The challenge
This NDIS plan-management and support-coordination provider helps people administer their plans and services. Processing invoices is part of that day-to-day work.
The provider wanted more invoices to progress automatically and fewer to require manual handling. The solution needed to read PDF documents and apply the business rules required to process them, not simply extract text from a page.
Document information
Extract useful data from PDF invoices.
Processing rules
Handle the business requirements needed for invoices to progress through the workflow.
What we did
8Squad combined MuleSoft RPA with AWS Textract OCR and a tailored logic layer to support invoice processing.
Extract invoice data
AWS Textract read information from PDF invoices.
Apply the required logic
A tailored layer within the RPA workflow refined OCR output and addressed processing requirements.
Connect the workflow
The solution included MuleSoft RPA and Salesforce Email-to-Case.
What changed
After four weeks of development, the project reported processing success above 60%, compared with a reported 10% straight-through rate for the previous tool. OCR, automation and tailored rules provided a route for more invoices to progress without manual handling.
Reported development processing success
Over 60%
After four weeks; comparison with the previous 10% rate requires like-for-like validation
| Focus | Goal | Delivered capability |
|---|---|---|
| Invoice handling | Increase processing without manual intervention | OCR and RPA combined with tailored business rules |
| Development result | Improve on the previous tool's reported 10% rate | Success above 60% reported after four weeks of development |
