Digital Imaging + Machine Learning + AI + RPA Bots combined to reduce supplier invoices processing
The client is one of the largest independent energy company in Argentina, with participation in the electricity and oil and gas value chain. The company is headquartered in Buenos Aires and has interests in power generation, electricity distribution, gas upstream and downstream.
The client is going through a process of expansion of its businesses and operations, both through the purchase of new organizations and the improvement of its processes and system components to accompany and support the management and administration of its work model.
For this, it is in constant research and development of innovative solutions in areas like vendor invoices processing to improve.
Accounts Payable area has 9 people dedicated to reading and interpreting 6,000 emails per month with attached PDF documentation, which must be uploaded to the SAP system as part of the “Preloading invoices” process.
Understanding the need, is where in association with Automation Anywhere Argentina, Ataway offers tools and working models that will allow them to carry out their new challenges with technology and intelligence, with COGNITIVE AUTOMATION.
Initial Project Plan
The project scopes the implementation of RPA and IqBot to automatize the processing of vendors invoices that are received by email or Ariba to the posting in SAP.
Project duration: 6 months
We face many challenges in this project:
- The complexity of the Argentina’s tax system with almost 30 different fields to be read from the invoices and in some cases with 23 possible values and behaviors in some fields.
- A lack of a standard email format or process to receive the invoices made us create a pre-classification bot to distinguish between invoices and other documents.
- The low quality of the documents to be recognized did not allow get to better success in straight forward processing.
Using Automation Anywhere RPA solution, Ataway team implemented Automation Anywhere IQBot that reads 6000 documents and classifies the legal documents to be processed. Then 50 different bots make the validations and posting into SAP. Another set of bots automatizes and manages the flow of the information and creates reports that are used for operation and maintenance of the process.
Achievements & Outcome
Results achieved: 70% of invoices analyzed, and 50% directly posted into SAP without manual intervention. Numbers are expected to increase as a result to the machine learning process of the invoices by the IqBot.