Name: AtData at Fintech Meetup 2024 – Fraud detection and prevention
Uploaded: 2024-03-25T14:12:50.979Z
Duration: 878.31
Description: At the Fintech Meetup 2024 conference in Las Vegas, FinTech Futures sat down with Brandt Hoffman, Sales Director, Frad Services at AtData, to discuss how financial services firms can strengthen their fraud detection and prevention strategies through email address intelligence.

Atdata_1.mp4

Atdata_1.vtt

ENGLISH

{{00:00:02}}
Hi, I'm Heather Sud with Fintech Futures.

{{00:00:04}}
We are here at Fintech Meetup and
I'm joined by Brandt with AtData.

{{00:00:07}}
Welcome.

{{00:00:09}}
Thank you, Heather, for having me.

{{00:00:10}}
Nice to have you.

{{00:00:12}}
We're talking about fraud today.

{{00:00:13}}
Yes.

{{00:00:14}}
There's a lot to discuss,
but let's start from the top.

{{00:00:16}}
If you'll tell me about yourself and
the background you have in fraud.

{{00:00:19}}
Yeah, absolutely.

{{00:00:21}}
So I'm very fortunate to have been
brought into this space by a wonderful

{{00:00:29}}
peer, and certainly somebody that
I've, leaned on for the past five

{{00:00:34}}
years getting into this space.

{{00:00:36}}
But a gentleman by the name of Scott
Stone who was a founder at chargeback.

{{00:00:41}}
com, he pulled me into this space after I
spent 10 years in architecture and design.

{{00:00:48}}
So I actually fit kind of outside so I had
a, different perspective and lens perhaps.

{{00:00:55}}
But it allowed me to get really, start
from the ground up and build up my

{{00:01:01}}
knowledge set as I started looking at what
payment fraud was, what a chargeback was.

{{00:01:07}}
I didn't even know what
that stuff was in 2020.

{{00:01:10}}
So I joined chargeback.

{{00:01:14}}
com.

{{00:01:14}}
I've always been in a selling capacity
and so I helped them build up what

{{00:01:20}}
they were doing in terms of their
sales processes closing some business.

{{00:01:24}}
And then ultimately we
got purchased by SIFT.

{{00:01:28}}
So for the fraud people in the community,
you'll be very familiar with SIFT.

{{00:01:33}}
Had a great experience there and it
really rooted I want to stay in this

{{00:01:38}}
industry and, just learn as much as I can.

{{00:01:41}}
And now I'm with AtData, which is
very exciting because I have some

{{00:01:45}}
platform experience, but now I'm
getting into more of that okay, what

{{00:01:49}}
does the data do for the fraud industry
specifically, which is very exciting.

{{00:01:54}}
Tell me more about this.

{{00:01:55}}
So with AtData, what is the
fraud data that you're using

{{00:01:58}}
and how is it being used?

{{00:02:00}}
Yeah so AtData is unique in that
we solely look at the email and the

{{00:02:09}}
email intelligence metrics that we can
derive from that single email address.

{{00:02:15}}
So email is nothing new to fraud.

{{00:02:16}}
It's been around for, I
would say, close to 15 years.

{{00:02:21}}
When it really started to gain
traction in the community for fraud,

{{00:02:27}}
investigators, teams that were
looking at risk in a different way.

{{00:02:31}}
And what's unique about what we're doing
with the email is that we've been a, an

{{00:02:38}}
enterprise data company for 23 years.

{{00:02:41}}
So we didn't get our start in fraud.

{{00:02:43}}
We in fact got our start in in the
martech side of the space where

{{00:02:49}}
our founder and CEO Tom Burke.

{{00:02:51}}
Essentially placed a bet the email
would be the, or our digital identity.

{{00:02:58}}
And I think it was a pretty
smart bet to make because we

{{00:03:00}}
use email on a regular basis.

{{00:03:04}}
And what it's allowed us to do is we we,
got our start in one aspect, but then

{{00:03:10}}
now what we're able to do is because
we process so many billions of emails

{{00:03:14}}
a month, we can start to look at what
type of products can we build with this?

