Client orders can come in different forms via web portal and mail. One of our clients was struggling to process all these orders as each had its own format and wouldn't deviate from this. Previously sales admins would insert the sales orders manually which created a lot of overhead doing this repetitive task. The main goal of this project was to automate the full purchase order to sales order process (output SAP) no matter how unstructured the PO document is.
To get the required structured data, the project uses an IDP tool to get an initial semi-structured output. Other cognitive services (such as Customvision.ai and Luis.ai) augment the process by extracting entities related to steel specifications and keeping the process clean from unwanted documents (non-po). Once the customer data is structured, a rule engine maps the customer data to the customer's language. The end result is a sales order booked in the SAP system. As many components play an important role in the full process, a dashboard is created to track and provide meaningful insights into the PO data
and transform your business to tackle the digital age more efficiently.