LEO: Automation for Accounting Teams

by Xpand IT

Automation solution for invoice processing and interpretation

LEO is a solution supported by Power Automate and Microsoft Teams, and its purpose is to help automate and digitise daily tasks in the Finance Department, such as processing invoices. We know that all Finance Departments need to process an enormous number of documents, and this is where LEO is indispensable to finance teams, freeing up precious time so that they can focus on other more complex tasks.

The process begins when a new invoice drops into an inbox: LEO reads the email, verifies that it complies with pre-established conditions (was it sent to the accounting team’s email, does it has a specific subject line, and other criteria – for example, does it have an attachment?) and if these conditions are met, the invoice is then processed using AI Builder. With AI Builder, we build and train models to recognise fields in the documents being analysed, such as, for example, the amount, the billing period and the name of the supplier. When dealing with invoices, one of the most important data to analyse is the amount. Clearly, we need to approve the amount on the invoice before moving it further into the organisation’s systems. A request for approval is therefore sent via Teams to Accounting, and a team member must approve it.

Once the invoice has been approved, and with the data extracted by AI Builder, it becomes possible to automate the entire process of assimilating the invoice into the organisation’s systems. Taking advantage of the RPA (Robotic Process Automation) technology that Power Automate gives us, we can register and process the invoice automatically in just a few seconds and without further need for human intervention. Finally, the Finance team can also use LEO to help them as a chatbot: LEO can answer a series of questions related to invoices. LEO can list all the invoices that have already been approved and give insights into what invoices have been rejected or approved over the last week, among other relevant questions.

Supported language: English. Per request, we can make the necessary changes to support your own language. 

At a glance