{{00:03:18}}
How does that apply to one specific
company or one specific segment

{{00:03:24}}
where we can start to build out
those robust APIs that are needed

{{00:03:27}}
for marketing type services or fraud
services is what we call our product.

{{00:03:35}}
So you're building out the APIs to
enable other FinTechs and others to use

{{00:03:40}}
this data that you've been building.

{{00:03:42}}
Tell me how it works
and what the process is.

{{00:03:44}}
Sure.

{{00:03:44}}
Yeah.

{{00:03:47}}
The email that we're collecting is
coming into us through our, network.

{{00:03:53}}
So again, we've got 23 years
of of data that we're seeing.

{{00:03:57}}
We're processing about 2.1
billion active emails a month.

{{00:04:01}}
That leads to about 6 billion activity
signals within that, within that data set.

{{00:04:09}}
What we do with that is we start to look
at what's going on behind the email.

{{00:04:15}}
How is the identity of
the email being shaped?

{{00:04:18}}
What is the velocity of it that
we're seeing in our network?

{{00:04:21}}
What is the popularity that
we're seeing in our network?

{{00:04:24}}
When was it last seen
when was it first created?

{{00:04:28}}
So we're tracking all of those different
types of data points that provide

{{00:04:33}}
essentially a risk score that a merchant,
a fraud team a FinTech could essentially

{{00:04:42}}
leverage as part of their own workflows.

{{00:04:45}}
So again we are not a fraud platform.

{{00:04:48}}
We are very much an enterprise
data company where we're passing

{{00:04:51}}
you back the data you need.

{{00:04:53}}
So you can either create products or
route them into existing products or

{{00:04:58}}
route them into existing workflows.

{{00:04:59}}
We're proverbially the way to
describe it as we're the premium

{{00:05:05}}
gasoline that makes the engine run.

{{00:05:08}}
Absolutely.

{{00:05:08}}
Then the FinTech or the merchant or
whoever chooses to use the data may

{{00:05:12}}
set their rules, set their boundaries
and build this, Fraud mitigation

{{00:05:16}}
strategy around what they're able to
find in your email addresses and what

{{00:05:20}}
other kinds of data are you sharing?

{{00:05:22}}
So yeah it's funny.

{{00:05:24}}
I get asked this a lot what else are
you, what else are you looking at?

{{00:05:28}}
And our simple response is email,
But we do on the fraud services

{{00:05:35}}
side of the house, we can look
at email IP, full name as well as

{{00:05:41}}
physical address because we've got.

{{00:05:45}}
Emails, we also have been collecting
physical addresses, so we can look at

{{00:05:50}}
like name anomalies physical address
anomalies tied to that specific email

{{00:05:55}}
address, which help us give more
clarity into why we may have scored

{{00:05:59}}
an email low, high, or in the middle.

{{00:06:03}}
We do run on machine learning models.

{{00:06:06}}
Those machine learning models
can be trained specifically

{{00:06:09}}
to your fraud demographic.

{{00:06:12}}
But typically the only, or not typically,
but really the only compulsory field.

{{00:06:17}}
that somebody needs to give
us is the email address.

{{00:06:20}}
The other information helps apply
context, but it's not necessarily

{{00:06:25}}
providing the main result.

{{00:06:28}}
Understood.

{{00:06:30}}
So you mentioned machine learning.

{{00:06:31}}
Yes.

{{00:06:31}}
I want to know more.

{{00:06:33}}
We have talked earlier about
fraud being this area that we are

{{00:06:36}}
all constantly always learning.

{{00:06:37}}
We need to stay on top of
everything that changes every day.

{{00:06:40}}
Absolutely.

{{00:06:41}}
What are you doing at AtData to make
sure the process is continually evolving?

{{00:06:45}}
What have you seen changing?

{{00:06:47}}
Yeah, so I this is, a great question.

{{00:06:51}}
Innovation is paramount in this
industry because fraudsters are

{{00:06:56}}
very adaptive, you could think
that they're always one step ahead.

{{00:07:04}}
And so we have to be we ourselves
have to put ourselves one step ahead.

{{00:07:10}}
We do leverage machine learning models to
help train specific fraud demographics.

