CHALLENGE

Atlantis Foundries was grappling with a critical issue – their Accounts Payable (AP) bot’s Straight Through Processing (STP) rate had plummeted from a solid 85% to a concerning 54%. This decline was having a direct impact on their workforce, who were forced to manually input invoices due to the bot’s ineffective functioning.
The existing Robotic Process Automation (RPA) platform, Automation Anywhere, was combined with IQ Bot for document extraction, while MS SQL was used for background activities and MS Outlook served as the source for supplier emails containing invoices.

Results

  • Manual Invoice capture time of 5 minutes is now being done by the bot in 90 seconds. Error rate reduction achieved through continuous processing and comprehensive reporting, available around the clock. Empower employees for advanced tasks and supplement capacity by automating routine processes. Automated handling of non-compliant invoices through supplier correspondence. Enhanced filing system: invoices meticulously archived with indexing for swift retrieval, aiding auditors and various reference needs. Big improvement in IQ Bot straightthrough processing (STP), surging from 54% to an impressive 91%, significantly reducing human intervention

Challenges

Atlantis Foundries was grappling with a critical issue – their Accounts Payable (AP) bot's Straight Through Processing (STP) rate had plummeted from a solid 85% to a concerning 54%. This decline was having a direct impact on their workforce, who were forced to manually input invoices due to the bot's ineffective functioning. The existing Robotic Process Automation (RPA) platform, Automation Anywhere, was combined with IQ Bot for document extraction, while MS SQL was used for background activities and MS Outlook served as the source for supplier emails containing invoices.

Solutions

The solution for Atlantis Foundries' AP bot challenges involved a comprehensive architectural redesign. The processing style was shifted from batch processing to handling one email and invoice at a time to prevent complete batch failures due to errors, while robust error handling was implemented to trap exceptions and log errors. The use of SQL for background processing was optimised by structuring it correctly and reducing unnecessary connections to enhance performance. Additionally, the presence of two separate databases was analysed for necessity, and actions like updating SYSPRO usernames were reassessed. This holistic approach aimed to restore the bot's functionality, ensuring efficient Accounts Payable processing and higher STP rates.