns according to your requirements. After saving this document, the AI tool will learn from your action, get trained and start auto-deducting tables from the next couple of documents onwards. 

‍

Step 2: Now, when you are happy with the table drawn, you may hit the process button. This creates a popup showing the data extracted from the table and again allows the user to edit the extracted data

‍

Step 3:  Staple also provides dropdown options for Standards headers. These should be used when exporting data to an accounting system (such as Xero). You can also map a different header if the tools has incorrectly mapped column headers in the first place. Also, Staple AI can auto-calculate the total amount from the data captured. This helps you cross-verify the total calculated by Staple with the total amount in the invoice. 

‍

Step 4: Once a table is selected, it can be adjusted to fit in all the required information. You can add, resize, and delete any rows or columns as they see fit. If the table was wrongly registered, you can delete the table by clicking in the trash bin that pops up once the cursor is hovered over the setting icon.

The user can create multiple tables as desired. The output of all tables is concatenated into a single master table when data is extracted.

Learning and feedback

Staple uses Artificial Intelligence to provide personalized results to users based on their actions. It optimizes the scanning algorithm to provide users with the most relevant outcome. Staple focuses on decreasing the workload of manually scanning documents and offloading them to artificial intelligence.

Every action performed by the user acts as a feedback to the AI. It learns and improves its algorithm to fit the requirements of the user. The AI collects this feedback and processes it to provide a smoother experience for the user.

Suppose, a document uploaded by a user has a table that contains some information. The AI detected the table but the user manually made some changes, this will be registered by the AI. The next time the user uploads a similar document, the AI will use its previous knowledge to scan the document according to the previous changes made by the user.

‍