Discover in our video how to utilize ABBYY Vantage to perform a sentiment analysis with large language models like ChatGPT(TM).
Hello. Today I’d like to show you how we can integrate ABBYY Vantage with large language models in this case ChatGPT(TM). So what you have in front of us is a very basic restaurant survey, and that’s really the kind of the concept of today’s demo is that we have the ability to get things like a large comments field and be able to grade that based on what we refer to as sentiment or what is the intent of that comment in this case. Is it positive, negative, or neutral? So that we have some generic grading of incoming documents and the feedback we have on those documents.
This is a very basic example of some surveys that we had from a restaurant. Really what we’re grading here is the comments field. That’s really what we’re trying to capture off the surveys. And then what we’re gonna do is we’re gonna pass that over to ChatGPT for sentiment analysis so we can understand whether it’s a positive, negative or neutral response.
And the workflow will look like this. A document comes into ABBYY and we’ll extract that comments field. We will then send that comments field to ChatGPT, and then of course we’ll review it.
I’ve already set that up for you and kind of show you some of these examples here. So in front of us is an example of a survey that we got and you can read the comment is quite positive. And so what we’re doing is we’re passing that comments to ChatGPT and ChatGPT is looking at the sentiment of that comments field and telling us this is a positive statement. We’ll look at another one here where potentially it was just, okay, the comment said my experience was just, okay. So that’s what we refer to as a neutral sentiment. And then of course we have the negative sentiment where we are looking at a comment that obviously says we’re not coming back to eat anymore here. And that is of course a negative sentiment. And then last, just for fun, we have a book <laugh> of a comment where, look, we’re just getting a substantial amount of detail here that tells us this person had a very negative experience. They’re telling us why. Said they’ll never come back again and try other restaurants. And of course we are looking for a negative grading on that statement.
So sentiment analysis is one of these ways that we can take and have ABBYY Vantage extract that data for us and pass that to a large language model where we can grade it for either summarization or sentiment. And in today’s demo, this was sentiment, a very basic example, but really in real life we would use this in some sort of digital mailroom experience where documents may be coming in and we’re looking for whether or not a letter was a complaint letter or a compliment letter and those sorts of things. So hope you enjoyed this very basic review of how we can integrate ChatGPT into ABBYY Vantage.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
ChatGPT is a pending trademark of OPENAI OPCO, LLC. All rights will be reserved.]
Learn in our demo how to manually crop and split documents within a review queue in ABBYY Vantage.
Hello. Today I’m gonna walk you through how we set up manual review cropping on a document. And this happens in places where we actually get three separate documents on a given page or even within a given file. But this is an example of what you see on the screen. I have one page with three separate IDs on them, and really what we would need is those to be three separate documents. So in ABBYY Vantage, we have the ability to allow a human to tell us where on this page the documents start and stop so they can crop them. And then from there we can set up the rest of the extraction and classification processes automatically.
So this is very common, like what you’re gonna see in business. Somebody just gives us three documents, three in this case IDs on a given page. But we actually need those to be three separate documents for storage and automation downstream.
So what we’re gonna have here is what we refer to as a process skill. And today’s process is gonna be what you see here on the screen. We’re gonna bring a document in, we’re gonna send that to a cropping review screen. So that’s the very first thing that’s gonna happen is we’re gonna ask the human to crop this image to make sure that we’ve auto separated it. Now what that does is it will create one document with three separate IDs. So if you remember here, when we crop, what’s gonna happen is we’ll have three separate IDs, but all within a single document. So that actually creates three pages. That’s not quite what we need, but good enough for the user. So we don’t wanna burden the user with having to assemble the document manually. So we will do that automatically. We will assemble that through what we call our assemble activity, using our classification step. We will then extract the information from those IDs and then of course we’ll review the results together. But the idea is, is that a human’s gonna come into Vantage, crop that document for us, so we know where those IDs start and stop, and then we’ll automatically assemble them here.
So let me show you kind of this process. So we’re gonna go ahead and upload this ID PDF, and that’s going to go to our process skill.
