Teach on Udemy

Turn what you know into an opportunity and reach millions around the world.

Learn More

Your cart is empty.

Keep shopping

Using Azure AI Document Intelligence

Name: Using Azure AI Document Intelligence
Rating: 4.5 (519 reviews)

Use portal, Python and C# to OCR (read) documents, extract key data, use pre-built and create custom models

Created byPhillip Burton • over one million students so far, I Do Data Limited

Last updated 6/2025

English

EnglishSpanish [Auto],

What you'll learn

Create and configure resources
Programmatically analyze data in forms
Create a template-based custom model
Configure model endpoints for document types
Extract key value pairs

Course content

9 sections • 54 lectures • 4h 3m total length

Introduction2:29
Master intelligent document processing with Azure AI Document Intelligence, extracting text, tables, barcodes, and key value pairs. Sign up for Azure, use general models, and train a custom extraction model.
Welcome to Udemy0:38
How to get the best out of your Udemy course3:05
Master the Udemy player and course tools to optimize learning: adjust playback speed and volume, use captions, access resources, navigate videos, take notes, ask questions, and view updates and certificates.
Do you want auto-translated subtitles in more languages?1:20
Curriculum4:14
Learn to build a document processing solution with Azure AI Document Intelligence by creating resources, analyzing forms, and configuring endpoints for template and custom extraction models using Python and C#.
Resources0:20

Creating an Azure account, and exploring Azure interface6:52
Sign up for a free Azure account with 12 months of free services and a $200 credit, and navigate the portal to prepare a document intelligence resource.
1. Creating a Document Intelligence resource5:25
Create a new Azure AI Document Intelligence resource by selecting a subscription, resource group, region, and unique name, then use the standard pricing. Access Document Intelligence Studio for analysis.
Using the Read model9:57
Explore document analysis with the Document Intelligence Studio, focusing on the Read model to extract text across languages and handwriting, and compare it with the Layout model and pricing.
7. Using the Layout model4:21
Explore the layout model, switch output to text or markdown, identify titles, section headings, selection marks, and enable key value pairs for structured results with tables and figures.
4. The Prebuilt models5:58
Learn about Azure AI Document Intelligence prebuilt models, such as invoices, receipts, identity documents, contracts, and US tax forms, and their supported file types like pdf and image formats.
Practice Activity Number 10:22
Practice Activity Number 1 - The Solution7:07
Practice activity 1 guides you through analyzing a fictitious invoice with three document intelligence models: read, layout, and prebuilt invoice, extracting tables, key-value pairs, and fields in the json results.

The Read model5:46
Learn to use C# and Python with the document intelligence resource, focusing on the Read model and the analyse result structure, including content, pages, paragraphs, languages, and confidence.
5, 6. Installing Visual Studio Code and copying connection details4:33
Install Visual Studio Code on Windows, macOS, or Linux; set up a Document Intelligence project; and copy the endpoint and keys from Azure to connect to the resource.

Connecting to the Read model using Python9:12
Learn to connect to the read model with Python by installing Azure AI Form Recognizer, configuring an endpoint and Azure key credential, and initiating analysis from a document URL.
14, 16, 20. Extracting the analysis using Python - Part 17:07
Learn to extract pages, lines, words, and paragraphs from Azure AI Document Intelligence with Python—loop through results, use enumerate and f-strings, and print content with confidence.
18. Extracting the analysis using Python - Part 24:38
Learn to extract analysis from Azure AI Document Intelligence using Python, printing word content and confidence, iterating over pages, lines, words, and paragraphs, and handling read model results.
Practice Activity Number 2 using Python0:40
Practice Activity Number 2 using Python - The Solution4:40
Install Azure AI Form Recognizer, set endpoint and key, and run a Python program to analyze documents. Use the prebuilt-read model and print page lines, words, and paragraphs with confidence.