{{00:07:17}}
One of the things that we've figured
out over the past few years is fraud

{{00:07:22}}
looks different to every single person.

{{00:07:26}}
And when I say person, any merchant,
any bank, any, whether you're selling

{{00:07:30}}
in Europe, whether you're selling in
Central America, South America, North

{{00:07:33}}
America, fraud is different for everybody.

{{00:07:36}}
So what we have to do is really
apply the context that we can

{{00:07:40}}
see from the email address.

{{00:07:42}}
We use machine learning logic
to help us train those data sets

{{00:07:45}}
to be as effective as possible.

{{00:07:47}}
But something that we're really
looking into now is the use of

{{00:07:52}}
generative AI within our own database.

{{00:07:56}}
So what can we start to extract
that maybe we haven't seen before?

{{00:08:01}}
How can we build out better
identity graphs that make more sense

{{00:08:05}}
and easier for teams to follow?

{{00:08:07}}
Which can be at lightning speed.

{{00:08:10}}
So we deliver, because it's
API based, we're a real time

{{00:08:14}}
fraud prevention solution.

{{00:08:17}}
That's what we do.

{{00:08:18}}
It's in real time, right?

{{00:08:20}}
Again it's paramount within the fraud
space to, you can't really let time pass.

{{00:08:27}}
You really have to be on top
of it to allow yourself to to

{{00:08:32}}
capture that fraud appropriately.

{{00:08:34}}
But also most effectively.

{{00:08:37}}
So specifically in the real time
scenario, that would also help you.

{{00:08:40}}
I imagine if you're seeing a mass
hits across an address or something

{{00:08:44}}
that's been a flag in that sense.

{{00:08:46}}
Yeah.

{{00:08:46}}


{{00:08:47}}
There's, the we, look at
it from both ways, right?

{{00:08:51}}
So we've got velocity, which could
be, maybe we've seen an email

{{00:08:55}}
once over the past six months.

{{00:08:58}}
That's great.

{{00:08:58}}
We're classifying that.

{{00:09:00}}
We're indexing that.

{{00:09:01}}
We're understanding what that
means in our own ecosystem.

{{00:09:04}}
But like you said, if we see something
that's been hit 500 times, and we

{{00:09:09}}
get an email back from somebody
we're going to pass that that's

{{00:09:13}}
going to really affect the risk
score because it's, strange behavior.

{{00:09:18}}
It doesn't fit the mold of Heather or
Brandt going on and shopping online or

{{00:09:24}}
placing an order on whatever it might be.

{{00:09:28}}
So we look at those behavioral traits
as well and index those against

{{00:09:32}}
kind of what we're seeing, which is.

{{00:09:36}}
Not normal behavior.

{{00:09:37}}
So for the banks and the FinTechs
that are listening right now and

{{00:09:42}}
the fraud vendors, if they want to
use your API to have access to this

{{00:09:47}}
data, what does that look like?

{{00:09:49}}
What are some of the
use cases you've seen?

{{00:09:52}}
Yeah, what's really neat about the data
side of the world is that we have really

{{00:09:59}}
an undefined vertical or an undefined
territory that we can work within.

{{00:10:04}}
Really, at any point.

{{00:10:06}}
Along any user journey where you're
taking in an email, whether it be on a

{{00:10:10}}
contact form on the banking side, it could
be if they're creating a new account,

{{00:10:16}}
wherever we see that email, we can take
that in, we can score that, and then

{{00:10:20}}
we can provide that back in real time.

{{00:10:23}}
So it makes it very easy to access it.

{{00:10:25}}
But the other way we look at it is
we're typically going to be a low cost

{{00:10:30}}
solution at the top of a funnel to where
we can start to really realize value.

{{00:10:36}}
Hey.

{{00:10:38}}
If we get a good score that's coming
in, maybe you want to pass that

{{00:10:41}}
person through without any friction.

{{00:10:43}}
Maybe you want to remove their
2FA if they don't need it, right?

{{00:10:49}}
There's a lot more expensive calls
than at data when it comes to APIs.