Now this should happen pretty quick. What it’s gonna do is it’s gonna send us right into that cropping activity that you saw there on the workflow. So we would just hit our review button here. And what we’re gonna ask the end user is to look at this document and tell us where these IDs stop and start. So no big deal. We’re gonna crop this image here using our cropping tool. And all we need the end user to do is just tell us where these documents start and stop. So this looks like a good ID. This looks like a good ID. And lastly, this looks like a good ID as well. And we’re just gonna go ahead and apply the crop.
So this is what the software’s done now. So we have a document that came in and this document has three pages. Like I said, that’s not necessarily always ideal. Typically in downstream processes, we would actually want these to be three separate documents, not a document with three separate pages.
What we’re gonna do is once again, carry down the workflow. So we’re right here at the review and crop. Now we’re gonna tell the software to go ahead and assemble that for us, and then we’ll go ahead and extract. So let’s just kind of release this from our queue. We’ll save and close this. We don’t need to extract any data yet. And then we’ll go ahead and complete it.
Now this task will be completed. So what we’re gonna do is we’re gonna kind of see that here. The software’s gonna continue processing that image. And then what we will do is we will have a queue that shows three separate documents.
So this is now what we have. So we’ve cropped it. And now you can see here I have three separate documents, each of those that we’ve manually cropped here, but now we’ve classified them correctly as identity documents and we’ve extracted the data from them correctly. So we can kind of see here on these documents here. Not only do we know the document type, but obviously now that we know the document, we can use our intelligent document processing extraction technology to extract the critical details off of those documents. So then what happens from here on is really up to you as the citizen developer to take these documents over to data and perform interactions downstream.
But the critical part, once again, is knowing the workflow. Is we have this cropping mechanism that allows the end user to intervene and tell us where documents start and stop. And that’s all we’re asking that end user to do. And then the rest here, we will handle ourselves through our activities in the ABBYY Vantage Suite.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
Adobe, Acrobat, and the Adobe PDF logo are either registered trademarks or trademarks of Adobe in the United States and/or other countries.]
Discover in this video how to perform Address Parsing in ABBYY Vantage.
Hello. Today I’d like to show you how we perform Address Parsing within ABBYY Vantage. So what I have here is obviously a very simplified version of a document that it has an address. And on this address we wanna extract the name, the street, the city, state, and zip.
What we’re gonna do is we’re gonna go ahead and create a brand new document skill in our Advanced Designer within Vantage. We’ll call this our Address Parsing Skill. And when we do that, the first thing we’re gonna do is we’re gonna go ahead and upload a document, just kind of as our reference document.
And the next thing we’ll do is we’ll go ahead and walk through the setup. We’ll go ahead and map some fields. When we map our fields here, we’re gonna have a few things. So we know that we’re gonna have a field that describes the address. So let’s just go ahead and start mapping some things here. We’re gonna call this the full address because it’s the whole thing. And then we know that we’re gonna want fields for each of these independent parts of the address. So let’s just go ahead and add some fields here for the name, the street, the city, state, and zip.
All right. Now that we have our fields set up, the one thing I like to do at this point is I like to make sure that we have our reference details completed as well. This gives the software the ability to compare what it extracts versus what we told it is really the truth. We call that our reference spot. And so what I’m gonna do is I’m gonna go ahead and find each of these fields just so the software when we do it automatically has something to compare itself against for the truth. So we know for the name, the name is gonna be located here. We know for the street. For the full street, we’re here. For the city, we are located here. For the state, we’re here. And of course for our zip, we are here. So we have the full address. But then of course we told the software what the truth is for the other fields that we wanted to automatically extract.
Now that we have that set up, let’s talk about the activities that we’re going to need. So we have within our software, the ability to extract based on rules. Then we will perform Address Parsing. Now the last part of the video will be to actually get the name on the address, which the Address Parsing Module does not give us, so we will use our Named Entity Recognition to get the name of the address.
So let’s go one by one here. Let’s go ahead and tell the software how to do the extraction. We want the full address. We want the software to know where to find the full address. So we’re gonna go ahead and just make sure we only map the full address here. And when we do that, we’re gonna go ahead and go to our search elements and we’re going to tell the software that we’re gonna draw this on the image.
