Evelyn

Transcript

Today I’m going to show you Evelyn. What makes Evelyn quite unique is that she has two skills that she’s really, really good at. So she’s an entity sanction screening analyst. And as part of that, she does name sanction screening alert review, and she does adverse media monitoring. So the first thing I’m going to show you is how she works as part of the name sanction screening alert review team.

Just to set expectations, this is how she integrates with LexisNexis, but she can also integrate with many other screening systems. But this would be what an analyst would come in and see pre Evelyn. They’d come into 136 alerts. They’d have no idea which one to look at. They’d have to go row by row and make a decision on each of them. Sucks up a load of time. In all likelihood it will take days. They won’t get through it in one day, and they’ll have very little capacity for anything else.

With Evelyn, it goes from 136 alerts to 22 alerts. So anywhere from an 80% reduction in the amount of alerts that your team has to look at. So what’s the benefit there? They can come and they understand that Evelyn has done all of this work for us. We now know that, okay, these 22 alerts here, they really require my attention and they require my expertise to work through. 

Evelyn and WorkFusion, we are really transparent with the work we do. Let’s go check what Evelyn can do. We would click here on Query. We would click on Record State, and we want to check all the work that Evelyn has dispositioned. We click on False Positive. The record is changed by WorkFusion. We go down here and we click on Submit Query, and it brings us to all the work that Evelyn has done on behalf of your team. So we have a Watchlist entity here, David Hernandez, or… Excuse me, we have a counterparty for the bank, David Hernandez. Hits on the Watchlist entity, David Hernandez. So pre-world, your team would’ve had to come in here. They would’ve had to look at David Hernandez versus David Hernandez. Go out, scour the internet, scour existing databases, scour external databases to really get the information they can to make a disposition. Evelyn does all this in the background. She’s really consistent. She does it 24/7, and she’s really capable. So we want to understand what she’s done. We can come in here to Record History and you can see the work that Evelyn has done, because we have a clear audit history, a clear traceable record of what she’s done, how she’s done it. And on top of that, Evelyn lets you know the decision that she’s come to. So the Record Status: False Positive. And with that she explains how she came to a false positive decision. So this hit is a False Positive. Screened Name matches hit name. Records have the same record type. They’re both individuals. This could also be individual versus entity, individual versus vessel or entity versus vessel. The genders match male. So she’s taking all of these factors into account, but there’s also dispositioning factors. The dates of birth are too far apart. The locations do not match. And she consumes all of this information, puts it through her machine learning models and then outputs that it’s false positive.

What’s the benefit here? This is just one case that Evelyn’s done, but she did this for circa 90 cases that your team wouldn’t have to do it. So 80%, she’ll auto-do this, she can auto-generate all of these in the background. And she’s really skilled at it. She’s really accurate with the close to 0% misclassification rate.

That’s what Evelyn looks like in a system like LexisNexis. To understand how she made those decisions, it’s useful to see what goes on underneath the hood. So again, we’re looking at David Hernandez over here on the left hand side versus Dave Herman. How did Evelyn make the decision? She’s comparing David Hernandez versus Dave Herman. Beaumont, Texas to Syria. And again, it’s really clear what she does. She’s really open with the team about how she makes a decision. There’s 96% confidence that this has been evaluated as a false positive. And how she did it? The input name, David Hernandez, does not match the hit name, Dave Herman. The input location does not match the hit location, Syria. And these auto-generated narratives, these auto-generated dispositions, Evelyn comes with all of these out of the box day one, but this isn’t the limit of what Evelyn can do. She can also integrate with third party data models. And we’re going to see that now.

Here, she goes above and beyond. And she actually goes out and enriches the data to use existing third party integrators, like LexisNexis so that it helps her make a decision. And she does all this in the background. So it’s not just a case that she’s comparing the information we have on file versus the watch list entity, she’s actually going out to LexisNexis, pulling that information, and what that information here is, Pueblo, Colorado USA, and uploading that to her decision model and then making decision based on that. So she’s making a more complete decision straight off the bat, out of the box. And we can see it here. Again, she’s really honest in what she did. So the input name, Maria Cristina Ferrero is a partial match to hit name, Cristina Ferri. Input location was missing. A search in LexisNexis SmartLinx yielded an address in Pueblo, Colorado. The input location does not match the hit location. So she actually brought that information from Colorado and used it as part of her decision matrix.

Evelyn can go to things that are beyond LexisNexis. She can also go to Google. So you can see here that Evelyn went out to Google and then she hyperlinked the information she had and put it in there as part of her narrative. So input name is a partial match to hit name, Guz Centro Commercial and Alex Guz Company. But the input location is a partial match to the hit location. A location search identified the location for the hit name and the input location does not match the hit location. So again, what did she do? She was given information by the bank, but that information wasn’t enough for Evelyn to make a decision. So she wanted to go out, she went out to Google, she pulled down information and then she was able to make a decision based off the work that she did.

There’s going to be some instances, and there are a lot of instances, where you do not want Evelyn to make a decision. She’s going to ask your team for help and she’s going to send it to them for a review. This is a real benefit that Evelyn has is that, it’s not just handed to your team. They have to start from level zero, but Evelyn will let them know why she’s sending it to them for a review and why she thinks this may be a high risk case or a high risk alert. So you can see here, this hit has been evaluated as “needs more information.” It needs more investigation. And why is that? It’s because the names have at least the partial match, Adrian Rawlins and Adrian Ralins. So Evelyn… That’s not something that you would want Evelyn to make a decision on, but she’s highlighting to your team why she is sending to you. So they don’t have to start from level zero. They don’t have a cold start. When this case comes to them, they have a head start in where they would’ve been before.

Evelyn, as I said, she’s really transparent and she’s really honest in what she does. As part of Evelyn and when you bring Evelyn into your team, you get prebuilt analytics with her. There are two major benefits of having an analytics view like this. It allows management and leaders to quickly understand how Evelyn is doing, what’s being automated, what she’s sending for false positive, and also what she’s escalating to your team for a review. It gives your leaders and it gives the operation leaders a quick snapshot of what Evelyn is doing each week. The same way you have a one-to-one with your team, this is your version of a one-to-one with Evelyn. With that, it gives you actionable intelligence. So you can understand that, okay, if the false positive rate is 83%, she’s doing really, really well. But if it’s a bit lower, you might want to tweak things a small bit and work with Evelyn to improve. 

This brings us to something about Evelyn. She’s really curious and she’s continuously learning and she learns on the job every day. She takes feedback from the system. Every decision Evelyn makes is fed back into the model and she continuously learns as she goes. The same way that your team learns day in, day out, Evelyn learns day in, day out. But the major benefit is she doesn’t forget. She’s always improving and she’s always taking information. And with that, she’s not just learning from your team. Evelyn can learn from the WorkFusion Network. So everywhere throughout the world, where Evelyn is working, it gets fed back into her. And she’s learning from different customer environments all the time. So that’s name sanction screening alert review.