{{00:10:54}}
And we really fit well in that kind
of, right at the top of the funnel as

{{00:10:58}}
you're acquiring that email address.

{{00:11:02}}
From that regard it's it's just very
simple in terms of integration is

{{00:11:07}}
we're a single line of script or
it's a single endpoint that you call.

{{00:11:12}}
We give you back the
data in a JSON format.

{{00:11:16}}
We're not a platform, we're not a clunky
integration and really we've doubled

{{00:11:21}}
down on the fact that we want to be the
leaders in email address intelligence.

{{00:11:26}}
And so that's where we put
ourselves and pride ourselves.

{{00:11:31}}
And just to that extent, I think it's
probably worth knowing for your community.

{{00:11:36}}
In North America, we recognize about 90,
just shy of 99 percent of all emails.

{{00:11:43}}
And what that means is we've
got a, just a massive amount of

{{00:11:46}}
data that we're already seeing.

{{00:11:48}}
But as we go out into the world, one
of the questions that we get asked

{{00:11:50}}
a lot is are you a global company?

{{00:11:53}}
Absolutely, because we can take in
email from anywhere in the world.

{{00:11:57}}
So we have different rates and different,
identity rates in, other countries.

{{00:12:04}}
But we're very much a, global
entity because email is

{{00:12:07}}
very much a global identity.

{{00:12:09}}
Absolutely.

{{00:12:11}}
Oh, the, in addition to the fraud, the
ability to ease that friction is a big

{{00:12:14}}
point that I hadn't thought of yet.

{{00:12:16}}
Yes.

{{00:12:16}}
We've all been in that case as a consumer
where a big tech will say I recognize you.

{{00:12:21}}
You may proceed.

{{00:12:22}}
And it's a breath of relief.

{{00:12:23}}
I think that it's interesting to have
both sides of that coin as we continue

{{00:12:27}}
to balance fraud and consumer experience.

{{00:12:30}}
Yeah, definitely.

{{00:12:32}}
I, fully agree with that.

{{00:12:33}}
What is it from what you've seen
in your career and at AtData

{{00:12:38}}
that you feel makes for a most
effective fraud prevention strategy?

{{00:12:42}}
I love this question because from somebody
coming from outside of the industry,

{{00:12:48}}
and now immersing myself into the the,
fraud community, the risk the risk

{{00:12:55}}
managers, your fraud prevention managers,
I think what's been most telling about

{{00:12:59}}
it is it's not a one size fits all.

{{00:13:02}}
solution in any regard.

{{00:13:06}}
I don't run fraud teams but I listened
to a lot of fraud teams and I try to

{{00:13:11}}
understand what their pain points are
and how they're affected by fraud.

{{00:13:16}}
And I'll say a, healthy fraud
prevention strategy is you've

{{00:13:20}}
got to have a lot of data.

{{00:13:21}}
You've got to have it coming from a lot of
different sources that are reputable, but

{{00:13:26}}
you also have to be able to orchestrate.

{{00:13:28}}
that in a proper way.

{{00:13:29}}
So what are the workflows
that you're writing?

{{00:13:32}}
What are the rules that
you're maybe leaning on?

{{00:13:34}}
How can you be very reactionary or
very reactive to a massive bot attack

{{00:13:41}}
or somebody spinning up hundreds
of thousands of emails at a time

{{00:13:44}}
to try to, for credential stuffing
and things along those lines.

{{00:13:48}}
So I would say from my perspective,
always just try to be learning,

{{00:13:57}}
see what's new in the market.

{{00:13:59}}
What is innovating in the space?

{{00:14:01}}
Everybody's talking about AI.

{{00:14:03}}
There's definitely a place for it.

{{00:14:06}}
We work with some companies that are
leveraging AI to respond and adapt the

{{00:14:10}}
rules that they're writing internally,
which is I think very efficient and smart.

{{00:14:15}}
And there's going to be a lot
more of that to come, but I think

{{00:14:18}}
for the big businesses out there,
again, always be exploring and

{{00:14:22}}
don't get stuck in the status quo.

{{00:14:25}}
Because this is an industry
that is really moving quickly.

