top of page
  • Writer's pictureIlia Zelenkin

How to Recognize Document Type Automatically - Logistics Use Case

Updated: Jul 8, 2023

The world of supply chain and logistics is a complex and interconnected web of activities that involve multiple parties, including manufacturers, suppliers, distributors, retailers, and end customers. In order for this system to function efficiently, it is critical that each participant has access to accurate and timely information.

One of the key components of this information is documentation. Documents such as purchase orders, invoices, bills of lading, and customs declarations play a critical role in the supply chain by providing a record of the transactions that occur between the different parties involved.


It is pretty clear that matter how many innovations will appear, the logistics industry will still rely heavily on documentation. And if you are in this space, you already know the impact of automation to document processing.

Why it is important to automate document processing

Besides the obvious things like increased efficiency, improved visibility, and cost savings you will be able to do 10 times more than you do today. I think there is a secret in almost all industries where there are still lots of documents available - the company growth is stagnating because the extra load requires hiring more people.


And in small and medium businesses hiring a new employee is a very risky endeavor. That's why many prefer not to grow to avoid hiring.

The first step in automating document processing

If you take a closer look at your processes, you will notice that the first step in it is receiving and opening a PDF file to understand what it is. We do it mechanically - we receive an email or form submission, with a bunch of files inside, and we open each file and check which one is what. This file is a Bill of Lading, and that one is a dangerous goods certificate.


You are basically doing document recognition. And that's what we are going to automate - we will create an automated document recognition scenario to stop wasting time on the same things over and over again.




Setting up the Bitskout plugin

1. The first step is to configure the Bitskout plugin. Go to the Templates library and search for a template called Detect Document Type.



2. Press Use Template and then choose the tool where you would use the plugin. For this example, we will use make.com. Click on the tool icon and press Next.




Your plugin is now ready. The next step is to build a make.com scenario. For this example, we will use a simple setup: an Airtable form where a customer can submit a bunch of documents, and we will recognize each document and set it's type in Airtable. Your case might be different - customers can send emails or they can create a shipment first and you need a shipment ID. In any case, Bitskout needs a URL with a file to recognize the document type. Once you provide the URL, it will automatically recognize the file and return the label for the document type.



Setting up the make scenario

In our example, we will use a webhook. This webhook will react to the new form submission.




The difference between the webhook and other methods is that webhook only reacts when the event happens, but Watch Records periodically checks for new records.


The next step is to get the record details. Basically, when a user submits the form, we need to get the list of files that were uploaded.


As we have multiple files, we need to use an iterator. The input array for the iterator would be the list of files from the File column that I've added to Airtable.






And then we need the Bitskout module to read the file and tell us what type of document we are looking at:




In this case, we are providing an Airtable URL, but if you use another setup, you need to provide a URL that is accessible from outside. Otherwise, Bitskout will not be able to read it.


And the last step is to create a record per each file with its type.


And that's it. Now every time a new form is submitted, Bitskout will read the file and return the document type. Thus you can build the next step in automation - extract data from a document into your fields.


For instance, you detect a Bill Of Lading. You can move the document to the "Bills of Lading" table and extract all the data from there.



This way you create a fully automated document extraction workflow.


Video How To




Conclusion

In conclusion, by creating such automation you can lay down a foundation for your productivity boost. By automatically detecting the document type you can then create separate folders per document, add new automations on top of that, and stop wasting your employee's time on such a mundane task.


And you should be looking at employee motivation - almost 57% of employees want to use A.I. for their work. Basically, they already know that such tasks should be automated. But before it was too expensive and that’s what we’ve changed with Bitskout. Now it is available to everyone.


So go ahead, try it out, send us questions, and get ready for the next step where we will build an automated data extraction based on document type.


3 views0 comments
bottom of page