So I want to start with what you get when you hire Kendrick.
When you hire Kendrick, you get an end-to-end workflow that allows you to ingest, validate, and then input identity documentation. That means you get the entire end-to-end workflow and you get our pre-built machine learning models, and this is where the intelligence comes in. So our machine learning models are pre-trained to ingest driver’s licenses, passports, national IDs, to perform any kind of actions such as rotating them if they’re flipped, kind of splitting them apart if they’re part of multiple document packages. And so Kendrick really provides the end-to-end process for doing the ID ingestion as well.
Kendrick comes with pre-built integration. So I’m going to show you our customer management system integration with PEGA, but also you have the ability to integrate via open API or even RPA to multiple different systems with a very standard interface. And then finally, when you hire Kendrick, you always get to see how he’s performing. So we have pre-built analytics dashboards for you to understand what’s happening in the workflow and statistics about how Kendrick is performing and even his interaction with his human teammates.
With that being said, I’m going to actually demo Kendrick. We are starting here in the PEGA case management system. In this scenario that I’m demo-ing, Kendrick is trying to select identity documents (you can see the requirements over to the right) for two individuals that we’re trying to onboard as a new entity at this bank. Kendrick can ingest the IDs from multiple different formats. If you recall on the slide before, Kendrick can intake IDs from email, through an upload document portal, or even through file share. So really he’s very flexible, exactly like his human teammates to be able to ingest IDs in any fashion that they might be received.
In this scenario, I have Kendrick paired with an additional digital worker because he is going to be validating the IDs, but also validating them against other information that we have on these individuals in the system. So I need to collect some requirements from Alison and Timothy for this company. And so I’m going to go over to email, and Kendrick is going to ingest this from my email inbox. So I’ve got an email from World Healing Church, and this is containing ID documents and a couple other documents I need for onboarding. What Kendrick does is he actually can interact with my inbox and he will automatically pick up information coming into here and move it into the processing folder so that he always takes the first pass and I only need to do things such as review escalations or errors. Really saves time for me. So you can see, for example, we have an ID here, Kendrick’s going to pick this up and actually start processing this through the workflow.
I’m going to flip over to our WorkFusion Control Tower. WorkFusion Control Tower just gives you a visualization into what’s going on in our automation process. This is our visual workflow representation and our visual workflow designer. And you can see what Kendrick’s actually doing behind the scenes. So he picked up an email from my mailbox, he downloaded all of the attachments, and he actually started to digitize these documents by running them through OCR. So WorkFusion doesn’t just digitize documents, but actually extracts key pieces of information from the documents. So you can see that Kendrick is classifying to understand if this is an ID or not an ID. Once he’s identified the IDs, he will run it through the relevant machine learning extraction model for that document type. Here, you can see, we have representation of extraction from driver’s licenses or passports or national IDs. He’s intelligent enough to select the correct model for that.
If Kendrick runs into any issues such as he’s unable to pull a key piece of information from an ID or oftentimes there’s a mismatch between the data that he pulled and maybe something that someone filled in, in a customer form or an online form, he’s able to escalate these issues to his human teammates. And that’s where, in the white, human workers can come in and really just review and handle exceptions.
Another benefit of using WorkFusion and then automation for Kendrick is he’s always collecting the results from his human workers to actually learn and improve his intelligence, such as classification or extraction going forward. I’ll show you what that experience looks like for someone who’s interacting with Kendrick. This is WorkFusion’s WorkSpace, and this is where human workers can teach and interact with the digital workers. It’s a very friendly user interface, and this is just where our workers will surface escalations. In this scenario, Kendrick processed the IDs for both Timothy and Allison that we received via email. He was able to extract the information and he was able to compare it against other documents and an internal system that are related to this onboarding workflow. For example, Allison may have filled out an online form with her name and you can see that though the ID matches, some of the other documentation does not match. This is really where Kendrick can provide quality assurance and double check that there are no typos or inconsistencies across the application. He’s also able to interface automatically with your internal customer systems to again, provide instant feedback about, is this person’s name the same as what I’m expecting?
Let’s dig a little bit further into Kendrick’s ability to extract information. In this view, you can see one of the passports provided with this application. And you can see around the text, it’s a bit faint, but there are little boxes around all the text on the document. So passport type, surname, et cetera. That is the OCR digitizing the text. But then Kendrick identifies the key information. So he’s able to find, for example, that Timothy Aaron is the first name and that Binder is the last name. And this way, instead of just getting a block of digitized text, we actually use machine learning to identify the key points. And this is all captured to be able to transmit into an end system. Just another example here of Allison. Again, even with the passport, even with these images and stamps covering part of the text, Kendrick is able to identify and extract the key information from these. So, in a straight through process scenario, where all the information matches and Kendrick is able to identify all this information and he’s able to check it against an internal system, the user would not actually have to go into this view. But for our demonstration purposes, I just showed a couple escalations that one might see.
Kendrick is also able to learn from this. If, for example, an individual has to make a correction, they can make it in the screen, and WorkFusion would capture that to be able to improve our extraction models in the future.
Lastly, we can view a dashboard on how Kendrick is performing, and this can provide really business-critical KPIs that you can monitor for this entire process. Again, in this scenario, it’s mostly focused on Kendrick’s ID, but also you can combine this with other documents that are being processed as part of onboarding applications. You can, for example, see the breakdown of types of IDs that Kendrick has received, types of other documents that Kendrick has received, and kind of you can see this breakdown over time. You can see in the middle bottom, how much handling time has to happen manually because there were exceptions. It’s a really great tool for you to visualize what is happening in Kendrick’s operations and even his interaction with his human workers.