Starting to create our C# program5:24
Create a C# console program in Visual Studio Code to connect to the Azure Document Intelligence service using Azure.AI.FormRecognizer, configure the key and endpoint, and run a Hello World example.
Connecting to the Read model using C#7:02
Connect to the document intelligence service with a document analysis client, supplying endpoint and key. Analyze a document from a URI using the prebuilt read model and await results.
14, 16, 20. Extracting the analysis using C# - Part 16:43
Extract data from an analyzeResult by iterating over document pages and lines in C#, using foreach and for loops, and printing page numbers and line content with interpolated strings.
18. Extracting the analysis using C# - Part 24:12
Extract and display document content by iterating over pages, lines, words, and paragraphs in C# with Azure AI Document Intelligence, showing word content and confidence.
Practice Activity 3 using C#0:54
Practice Activity Number 3 using C# - The Solution4:46
practice activity 3 in c# guides creating a console app, installing azure ai form recognizer, configuring endpoint and key, and analyzing documents to extract pages, lines, words, and paragraph content.

The Layout Model3:39
Explore layout model with Python and C#, focusing on analyzeResult, tables, and section marks; filter by low confidence using selection marks, and move endpoint and key to a separate file.
8. Extracting data using Python5:47
Use Python to apply the Azure AI Document Intelligence layout model, extracting titles, headings, and checkboxes, and demonstrate looping through selection marks with confidence filters.
Using Environment variables and a local file using Python6:26
Move the Azure key and endpoint into a .env file, load them with load_dotenv and os.getenv, and switch to analyzing a local image with a pre-built layout model.
8. Extracting Data using C#8:17
Expand your C# code to use the layout model by switching to prebuilt-layout, access page selection marks, and analyze a local image via a file stream with AnalyzeDocumentAsync.
Practice Activity Number 4 (using either Python or C#)1:18
Practice Activity Number 4 - The Solution4:28
Practice activity 4 shows setting up Python and C# projects, configuring env with key and endpoint, and running analyze document with the prebuilt-layout model to extract confident data.

The Invoice Model3:22
Learn to use the invoice prebuilt model with Python and C#, fetch a pdf from a url, and extract table data and fields like customer id and address.
Extracting Table Data Using Python4:53
Learn to use Python with the Azure AI Document Intelligence invoice model, switch to prebuilt-invoice, and extract a four-by-four table of quantities, descriptions, unit prices, and line totals.
Extracting Key-Value Pairs Using Python7:14
Extract key-value pairs from documents in Python by iterating result.document and document.fields, printing field values with f-strings, and handling missing keys. Display only string-valued fields to focus on usable values.
Extracting Table Data Using C#7:04
Extract table data from invoice PDFs using C# with the Azure AI Document Intelligence prebuilt-invoice model; iterate through DocumentTable, access cells, and print contents by rows.
Extracting Key-Value Pairs Using C#6:52
Extract key-value pairs from an invoice model in C#, looping over document fields and safely retrieving values with try get value to output existing keys and string results.
Practice Activity Number 5 (using either Python or C#)1:22
Practice Activity Number 5 - The Solution4:21
Copy and clean project files, then run C# and Python code to extract customer names and table data using endpoint, key, and TryGetValue.

13. Introducing custom extraction models3:57
Explore template-based custom models in Azure AI Document Intelligence, training with five examples, handling PDF and image inputs, and preparing storage and format requirements for effective custom extraction.
2, 3. Creating an Azure Blob Storage Account, and upload files to a container5:34
Create a storage account and a blob container, enable anonymous access, upload sample invoices for training, and prepare for a custom training model.
9, 10, 11. Creating Model7:53
Upload training documents to a storage container, create a custom extraction model in Document Intelligence Studio, configure resources and storage, run a layout, and define fields and tables.
10. Labelling Data6:19
Label data across multiple documents by auto labeling tables and drawing regions to capture date, items, sales person, subtotal, invoice number, tax, and total for Azure AI Document Intelligence.
11, 12. Training and Testing Our Model4:20
Train and test a template-based Azure AI Document Intelligence model using custom-addatum, selecting template mode, validating with a sixth document, and achieving 99.5% accuracy with guidance on confidence and retraining.
Using our custom model using Python and C#2:55
Demonstrate using your existing Python and C# programs to call a custom Azure AI Document Intelligence model by updating only the model name, then run the code to view results.
15, 19. Using documents from an Azure Storage Account4:13
Upload and access PDFs from an Azure storage account by generating a SAS URL, then load the document in Python and C# to observe consistent output.
17. Using model results to calculate a total in Python5:50
Transform model results into a running total in Python by isolating column three, filtering dollar values, converting to float, summing, and formatting the grand total.
17. Using model results to calculate a total in C#6:02
Learn to compute a running total from line totals in the third column by cleaning dollar amounts, parsing to double, and formatting the final result as currency in C#.
Practice Activity Number 60:33
Practice Activity Number 6 - The Solution8:08
Configure an Azure storage account, enable blob anonymous access, upload reports, and train a custom Document Intelligence model to extract report date, conclusions, and a project overview table.