So we’re telling the software, “Hey software, this is where we want you to pull the full address from.” Just because I know what’s gonna have to happen here. This is gonna be a full paragraph. So not just a single line of text, but we want the software to grab that whole region here. Now we have what’s called a paragraph of text, which is on that specific region on the document.
So at this point we’re gonna go ahead and test the activity. This should be a pretty obvious one for the software here. We’re just telling it where to pull that paragraph of text. We have a hundred percent results and that’s because the software now compares against the truth. If you remember back in one of the previous steps just a minute ago, we told the software where to find it. Now it’s telling us where it found it automatically and it’s gonna go ahead and tell us if there’s any difference, which in this case there is not.
So now we told the software where to find the full address. Now we want to teach it how to parse the address. So we’ll go ahead and add our Address Parsing option here. And the software’s gonna say, “Okay. Hey, where do I find the full address?” And we’re gonna say, “Hey, you find it from this field.” And of course we’re going to tell it the things that we know it will find. Now just out of experience, we won’t be able to find the name as part of the Address Parsing. We’re gonna come back and do that. So we’re gonna find street, city, state, and zip. And we’ll go ahead and set up our mapping here.
And what we will do here is we will go ahead and test the skill. So the software automatically comes with intelligence that can take that full address field and give us the result here. And now we have our results. We can see where the software was able to extract that data for us. And as you can see it found street, city, state, and zip. Of course it did not find names. So let’s go ahead and now that we know the software can find and parse the address, let’s go back and say let’s teach the software, how we can extract that name.
What we’re gonna do here is go ahead and use our Named Entity Recognition. This gives us the ability to reference the full address here, but in this case what we’re going to do is just grab the name from the documents. So we’ll call this the organization and the software is gonna use the full address and give us the organization here and put that into our name column. Let’s go ahead and test the skill now.
And now at this point we have a hundred percent accuracy. So the software’s using what we told it as the truth and now it’s going to extract the information on that name for us.
So at this point we have taught the software where to locate the full address. And then using our Address Parsing Activity, we were able to get the street, the city, state, zip, we could even get things like country. And then to actually get the entity on the address, we used our Named Entity Recognition. So we have just full complete control of how this Address Parsing takes place. Now what we could do is we could obviously deploy this skill like we do in other situations with the Advanced Designer and use this technique to get the address details and the entity on that address.
Hope you enjoyed this video. If you have any questions, please reach out to us.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
Adobe, Acrobat, and the Adobe PDF logo are either registered trademarks or trademarks of Adobe in the United States and/or other countries.]
Watch in our video how to use ABBYY Vantage to convert a customer purchase order to an EDI 850 Format.
Hello. Today I’d like to show you how we can perform OCR on an incoming customer purchase order and create an EDI 850 file out of that customer purchase order. So what we have here is a very standard flow in ABBYY Vantage. We’re gonna bring a document in, we’ll of course extract information from that purchase order. And what you’ll see here is that we will generate that EDI file. And then for today’s demo, we’re actually gonna place that EDI file as a field to the actual document, just so you can visually see what that file looks like and you can also see the actual document as well. For today’s demo, we’re gonna throw that into a review queue. Of course this is optional in any traditional implementation. But for today’s demo, we’re gonna throw that into a review queue so that we can see the results.
Now I have a pretty basic customer purchase order that I’ll show you. And just kind of what you would expect. It has customer information, it has shipping information, of course it has line items and those sorts of things. So what I’m gonna do is we’re gonna go ahead and put that into our try any skill page. We’re gonna drop that in. We’re gonna go ahead and select our 850 skill.
And that purchase order will now be processed. ABBYY Vantage is gonna extract that information and then eventually we will get a link so that we can review that document in a manual review queue.
Okay, that document is now completed. So we can click our review button here. And what I’ll wanna show you here is, of course we have the document and the purchase order in the system, but just for fun, we went ahead and threw that EDI string here onto the actual document itself, just so we can see that and kind of see it in play with how this document works. So we have obviously our header information and then of course our transactional information here.
