President R. Kelly Haws and Sister Connie Haws | BYU–Hawaii Devotional (Livestream)

[0:00]Hello, everybody. My name is Matt Modrick, and I'm a solutions architect with AWS. I'm here today to give you a quick run-through and demonstration of the new document understanding solution. What is the document understanding solution? It's a reference implementation that uses artificial intelligence to automate document processing. So whether you're extracting data from documents or you're classifying documents, this solution allows you to apply artificial intelligence to accelerate and automate those business processes. How does this solution work? Well, it's very simple. You upload documents to an S3 bucket. Once those documents are uploaded, it kicks off an event, which invokes a series of step functions. And within those step functions, we're invoking AWS services like Amazon Textract, Amazon Comprehend, and Amazon SageMaker to apply artificial intelligence to those documents to extract the data and classify those documents. Once the data has been extracted, we store the output into an Amazon S3 bucket, and we also update the metadata in an Amazon DynamoDB table. From there, we have an API endpoint that allows you to integrate with other business applications and systems. For the purpose of this demo, we're actually going to use our own web user interface to view those documents. So let's jump into the demo. All right, so this is the document understanding solution home screen. From here, you can see all the documents that have been uploaded into the solution. You can also see the document status. So whether or not the document has been successfully processed, it's still in process or there's some type of failure. We also have the option to search documents using keywords or values, which I'll show you in a little bit. And then we also have the option to upload a document, which is what we're going to do right now. So I'm going to go ahead and click the upload document button, and I'm going to choose my W2. So here you can see the file name W2. You can also specify the document type. We have two built-in document types already, an invoice and a receipt. But for the purpose of this demo, I'm going to show you a W2 form, and it's going to be a custom document type. So I'm going to go ahead and click upload. And here you can see the W2 has been uploaded and it's in process. So what's happening now is it's kicking off those step functions and it's going through that workflow that I showed you in the architectural diagram. So it's invoking Amazon Textract and Comprehend and SageMaker to analyze this document. So I'm just going to give it a minute or two to let it process. All right, and you can see that the document status has been updated to successfully processed. And you can see the document type is a W2 form. So from here, I'm going to go ahead and click on the document. And here you can see the document on the right-hand side, which is a W2 form. And on the left-hand side, you can see all the data that's been extracted from the W2 form. So you can see the classification, which is a W2 form. And then you can also see all the data that's been extracted by Textract. So for example, if I wanted to find the employee's first name, I can see that's Matt. I can also see the social security number, which is all redacted. And that's actually a feature within the document understanding solution. So any PII that's found in the document, we're actually redacting it and blurring it out. We also have the employer's identification number, which is also redacted. Again, this is a feature of the document understanding solution. So as you can see, all the data has been extracted from the W2 form and it's presented in a nice JSON format. This allows you to either update the values or copy the values and integrate with other business applications and systems. So I'm going to go ahead and close out of this. And now I'm going to show you the search functionality. So earlier I mentioned that you can search documents. So for example, if I wanted to search for my first name, Matt, I can actually go ahead and type Matt in the search bar, click search, and it'll pull up my W2 form. This is a great feature if you're looking for specific documents or you want to categorize and search for documents within the solution. And that's a quick run-through and demonstration of the new document understanding solution. If you're interested in more information or you have any questions, feel free to visit our public facing website, or you can check out our GitHub page. Thanks for your time, and I hope you enjoyed the demo.