How to use Microsoft AI Builder to Extract Data from PDF

In this step-by-step tutorial, learn how to use Microsoft AI Builder to extract data from a PDF document. Along with extracting text fields, tables, and text boxes, you can use AI Builder to determine text sentiment, identify objects in images, and much more. Once we extract data from the PDF you can automatically insert the data into a spreadsheet, Microsoft Teams, or any other connector. You can leverage your new AI model in Power Automate and Power Apps.

We begin by entering the fields we want to extract from the PDF. Next, we upload sample documents into a collection. With the documents uploaded, we tag the documents to teach the computer what data is contained within and where the data is located. We then train our model based on the sample documents and tagging. We then upload a sample document to ensure that the model works as we expect. Lastly, we create a Power Automate flow that leverages the model. At the end, you’ll be able to automatically extract data from a PDF document.

• Additional resources
  • Microsoft AI Builder License details: https://powerapps.microsoft.com/en-us…
  • Get to AI Builder using Office.com: https://www.office.com
  • Introduction to Power Automate video: https://youtu.be/dDO4Y4aDYXw
  • Introduction to Power Apps video: https://youtu.be/2RzCbd5XgJg
• Timestamps
  • 0:00 Introduction
  • 1:02 Caveats
  • 1:25 Get to AI Builder
  • 2:17 AI Builder overview
  • 2:42 Choose info to extract
  • 3:38 Create collection
  • 4:22 Tag documents
  • 5:06 Train & publish model
  • 6:04 Use model in Power Automate
  • 8:26 Run test
  • 9:15 Wrap up