So this is a very standard approach. Now what we do with this EDI file is a hundred percent configurable. We can pass the actual file to some downstream service. We can save the 850 file independently besides the document. So we just have a lot of control. The secret here is that we wanna get that document in EDI 850 format so that we can process downstream in an automation step.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
IrfanView is a registered trademark of Irfan Skiljan. All rights reserved.]
Discover in our video how to convert an Explanation of Benefits (EOB) document to an EDI 835 Format in ABBYY Vantage.
Hello. Today I’d like to show you our EDI Connector, specifically targeting EDI Format 835 to process, typically an Explanation of Benefits document. What you see here is a very typical process within ABBYY Vantage. So a document or EOB in this case will come in. We will extract the critical information off of that Explanation of Benefits document and we will generate an 835 format to represent the data that we’ve extracted from that document. So pretty standard flow here from an OCR perspective.
So what I have is an EOB. The EOB has typically header information. So details here at the top and then of course we have our claim information or the transactional detail. So what we’re gonna do is put this into ABBYY Vantage and then what we expect on the output is an 835 formatted file.
So let’s go ahead and do that. I’m gonna use our try any skill site. I’m just gonna go ahead and drag and drop that sample. I will select the Process Skill that we’re going to use today, which is our 835 Connector Skill, and the software will start processing that document for us.
That document is now processed and let’s go ahead and see the output file here that we have for it. I’ll share it from this screen. And here’s an example of that EDI output for that specific file. So we have our 835 format there. We have our different segments here outlined in the EDI format that’s expected, and we can now use that EDI information to populate our backend systems, or pass through our downstream step of the automation process here.
So this is a really cool experience, a very simple way to convert an Explanation of Benefits document into a very standard file format that can be used in your downstream EDI steps.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
Adobe, Acrobat, and the Adobe PDF logo are either registered trademarks or trademarks of Adobe in the United States and/or other countries.
Cigna and the Cigna Logo are registered trademarks of CIGNA Intellectual Property, Inc. All rights reserved.
Notepad++ is a registered trademark of Don Ho. All rights reserved.]
Watch in this demo how to use ABBYY Vantage to convert a Purchase Invoice to an EDI 810 Format.
Hello. Today I’d like to show you our EDI Connector for ABBYY Vantage. And with this EDI Connector we’ll do, it will give us the ability to take an incoming document, in this case an invoice and produce an EDI 810 formatted file. So a document’s gonna come in through a sample site that we have built. We’re gonna extract the invoice details. We’re gonna create that EDI 810 file. And then we’ll produce that here for us to look at through the output stage.
So what I have here is just a very common invoice. We have some header information, we have some line items. And what I’m gonna do is go ahead and upload that to ABBYY Vantage. Just gonna go ahead and upload a document. And we’re gonna go ahead and select our skill. And what’s gonna happen here is the software’s going to take that document in, it’s gonna extract the details, and then it’s gonna produce that EDI file.
Okay, that file has now been fully processed and I just got my EDI file for it. And I’ll share that with you here on this screen. And what you’ll see here is I have my header information as described in the 810 requirement. And then of course we have line item details also described here as well. So what we can do with that file, we can be very creative. We can call an external systems, we can put it on a file directory to be consumed by another third party. But the secret sauce here is that we have that EDI format and can utilize that in our downstream steps.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
IrfanView is a registered trademark of Irfan Skiljan. All rights reserved.
Notepad++ is a registered trademark of Don Ho. All rights reserved.]
Learn in our ABBYY Vantage Demo how you can transform healthcare clinical/administrative data to a Health Level 7 (HL7) Format.
Hello. Today I’d like to walk you through our HL7 Converter Skill for ABBYY Vantage. And the process will be very similar to what you see on the screen here. We’re gonna bring in a document. We’re going to extract the data off of a clinical form, an administrative healthcare related form, so that we can get that HL7 data. We can route that document to human review if we’d like, and then we will output, of course, our HL7 format.
