OCR bot development for reading downloaded docs

By lmartinez, 13 March, 2023

Challenge

 

The client is an intermediate company, they are a cargo and transportation company. In their day-to-day operations, they handle the load of various clients from departure to destination. Clients always send you a document in a standard format, which is what they present to the ports or customs where all the information about the cargo they carry is located.

 

That document may contain item information. The documents for each load can have one product or one hundred.

 

FCI has its own internal platform, they manually insert these documents, which became quite a tedious job, so they decided to automate that process and save several hours manually typing the documents on a daily basis.

 

Our solution

 

Rootstack experts immediately got to work. For FCI, a bot with UiPath was created, with the aim of reading the documents generated by clients, which they call settlements. The documents for reading can already come in digital formats for download, or scanned from their physical format.

 

fci

 

FCI stopped processing manual documents on its platform thanks to the creation of the OCR Bot by the Rootstack team. To create it, they used an OCR from Amazon called Amazon Textract, which is a plugin that is available for UiPath. The OCR configuration was done and all the documents began to be read, thus obtaining the ability to read scanned or downloaded documents.

 

A normal OCR could only read the downloaded documents, with the OCR created by Rootstack, FCI got a bot capable of reading scanned documents that might be unreadable.

 

 

Amazon Textract was chosen because it is the most complete OCR and the one that offered the greatest support to the problem that FCI presented with its scanned documents and those with difficult visibility. For example, in a document with 10 products, the OCR reads them line by line and they are inserted into the FCI software.

 

The Amazon Textract was used with the community version of UiPath, this is free.

 

In addition to this, a function was developed to be aware of the possible failures of the Bot, which, when unable to process a document for some reason, immediately sends an email to the person or persons assigned to this process and so they are aware. 

 

 

Technologies

 

UiPath is one of the most widely used tools for building RPA, or Robotic Process Automation, technologies and bots that help businesses speed up tedious processes and maximize productivity.

 

For its part, Amazon Textract "is a machine learning (ML) service that automatically extracts text, handwriting, and data from scanned documents. It goes beyond simple optical character recognition (OCR) to identify, understand, and extract data from forms and tables", according to the definition on its official website.

 

Content
Testimony
Author Role
CFO
Author
Construction Company
Quote

“Mastering business data is key to define strategies that have an impact. Working with Rootstack, we were able to take control of our data and analyze it purposefully to help us advance in the right direction.”

Author Role
CTO
Author
Financial Institution
Quote

“Working with Rootstack has been a wonderful experience. They have truly helped us improve our site design and performance, allowing our business to grow and stand out in the online world”

Title
Testimony
Margin
With Margin vertical
With title
Off
Title
Check out some of our success stories!
Text Alignment
Left
Text Color
Dark Indigo
Items Alignment
Center
View
Success Stories
Item style
Success Story Card
Show Filters
No
Margin
With Margin vertical
With Background
Off
Title
Our main services
Text Alignment
Center
Text Color
Dark Indigo
Items Alignment
Center
View
Services
Item style
Service Card
Show Filters
No
Margin
With Margin vertical
With Background
Off
PDF Document
Document
Weight
2
Image
Hero
Title
OCR boot development for reading downloaded and scanned documents
Image
Image
Description

A bot with UiPath was created, with the aim of reading the documents generated by clients, which they call settlements.

Text Color
White
Text Alignment
Left
Size
Medium
Overlay effect
Hide overlay effect
Margin
With Large margin bottom