Modernize document workflows with intelligent processing

Bruce: Thanks. Can everybody hear me? Ok. All right, I can hear myself too. One of the things is I have to be careful what I say because everything appears on screen. So this is gonna be the PG version today.

We're going to talk about document processing, but I want to talk a little bit about workflow in general. Most businesses, their data can be processed by RPA. So what RPA is robotic process automation. 80% of data in companies today is unstructured. So you have information for HR you have information for finance, you have information for your legal department. It's all over the place.

Look at this by 2025 we're gonna have 100 and 80 zetta bytes, zetta bytes of data. What is that? Put it this way? One zetta byte of data is 1 billion terabytes. There's so much data out there. So being able to automate the process if it's very important. Since 2010, we've had naughty times the data being created and I can attest to that because my wife has two terabytes of data on her own cloud storage. Crazy.

The McKinsey Global Institute estimates over the next 5 to 10 years, you can have 15 to 40% of our world economy increase by AI. So all the stuff you see here today is going to contribute to rapid growth in our economy.

This is a quote that everybody can read that I found in one of the MIT journals. And basically what it's saying is this in the seventies and eighties, when computers were coming up, people that adopted that technology tended to do well, companies tended to do well, the same thing with AI. So AI is new over the last 3 to 4 years and it's going to be with us the next 1520 years. So it's basically saying embrace it if you embrace it and you're good with the change, good things will come of it.

So invasion, our vision is make digital transformation, attainable to everybody. So what does that really mean? Two things? One, we try to eliminate infrastructure. So we're a SaaS platform. So we eliminate file servers, print servers and two, we try to automate business processes. So making things more efficient.

So why do we come up with this product? Right? Why do we come up with this solution? Well, in the marketplace, we found that there's a lot of different products out there. Some do workflow, some do document management, some do online forms, some do print management, but none do it all together in a single platform. So at vision that's what we created. We created a single SaaS platform that includes data capture, digital signature workflows, content management, and print management. And I want to stress this, we right now to the best of my knowledge are the only platform that has this end to end solution in one's product, there's other products out there that does some workflow, does some capture, but they don't do the print management. There's some print management products that maybe do a little bit of scanning but don't do the digital signature or online forms. So that's a big differentiator. So if you take one thing away, single platform, ok.

So before I get into what we're gonna talk about today around intelligent document processing with AWS Textract, a couple of things about the product or the company. Uh we're a Saint George based company. We've been around roughly 20 plus years. We have literally over 10,000 clients. We just hit our 10,000 client, I think sometime three weeks ago, roughly it was some city in, in Florida. So we've been around a while. We've been doing print management a long time. We've been doing document management and workflow a long time. But one of the things that has changed is the fact with the Textract product, the ability to take documents and instead of having people key in information to automatically extract that information and kick off document workflows.

So this is our standard interface. This is what you see here. And if you were an admin user, you would see this interface as a typical user. If you're only involved in workflow, you'd only see the workspace app. If you're only involved in storage or scanning, you'd only see the capture or storage app.

So let's talk about how you get documents in Vision. So when we're talking physical documents, you can get documents in by everybody knows what a copier is, right? You can scan them in, you can have a desktop scanner. In this particular case, I'm just pretending I'm importing some financial documents. These happen to be invoices. What ends up happening is I send them to a document folder that is monitored by AWS Textract. So as the documents come in, they automatically get sent to AWS Textract and then Textract as you see in the bottom starts sending us back the information with key metadata fields.

So in this case, when I send it to Textract, it's automatically grabbing the invoice number, the vendor name, the amount this is without any type of scripting with any type of zonal OCR. It just we're sending the raw document there and it's sending us back the data.

When all the documents have finished processing, then what Vision can do is then we can automatically send them off to the appropriate workflow. But before we do that, we wanna do some type of validation. So when data gets extracted, unless you're going to basically integrate with a line of business where you do a look up. So let's say i read invoice number 1234. If I want to make sure it's accurate, I want to go to my Microsoft Dynamics GP database and make sure it matches if it matches. I kick off the workflow. If not, you go through some type of validation.

So invasion, we have a very quick interface. So let's say you had four or 500 documents that you just processed invasion. You can bring up the validation document view. It brings the document on the left hand side, you can quickly see any errors, it will highlight them in red or yellow. If it's a low confidence level, you'll be able to scroll through the whole batch, quickly update them. And then once that process is completed, you can kick off a workflow.

But the other great thing with Textract is it does handwriting. So I've been in the business since 95 is when I started in the workflow document management. space long time, handwriting has always been a challenge even up to three or four years ago.

So when we talk about a use case that we have coming up, handwriting was particularly important. And you can see here this is some handwriting, some of it's uh printed, some of it's written and you can see the accuracy. It's pretty amazing like the fact that on a typical document, you can get 85 to 90% accuracy in handwriting is mind blowing because that wasn't there five years ago.

And once again, just another example of some handwriting of a poor quality.

So after the documents have been validated, then they can get sent off to the various workflows. So here is where AI really helps, right? AI helps with the extraction. AI can help with the determination of what workflow it should go to. And in the end AI basically allows you to improve your process somewhere between 80 to 90%.

So I'm gonna walk through a use case. Look at this, I'm good on time. I got 11 minutes and 40 seconds. So you guys better have some questions. Ok?

So in the use case, um this is actually one of our clients that joined us probably 2.5, 3 years ago. So this is a large county uh department of family services and they came to us with a problem, right?

So I'm just gonna click up the whole slide here so we can view it. Ok?

So they had an existing case management system. They were using an existing product. They were having bottlenecks in the sense that they were using barcode cover sheets. They were still having a large percentage of errors, roughly about 20% of errors per day and all their information was time sensitive. And in the end, all the documents that get processed had to get up to a court system and there was no automation to the court system.

So it basically made it kind of hit and miss they get their documents in. There was lots of errors, lots of manual interaction. Did it get sent to the court system? Did it not get sent to the court system? And guess what? It was a lot better than it was 10 years ago when it was manual, right?

So when they came to us at Vision, they basically said, well, how can you improve this with AWS Textract? So what we ended up doing is we got with the team at AWS, we put our heads together and we built a better mousetrap, right?

So what we ended up doing was the case manager still can basically scan in or upload the files. We came down to 5% from 20%. I'm gonna explain that in a second. So it it doesn't get down to zero. I'm not here to say, oh no, zero, no, it's realistic. There are still some data issues and that's why validation is important.

But in the end, what we were able to do is between our interface and the fact that there was still 5% errors, we basically increased their process by 95% efficiency overall. And when they were dealing with mission critical documents that is critical.

And the other thing that they did with our solution is it basically tied in to their court system. So now they basically had a streamlined process where they could scan, they could automatically extract, they can validate and then the information move seamlessly into the court system.

So that was the best i can do in a 10 minute presentation. I wanted to open it up for questions because when you're dealing with workflow, right? Or automation, right? AI can do a lot of different things, right? So AI is not just about data extraction.

Why I brought up the thing at the beginning of RPA robotic process automation is because Vision, we do a lot of repetitive system tasks behind the scenes. So I'll give you an example. Let's say you're in an accounts payable application and vendors email you your invoices. Well, with Vision, you can have an RPA process that monitors an email address, imports that document automatically extracts the data and then based on the data extraction, sends it off to the correct person for sign off. That's an example.

Now compared to other presentations today, we try to right size in the sense that we're not going to get into the weeds technically and uh how, how everything is built but invasion. Everything is simple. Everything from building a workflow to building an online form to setting up your routing rules is all via drag and drop or using drop down list.

So thank you for your time. Awesome, great. Thank you Bruce. Thank you. Vision