We are all familiar with situations in which we or a third person spend part of their time processing data from one medium to another or from one information system to another.
For example, imagine that we work in the accounting department of an organization and one of our tasks is to enter the invoices of our suppliers into the ERP of said organization, which reaches us in PDF format via email.
The traditional working method consists of:
- Download email invoice
- Access the ERP
- Manually enter the invoice data to finalize the posting of said purchase
We will all agree that it is a tedious process. However, by using the AI Builder feature, we will be able to improve the efficiency of this process and therefore the productivity of our employees.
What is AI Builder?
AI Builder is a feature of the Microsoft Power Platform that helps us automate processes and predict outcomes using pre-built Artificial Intelligence models. That is, users can implement the capabilities of such models without having to write code.
AI Builder’s pre-built AI models can be used within Power Apps applications or Power Automate automated workflows, so we can find this feature in both Power Platform applications, as seen below.
Invoice Processing Model in AI Builder
In this article, we are going to focus on one of the pre-built AI models offered by AI Builder: invoice processing.
As we have seen previously, we can access AI Builder models from Power Apps and Power Automate. In the example developed below, we access Power Automate. Once inside Power Automate, we select the Invoice Processing model.
Next, the AI Builder itself will show us how the selected model works.
In this case, we see that the model uses an OCR (optical character recognition) component to identify data in a document and be able to process it in another format, as we can see in the Extracted information section.
When creating this Invoice Processing model, AI Builder will present us with a wizard that will guide us in generating it.
First, the model will ask us if the documents to be processed by the OCR component of the model are structured, that is, if they have a fixed structure, such as a company invoice model, or not.
In this example, we’re going to assume that our supplier’s invoice model will be the same with all invoices that the AI Builder model will process.
Next, we will need to indicate what information, in our case, fields we want the AI Builder model to retrieve.
At this point, we must provide the model with a set of documents so that the model can be trained, identifying the format of the documents and detecting the information that we want to extract.
But first, so that the model can start learning, we must manually mark the information that we want to extract from the documents that we have provided to the example model. Once this step is complete, we can start training the model.
When our model has finished its training, AI Builder will show us the accuracy of its prediction.
At this point in the example, we’re ready to embed and use the model within a Power Apps app or within a Power Automate flow.
Flows in Power Automate
We could set up a flow where, by dropping an invoice document from a specific vendor into that vendor’s folder on our Sharepoint Online site, the flow would call AI Builder‘s invoice processing model to pull the desired information and place and store it wherever we wanted: an Excel table, a SharePoint list, a Dataverse entity, etc.
What benefits does AI Builder bring us?
The main benefit of using this function is clear, it will allow us to do more with less. It is a simple idea, but very powerful.
The more time we can free up for our teams with task automation, the more time they can spend on tasks that add the most value to their organizations.
In addition, freeing our team from repetitive and boring tasks will allow us to have time to tackle creative and more motivating tasks, improving the efficiency and satisfaction of our employees.
Nico García Fernández – Senior Product Owner at Itequia