Let me kind of explain to you what we do within ABBYY Vantage. So within ABBYY Vantage, we will take our administrative forms and we will teach the software how to extract information that is critical on those documents, such as what you see here. This is a generalized consent form for a document, and we just wanna extract this information so that we can pass this between software systems. I’ve set up a sample here and as part of our HL7 conversion skill, when a document comes in and we’ve passed a document through this workflow, we will convert that data to HL7 format. So what I’ve done is I’ve loaded a sample here and I’m gonna go ahead and process this document.
The software is now uploading that sample and it is processing that for us. All right, that document is now processed. I have an HL7 format output here that I will share on my screen, and what you’ll see here is the HL7 format for that given document.
So this is a very common transaction process. The document comes in, we’ll extract that data and we’ll produce that HL7 format so that we can pass that to a downstream system or a web service or backend process.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.]
Discover how to convert a Health Claim Form to a usable EDI 837 format in our ABBYY Vantage demo.
Hello. Today I’d like to show you our conversion to EDI, specifically format 837. Eight hundred thirty-seven is the format that we use for health claim forms. And what you see in front of us is ABBYY Vantage. So within ABBYY Vantage, we will teach the software how to classify and extract data from a health insurance claim form such as a CMS-1500, maybe a UB-04 or a dental form. Those would be the most common ones that we see here in our OCR industry. And our goal is to take that data that we’ve extracted and produce a standard EDI 837 format.
And as part of that process, what we’ll do is we will create what’s called a document skill in ABBYY Vantage. And we offer one here kind of out of the box where kind of a starter skill. You can get extraction up and running very quickly here within ABBYY Vantage by just lassoing fields and teaching the software where these data points exist on a health insurance claim form.
Once we do that, we’ll create what’s called a process skill where we’ll teach the software about how we’re gonna get documents in, how we’re gonna get documents out and if or when we want a human in the loop to review the extraction before we export that data. But the important part of today’s demo that you’ll see is that we’re going to drop a document in to our skill, we’re gonna extract the EDI format and then we’re gonna send that to our output.
So I’m gonna just gonna go ahead and drop our CMS-1500 sample form. I’ll show that to you here. It’s just obviously a very standard in this case, we got a fillable PDF and we understand that’s not always the case here in real life, but here we have a standard EDI fillable form. So we’re gonna go ahead and publish that here to our try any skill page. Now our software is processing that form and what we’ll expect on the output is an EDI 837 format.
Alright, our results are ready. So I’m gonna open up our EDI format file that we just received as an output to this document and bring that over here to our screen. And what you’ll see is what you would very much expect here. We have our output file, it’s an EDI 837 format, and it has the specific format that we’re expecting so that we can pass this data and the extracted information downstream to our backend processes and or systems that are expecting this EDI format.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
Adobe, Acrobat, and the Adobe PDF logo are either registered trademarks or trademarks of Adobe in the United States and/or other countries.]
Watch in our demo how to split a document that contains multiple invoices with ABBYY Vantage.
Hello. Today I’d like to show you some document splitting in ABBYY Vantage. And the cool part about this is that the software has the ability to split based on text that is found on the document. And one of the ways that we typically use this sort of functionality is with invoices. So a document comes in that may have multiple invoices in it. So let me show you an example of one here.
In this PDF I have just a few different invoices. They’re completely different organizations, completely different invoices, but these are coming in and we need to be able to automatically split them. So what the software can do is it will read these documents and based on logic and today’s logic is pretty simple. We’re gonna look for extracting the invoice number and splitting it when we see a difference. But we can get pretty advanced in the way that that sort of logic is applied.
So what will happen in today’s demo is a document is gonna come in. We will use the assemble skill to split a document. Now we have what’s called a document splitter skill that’s used to split it and we have full control over that splitting logic, and I’ll show that a little bit to you behind the scenes here in a second. But once we’ve split it, we will then formally extract using our Invoice US Skill. So we’re gonna split using our splitter skill. Now that we’ve split successfully, we will extract using our document extraction skill. And then just for review purposes, we’ll actually come in here and we’ll see that document here that’s ready for review and the reason why we’re doing this in today’s demo so that we can see that splitting actually taken place. Alright, let’s go ahead and fire off a splitter skill.
