Processing invoices in a construction company

About the client

The company is a big construction enterprise, located in Canada.

Motivation to implement RPA solution:

They have a problem handling over 20 thousand invoices from contractors per month. The person in charge of this is not able to keep up with the amount of work.

Every invoice comes via email and is in PDF format. Each invoice is different, with no universal template. This requires the employee to read each invoice and put the necessary information in an ERP system in the correct order according to the date of payment. After the invoice has been processed the employee makes the transfers.

In the automation phase we have:

• developed a methodology for automating existing systems conducted tests of solutions
• measured the parameters of the solution according to the specific needs of the customer
• implemented improvements
• launched production of the robot

Tasks performed by a robot:

We set up our software robots to be triggered by an incoming email with an invoice. Then the robot is checking the email sender address in the database to see if they are the company’s clients.

The robot then downloads the invoice in a PDF file. Each file had a different structure so in order to extract the data we created an Add-on for reading unstructured PDF files.

Then the robot is putting the extracted information into the ERP system accordingly to the date of the invoice and generates a report for the employee with all the updates.

Results:

The result is that the robot is taking care of 80% of the business process. The robot has no limit to the number of invoices it can process, it can handle all the invoices and never gets tired or makes mistakes.

The employee’s job is to work on reports generated by the robot and make transfers.

• A larger number of invoices can be processed
• No mistakes are made, and the process is standardized
• Freed up employees’ time to focus on higher-level activities
• No need to hire additional employees

• Savings on hiring another employee are approximately 40 thousand USD per annum
• Automation took a week to develop and another week for testing, adjustments, and implementation
• Saved approximately 120 hours of employees time per month
• The expected ROI per annum is at least 75%