The challenge in invoice processing

Issues due to excessive human involvement:

  • Forgetting/ procrastinating invoice generation – Low efficiency, delayed payments
  • Infrequent follow-ups/ unintended invoice sharing – Delayed payments and unhappy clients
  • Tedious data extraction/ inaccurate verification – Wrong invoicing
  • No appropriate data backup – Threat of permanent data loss

The solution

Nalashaa’s team, after analyzing the workflow and loopholes, automated and optimized the invoicing process with RPA bots (using UiPath).

The process

Auto-reading of invoices: Bots identify and classify invoices and read invoice data in predefined formats with an accuracy of 98%-99%. After this, the structured data is stored on the cloud.

High efficiency: As bots work round the clock, all delays are eliminated and individuals are relieved of any mundane jobs. Therefore, no incorrect invoices are processed.

Accurate validation – Bots validate invoices based on predefined criteria, update invoice data on ERP, and send notifications for the same. For failed validations, they auto-route invoices to the concerned individuals for further actions.

Easy reporting: Bots were coded to send a report for each invoice processed and a consolidated one, in a preset format, at the end of the day.

Foraying into ML - The validation reports are used as feedback for machine learning capabilities proposed to be built into the bots.

The Payback

Technology Stack

  • UI Path
  • Google Tesseract
  • .Net
  • ML