So now that we fire that off, the software now is going to perform that splitting. So it’s looking for differences in the documents and will auto split those documents. And then based on our process skill, we will route those for human review. So what I’m gonna do here is just monitor our skill monitor in ABBYY Vantage. And you can see we now have an invoice that is in the starting stage. And when that’s ready for review, it will go to what we call the Manual Review Stage. So that is our human in the loop step so that we can see these documents.
So now we have that transaction ready for human review so that we can see the splitting. I’m gonna go ahead and open up that transaction. A couple of things I want you to know is that within this transaction, I now have four separate documents. So if you remember, I had a PDF that had four pages in it and they were each individual documents, and now ABBYY Vantage has split that up for us. Not only that, but we have then after we split, we have then applied against each document, our invoice document extraction skill. So now I have every invoice, it’s content that we read off of it, and of course the line items and those sorts of things on an invoice that we would typically want to extract. So this shows us the power of splitting.
Now in practice, document splitting can be much more advanced. We can use page numbers. We can use differing text. We can look at every single page. So let me kind of show you then behind the scenes how this really gets implemented. So if we go back and we look at our process skill, you’ll recall that we have a document splitter skill that we referenced here in that process skill.
Now, if we go and we look at our invoice splitter skill, it’s pretty simple. We’re going to of course perform OCR on a document as it comes in. We will look for different extraction rules. And in today’s demo, once again, this is very basic. We’re looking for an invoice number. We’re looking for maybe something that indicates the first page of a document, but once again, this is all customizable at this point based on your document type. And then we have a splitter script, which gives us the ability, frankly, to go as customized as we want to for the splitting. So we give you a sample script here to start out with, and this is definitely something we can consult with, but here’s where we kind of tell the software, here’s when a new document starts and stops, and then what to do with merging documents into previous invoices if we don’t see a differing number of, in this case, invoice numbers.
So just a lot of control here that we have now that we have the ability to classify, extract the text, and then perform the splitting. And we have full control over then how that batch is kind of formulated here in Vantage, which eventually gets us to the ability to have just a beautiful batch that’s outlined here as we would expect it to.
So a little bit behind the scenes, but a lot of this is really just low-code, no-code options here where we can assemble automatically. And honestly, within minutes we could have documents coming in, in this case invoices just being split and then routed as needed for our downstream steps. So I hope you were able to take this content and apply it to your business use cases here. Just a lot of power at our disposal.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
Adobe, Acrobat, and the Adobe PDF logo are either registered trademarks or trademarks of Adobe in the United States and/or other countries.]
Watch our demo to learn how to extract document data with our ABBYY Vantage Connector and send it to Snowflake® to be stored.
Hello. Today I’d like to show you our Connector for Snowflake®. Now what this will do is take a document and its data that we’ve extracted from ABBYY Vantage and pass this over into our data warehouse and Snowflake®.
This is a very typical workflow for Vantage. So a document will come in, we’ll extract that data for the document, and then we have our send to Snowflake® Connector here.
It’s pretty simple. So what I have here is a warehouse and a database in Snowflake®. And as you can see here, I have a table called invoice item as an example. And as we flow a document through ABBYY Vantage, we will then populate this as our output step here. So I’ve preloaded a document here that’s ready to be processed, and I’m going to select our Snowflake® Connector, and that will start processing now. When I refresh our Snowflake® table here, we will then see that updated with the line items that we’ve extracted from that invoice. So once we’re done processing here and we see this status updated, we’ll go ahead and refresh our Snowflake® screen.
All right, so now this is processed. I’m just gonna go ahead here and click our execute button. And you’ll see here we now have extracted that information from Vantage, and we’ve populated our database here in Snowflake®.
[Music- “Engineered to Perfection” performed by Peter Nickalls, used under license from Shutterstock.
The word “Snowflake”, “Snowflake” and the “Snowflake” logos, and other marks, logos and titles are registered and/or common law trade names, trademarks or service marks of Snowflake Inc. Other materials such as the text, content, and product previews are also owed by Snowflake Inc. and/or its Owners. Snowflake Inc. reserves all other rights.]