Deleting Azure resources1:27
Learn to clean up Azure resources by deleting individual items like the Document Intelligence service or storage accounts, or remove everything with a resource group deletion.
What's Next?2:41
Learn to create and configure Azure AI Document Intelligence resources, analyze forms and invoices, build template-based models, and extract key-value data, then prepare for the Microsoft Applied Skills assessment.
Congratulations for completing the course1:01
Review creating an azure document intelligence resource and analyzing documents with read, layout, and prebuilt models. Practice with python and c#, test a custom extraction model, and prepare for credential.
Bonus Lecture - AI Microsoft Applied Skills0:10

Requirements

There are no requirements for watching the course.
We'll look at how you can get a free trial of Azure for use in this course.
If you want to use the code in this course, you will need a Windows machine or some Linux systems. You cannot currently use this code on MacOS, Android or iOS devices. However, you can watch the videos in this course on these devices.

Description

Microsoft Azure allows you to extract printed and handwritten text from your documents. Among other things, you can extract barcodes, formulas, font styles, tables and check boxes.

This course was created to cover the APL-3002 Microsoft Applied Skills credential for creating an intelligent document processing solution with Azure AI Document Intelligence. However, in June 2025, this Applied Skill has been retired.

It will also help with the Microsoft exam AI-102 "Designing and Implementing a Microsoft Azure AI Solution".

Please note: This course is not affiliated with, endorsed by, or sponsored by Microsoft.

What do people like you say about this course?

Duncan says: Excellent course, especially for beginners with Python (with previous coding experience in another language). I passed the Microsoft Assessment after this with 100%. The course doesn't give you answers word for word but rather teaches you the skills just like a good course should.

Adam says: The instructor knocked it out of the park with this one. Sure, the concepts were pretty simple and that makes it easy, but I've seen plenty of instructors mess this up. The instruction was clear and concise and I had no problem following along.

Li says: This is really a course that fully meets or even exceeds my expectations. The instructor explains each function in great detail, and I have to do all the exercises again. I am really lucky to be able to find this course. Yes! Recommended to you who need to use Document Intelligence!

What will we do in this course?

First, we’ll sign up for a free Azure subscription. Then, we’ll create a Document Intelligence resource, and use it to analyze documents. We’ll use the general-purpose read and layout models, before looking at prebuilt models such as the invoice model .
After that, we’ll analyze the data using Python and C#. We connect to the Document Intelligence resource and analyse documents both on the Internet and on our computer, extracting paragraphs, tables, key value pairs, and more.
Finally, we’ll create our own custom extraction model. We’ll upload documents into an Azure Storage Account container, train the model based on selected fields, and test it using the portal, Python and C#.

There are several practice activities throughout the course, so you can be sure that you are learning.

By the end of the course, you'll be able to analyse your own documents in the portal and using Python and C#, and perhaps even take the official Microsoft assessment. That would look great on your CV or resume.

Who this course is for:

This is for you if you want use Document Intelligence to OCR (optical character recognition) documents, and retrieve specific data such as invoice dates, addresses and comments
This course will guide you through the skills required for the Microsoft Applied Skills credential "Create an intelligent document processing solution with Azure AI Document Intelligence".

Using Azure AI Document Intelligence

What you'll learn

Explore related topics

Course content

Introduction6 lectures • 12min