Скачать или смотреть Fraud and Financial Crime Update (New Account Fraud & Digital Account Opening)

Thanks for watching our fraud and financial crime update! For more around new account fraud, digital account opening, application fraud, data analysis for identity confidence, and fraud risk management, please subscribe:    / @feedzai  

To learn about Feedzai's New Account Fraud solution, click here: https://feedzai.com/solutions/new-acc...

CHAPTERS:
0:00 A Flood of First-time Customers
1:09 Entity-based Knowledge vs. Changes in Behavior
1:33 Determining the Presence of Data Drift
2:17 Detecting Guided Behavior

TRANSCRIPT:
Continuing the discussion from last week on identity and entity, let's see how it applies to new account fraud.

My name is João and this is Feedzai’s Fraud and Financial Crime News Weekly Update.

It's easy to think about behavior, entities, and identities, and all of that in the context of, “Yeah, yeah, I know this customer and I have actually six months, one year, or ten years of history on their behavior.” But what if it's the first time this person’s actually logging into their account? What if it's the first time they're accessing your mobile banking website? That makes things a lot more complex.

You have these economies, for example, Brazil, where a lot of new financial services, way leaner than before, are appearing. You have Pix on the one hand, Open Finance on the other hand. The scale of Brazil is very large, we’re talking about hundreds of millions of people trying to open a new account. It's easy, so convenient, more agile than it used to be. At this scale, new account fraud really becomes a problem.

And here we turn from this idea of the overall behavior to how much data do I need to be confident that I'm understanding this identity.

First time a person accesses some website, what do I know about them? Not that much. Maybe the time of day is relevant, some geographic information, the type of device. But these are like entity-based knowledge.

But then we start seeing how the person interacts. Click here, click there, the way the mouse moves. And if it looks suspicious, or if we just don't have enough confidence on making a decision, then what we could do is introduce more steps.

So a person selects something, fills in some form - it's actually quite super fast. Presses ‘next’, what do we do? We can either continue and assume the risk or we could introduce a new step requesting more information. And it's not exactly about the information itself that you are requesting, but will the behavior change?

Changes in behavior might happen when the person is being guided. You’re on your phone, you don't know, but you're probably getting scammed. And then something unexpected for the criminal happens. So they know the form, they know what to expect. But then something changes and that will introduce some noise in their behavior. It's no longer a systematic thing. Whereas a legitimate person will just probably fill in every page in a normal way as expected.

Specifically speaking, how can we think about detecting that someone is being guided? That's actually an interesting topic and it's all about what things do we do when being told to do something versus when we do it on our own?

With mobile devices this one thing could happen, which is, you're on the phone and you need to speak with a person. So there are going to be pauses in the way you fill in a form, for example. There's also going to be movements with your device that are very characteristic of this. You are listening to someone, then you put the device down.

But these are all the different signals that one could use to understand this behavior.

Let us know in the comments below how your institution has been looking into new account fraud.

Thanks for watching. This has been your Feedzai Fraud and Financial Crime News Weekly Update.

It's not a cat analogy, but maybe I think I can work on that.