The Process

The client is a leading global industrial manufacturer. Invoice processing is a crucial problem for the client due to the hundreds of formats of the invoices. The Accounts Payable team then extracts the data from these invoices for entering in the Enterprise Resource Planning (ERP) system. However, the variability of the formats makes this task tedious and repetitive. Also, the frequency of changes to the invoices makes standard automation an expensive solution. The additional checks performed by the Accounts Payable team also had manual entries in the “Goods Received” register. This also added complications to the process, which caused delays and errors. The client was looking at a solution that could take care of these issues.

The Solution

The solution involved applying Optical Character Recognition (OCR) to convert invoice documents to text. This digitized data was then fed into Aiwozo’s AI/ML model to extract relevant information from the invoices. Historical and real-time data was used to train the Artificial Intelligence (AI) model, and accuracy kept improving over a period of time. Any exception in identification was flagged to humans, and the corrected information was used to enhance Aiwozo’s Artificial Intelligence (AI) model. Once the extracted data was verified, it was then handed over to the Aiwozo Bot for entering into their ERP software and processing of the invoice.

See Aiwozo DocuBot in action:

Challenges Addressed