on 04-10-23 10:28 AM - edited on 12-04-24 01:18 PM by Michael_S
Each day, approximately 700 invoices are delivered to a designated email inbox by a wide range of vendors, each using their own unique formats. These invoices fall into five distinct document types, each of which undergoes its own dedicated automation workflow process before being ultimately uploaded to the target invoicing system.
Step 1: Discovery
We conducted a series of sessions with the team to gain insight into the existing manual invoice processing workflow as well as examine the OCR tools in use. Additionally, we observed that the team was manually inputting data based on their knowledge and experience. Decision-making criteria included factors such as the total amount, line-item description, invoice country, and various document types.
Step 2: Design
The comprehensive solution leveraged Blue Prism and ABBYY to handle the entire invoice processing cycle, encompassing loading the invoices, conducting processing tasks, validation through Intelligent Document Processing (IDP) as well as in Blue Prism, and finally, registering the data into the Accounts Payable (AP) system. Specific workflows and rules were designed for each document type to ensure accurate data extraction and processing.
Step 3: Build
Blue Prism serves as the centralized automation platform responsible for several key tasks within the workflow:
This comprehensive approach ensures efficient and accurate invoice processing while minimizing manual intervention and improving communication with vendors.
Step 4: QA
The development of this solution was a multi-phase journey that required several iterations and adjustments to meet the expected standards. Given its complexity and the numerous components involved, the solution underwent multiple changes, fixes, and enhancements to ensure it met its intended objectives. This iterative approach allowed for continuous improvement and refinement, ultimately resulting in a robust and effective system.
Step 5: Run
Utilizing the agile methodology was a strategic choice to manage and develop this project effectively, given its inherent complexity. This approach enabled the team to continuously address issues, make necessary fixes, and deliver incremental improvements. Through multiple refinement cycles and enhancements, we were able to achieve the highest possible document automation quality while ensuring the long-term sustainability and maintainability of the system. Agile practices were instrumental in adapting to evolving requirements and delivering a robust solution.
Helping business users understand how to collaborate effectively with automation robots was a critical aspect of this project. Adapting to changes in the existing manual processes presented its own set of challenges. Managing invoice volume and package versions emerged as crucial components of the project's success.
The project's ultimate outcome was highly favorable. It resulted in a well-optimized invoice automation process, leading to significant savings in terms of human resources and the elimination of extensive manual efforts. This streamlined approach substantially reduced errors and allowed the team to concentrate their efforts on handling problematic invoices that couldn't be easily processed by automation, while straightforward invoices continued to be processed seamlessly in an automated fashion. Overall, the project achieved its objectives efficiently and delivered valuable benefits to the organization.
The success of a document automation project relies on several key factors:
------------------------------
If I was of assistance, please vote for it to be the "Best Answer".
Thanks & Regards,
Tejaskumar Darji - https://www.linkedin.com/in/tejaskumardarji/
Technical Lead
------------------------------
Love the use of IDP here @Tejaskumar_Darji !
Did you find that most of the invoices this process handled were in a similar format / layout?
Yes, multiple vendors with varying layouts.
Thanks for sharing your experience automating the AP Invoice processing use case.
I'm particularly interested in understanding the decision-making process that led you to Abbyy. Were there specific features or aspects of Abbyy that stood out for the AP use case? Which specific Abbyy product/service you are using for invoice data extraction? How many documents per format you had to use to train the model, or did it work out of the box?
Additionally, could you share some details about the other IDP tools and platforms you evaluated before finalizing Abbyy? Understanding the comparative analysis can provide valuable insights for others exploring similar solutions.
Implementing new technologies often comes with its set of challenges. Were there any specific technical hurdles you encountered during the implementation of Abbyy for your AP use case? How did you address these challenges, and do you have any tips or best practices to share with the community?
Thank you once again for sharing your experiences. Looking forward to any additional details you can provide. I believe your insights will benefit many of us who are considering or currently implementing similar solutions.
Interesting! I had a similar project for AP team here in our org., but we used Microsoft's Document Intelligence for IDP. Collaborating with business users and managing manual process changes were key challenges. Your insights on IDP platform selection and stakeholder management are spot on!