{{00:14:27}}
Grant, it's been an honor talking to you.

{{00:14:29}}
Thanks for all the information
and keeping us informed.

{{00:14:32}}
Yeah, this has been great.

{{00:14:33}}
Enjoy the rest of your show.

{{00:14:34}}
Thank you so much.

Hi, I'm Heather Sud with Fintech Futures.

We are here at Fintech Meetup and
I'm joined by Brandt with AtData.

There's a lot to discuss,
but let's start from the top.

If you'll tell me about yourself and
the background you have in fraud.

So I'm very fortunate to have been
brought into this space by a wonderful

peer, and certainly somebody that
I've, leaned on for the past five

But a gentleman by the name of Scott
Stone who was a founder at chargeback.

com, he pulled me into this space after I
spent 10 years in architecture and design.

So I actually fit kind of outside so I had
a, different perspective and lens perhaps.

But it allowed me to get really, start
from the ground up and build up my

knowledge set as I started looking at what
payment fraud was, what a chargeback was.

I didn't even know what
that stuff was in 2020.

I've always been in a selling capacity
and so I helped them build up what

they were doing in terms of their
sales processes closing some business.

And then ultimately we
got purchased by SIFT.

So for the fraud people in the community,
you'll be very familiar with SIFT.

Had a great experience there and it
really rooted I want to stay in this

industry and, just learn as much as I can.

And now I'm with AtData, which is
very exciting because I have some

platform experience, but now I'm
getting into more of that okay, what

does the data do for the fraud industry
specifically, which is very exciting.

So with AtData, what is the
fraud data that you're using

Yeah so AtData is unique in that
we solely look at the email and the

email intelligence metrics that we can
derive from that single email address.

It's been around for, I
would say, close to 15 years.

When it really started to gain
traction in the community for fraud,

investigators, teams that were
looking at risk in a different way.

And what's unique about what we're doing
with the email is that we've been a, an

We in fact got our start in in the
martech side of the space where

Essentially placed a bet the email
would be the, or our digital identity.

And I think it was a pretty
smart bet to make because we

And what it's allowed us to do is we we,
got our start in one aspect, but then

now what we're able to do is because
we process so many billions of emails

a month, we can start to look at what
type of products can we build with this?

How does that apply to one specific
company or one specific segment

where we can start to build out
those robust APIs that are needed

for marketing type services or fraud
services is what we call our product.

So you're building out the APIs to
enable other FinTechs and others to use

Tell me how it works
and what the process is.

The email that we're collecting is
coming into us through our, network.

So again, we've got 23 years
of of data that we're seeing.

We're processing about 2.1
billion active emails a month.

That leads to about 6 billion activity
signals within that, within that data set.

What we do with that is we start to look
at what's going on behind the email.

How is the identity of
the email being shaped?

What is the velocity of it that
we're seeing in our network?

What is the popularity that
we're seeing in our network?

When was it last seen
when was it first created?

So we're tracking all of those different
types of data points that provide

essentially a risk score that a merchant,
a fraud team a FinTech could essentially

We are very much an enterprise
data company where we're passing

So you can either create products or
route them into existing products or

We're proverbially the way to
describe it as we're the premium

Then the FinTech or the merchant or
whoever chooses to use the data may

set their rules, set their boundaries
and build this, Fraud mitigation

strategy around what they're able to
find in your email addresses and what

I get asked this a lot what else are
you, what else are you looking at?

And our simple response is email,
But we do on the fraud services

side of the house, we can look
at email IP, full name as well as

Emails, we also have been collecting
physical addresses, so we can look at

like name anomalies physical address
anomalies tied to that specific email

address, which help us give more
clarity into why we may have scored

Those machine learning models
can be trained specifically

But typically the only, or not typically,
but really the only compulsory field.

that somebody needs to give
us is the email address.

The other information helps apply
context, but it's not necessarily

FinTech Futures

Expert Insight

At the Fintech Meetup 2024 conference in Las Vegas, FinTech Futures sat down with Brandt Hoffman, Sales Director, Frad Services at AtData, to discuss how financial services firms can strengthen their fraud detection and prevention strategies through email address intelligence.

