Task: The client had a clear business need for process automation in an area too complex and not covered by most RPA vendors – intelligent decision-making, many diverse and changing data sources, updates to core financial systems. Two previous projects had failed with leading RPA vendors.
Many client staff are located around the world and incur travel expenses which, in certain circumstances, are rebillable at different rates depending on the individual client contract. There are thousands of different contracts which dictate the approach. A customer invoice may need to be generated and, in all circumstances the GL system updated with a receivable or an expense. The entire process is very complex. Expense data comes from many different sources, both structured and unstructured, including pdf’s, handwritten chits, bank generated data files, etc. Thousands of pdf pages need to be searched, contracts analyzed, etc. The unstructured data needs to read, treated, structured, then integrated with other applications, e.g. S4/HANA.
Duration: The project was successfully completed in 8 weeks. The previous two failed RPA attempts had taken nearly a year. The project involved OCR (sometimes from handwritten notes) text recognition from pdf; unstructured to structured conversion, automated scheduler, triggers/ events in workflow, and intelligent decision-making.
Key differentiations / Value: There was significant value in avoiding time consuming manual work, ensuring accurate client billing and minimizing reputational damage. Following two failed attempts with leading RPA vendors, the client and consultant needed confidence in the outcome. The r4apps ability to read data from various pdfs, even if only “graphics”, the ability to identify relevant information out of tables with heterogenous structure, combine data sets in a new data model to search list, update S/4HANA, etc provided that assurance.