Invoice Processing with IDP

The below information is a problem statement and the implementation life cycle for an Account Payable automation project. With the assessment we will demo the extraction accuracy with your own documents. This will also display the audit trail to you showcasing the accuracy and the flow to get to the results giving you the comfort that the solution has strong underlying technologies and algorithms from leading players in the market.

In the session we also discuss the potential design of the Power automate solution for Invoice processing.

Problem statement: The Accounts Payable automation allows all the documents encoded in a human tailored consumption and the solution extracts fields from various document formats. Dealing with different creditors means that any different formats of invoice will be submitted, thus the solution will be content-driven machine learning, and this allows to handle any unseen documents. The goal of the Accounts Payable Automation is to automate some parts of the business operation with the help of intelligent Document processing (IDP) technology to reduce human intervention, to a certain degree not all the aspects of the account payable business will eliminate human intervention. For example, in some cases, the exception may occur, and humans may need to decide and fix the issues and verify the correctness of the invoice prior to the payment of the invoice.

In summary AP automation advocates for the following when it comes to processing invoices automatically.

• Interpretation of invoice with the aid of IDP technology to extract data from invoice submitted in a format of PDF. • Validation of invoice data: using machine learning to teach the system to make decisions about the correctness of an invoice. • Assistive technology, to a certain extent some form of user intervention may be required to deal with exceptional situations. For example, in the case where invoice format has drastically changed from the initial usual invoice.

The implementation of the AP Automation solution will drastically improve the business process of accounts payable where the automation part of the invoice documents will enormously assist with the accuracy of invoices created and disbursement to the creditors.

Solution delivery life cycle:

• RPA Solution design • IDP Solutions design • Integration design • Field extraction validation and verification setup • Classification setup • Business, validation rules and exception design en documentation • Model training and testing • Rules implementation • Human in the loop setup • RPA solution development • Integration • UAT • Deployment Business Rules:

When looking at the business rules it is important to consider: • What additional data is required to run / apply a rule on the extracted data? • Is the data easily obtained (can we pull it from a system or do the business maintain the data in Excel)?