Fall.SPINT.012.mp4

Back

Or filter by your interest

Browse all our content

Finovate

Unlock all the content with only {{price}}

Fall.SPINT.012 David Snitkof.vtt

English

{{00:00:05}}
David Penn: Hi, I'm David Penn,
Research Analyst with Finovate.

{{00:00:07}}
We're  here at FinovateFall
2023 at Times Square, New York.

{{00:00:10}}
Joining me is David Snitkof.

{{00:00:12}}
He is the Senior Vice President
of Growth with Ocrolus.

{{00:00:14}}
Thank you for joining us today.

{{00:00:15}}
Thank you for having me.

{{00:00:16}}
Fantastic.

{{00:00:17}}
I thought before we began, maybe you
could tell us just a little bit about

{{00:00:19}}
the work that you do at Ocrolus.

{{00:00:20}}
David Snitkof: Sure.

{{00:00:21}}
So Ocrolus is an AI driven document
automation platform that powers

{{00:00:25}}
the online lending ecosystem.

{{00:00:26}}
So we work with consumer lenders,
small business lenders, mortgage

{{00:00:29}}
lenders, and others to really automate
their document driven workflows.

{{00:00:33}}
David Penn: Okay excellent, Now we're
talking quite a bit about AI and machine

{{00:00:37}}
learning here at, Finovate, just like
we are in the entire tech industry.

{{00:00:40}}
What would you say are some of the
biggest misconceptions that lenders

{{00:00:42}}
have when it comes to bringing AI and
machine learning, these sorts of enabling

{{00:00:46}}
technologies into their workflows?

{{00:00:48}}
David Snitkof: The interesting
thing is in lending, people have

{{00:00:50}}
been using machine learning in
one form or another for decades.

{{00:00:54}}
In some way, a lot of stuff's
been here for a long time.

{{00:00:57}}
In another way, there are some new
capabilities, but I think one of

{{00:00:59}}
the misconceptions is that there
has to be a lot of magic that

{{00:01:03}}
there has to be something that's
totally new or totally outside the

{{00:01:07}}
reach of a lot of institutions.

{{00:01:09}}
So every single bank, every single
credit union, even every non bank

{{00:01:12}}
lender is thinking about how can
they use AI to their advantage or

{{00:01:15}}
will they be missing something?

{{00:01:17}}
And I think what a lot of
people don't realize is you

{{00:01:20}}
can start using AI tools today.

{{00:01:22}}
You don't need to have AI
experts in your own company.

{{00:01:25}}
Although it helps, you don't need to
do incredible innovation internally.

{{00:01:29}}
You can partner with companies
that have already been building AI.

{{00:01:32}}
Ocrolus has been building document
automation AI for years and years now.

{{00:01:37}}
We have processed literally
hundreds of millions of pages

{{00:01:40}}
of verified document data.

{{00:01:42}}
So if you want to get started with
an AI based lending solution, you

{{00:01:46}}
don't have to build it all in house.

{{00:01:47}}
Sometimes you can get even
farther by partnering.

{{00:01:49}}
David Penn: You mentioned a couple
of major challenges certainly within

{{00:01:53}}
this space and beyond the idea of
having to take a big bite to start.

{{00:01:57}}
This idea that "oh my gosh, I'm going
to have to do this all on my own."

{{00:02:01}}
I'm wondering, keeping those in mind.

{{00:02:02}}
Are there particular challenges that
some of the folks that you've worked

{{00:02:05}}
with have had to overcome in 23?

{{00:02:07}}
And how are some of the ways that
AI and machine learning have helped

{{00:02:09}}
them to overcome those challenges?

{{00:02:10}}
David Snitkof: So 2023 has been a
very challenging year for lenders.

{{00:02:13}}
Higher cost capital, increased
delinquency, fraud has been up

{{00:02:17}}
in multiple different channels.

{{00:02:19}}
And if you really think about
what is the business of a lender?

{{00:02:21}}
Now, there's many different types
of lenders, many different lending

{{00:02:23}}
products, but ultimately, business
model wise, they're all the same.

{{00:02:27}}
They make money on interest and fees.

{{00:02:29}}
They spend money on operational
expenses and cost of capital and

{{00:02:33}}
cost of acquisition, and they lose
money on credit and fraud losses.

{{00:02:36}}
So you have to affect one of
those levers to a really matter.

{{00:02:39}}
And because of increased cost of
capital and increased delinquency, what

{{00:02:42}}
we've seen a lot of lenders focusing
on this year is operational expense.

{{00:02:47}}
How can they have a much
more efficient operation?

{{00:02:51}}
How can they use less human
work in doing their processing?

{{00:02:54}}
How can they make decisions
faster, more efficiently?

{{00:02:57}}
So what the best lenders coming to us
have been able to do is say, if I have a

{{00:03:01}}
document driven lending process, how can
we make sure that we only need to use our

{{00:03:07}}
own humans when it's absolutely necessary,
when it's an exception workflow?

{{00:03:11}}
So what we've been seeing people doing
a lot is automating a lending process,

{{00:03:16}}
digitizing a lending process, so that they
can do it in a very cost effective way.

{{00:03:21}}
Only using human intervention
where it's absolutely necessary.

{{00:03:25}}
David Penn: Now you mentioned
things like automation in terms

{{00:03:27}}
of these processes and so forth.

{{00:03:28}}
I'm thinking in terms of
enhancing the customer experience.

{{00:03:31}}
How do some of these technologies
work their way through so

{{00:03:33}}
that the customer feels the
improvements that are being made?

{{00:03:36}}
David Snitkof: So I think the
bar for customer experience

{{00:03:38}}
in 2023 is higher than ever.

{{00:03:40}}
Throughout people's entire lives,
not even in financial services.

{{00:03:44}}
They are used to services where they
get service very quickly where it's all

{{00:03:48}}
digital, where it's all mobile, where it's
very easily acceptable and the switching

{{00:03:52}}
costs are very, low, especially in the U.

{{00:03:56}}
S.

{{00:03:56}}
Banking system in fintech.

{{00:03:58}}
It's not very hard to switch between
providers, so people will demand a

{{00:04:02}}
customer experience where they get
quick decisions in a transparent way

{{00:04:07}}
and where they get the product that's
right for them and on the right terms.

{{00:04:09}}
And so where I think AI comes into being
is where people can get recommendations

{{00:04:15}}
for the right product at the right time,
but really more importantly, where a

{{00:04:18}}
lender, any financial services provider
is able to deliver a process that is

{{00:04:23}}
fully automated as much as possible.

{{00:04:26}}
So they get an answer back
quickly, even if the answer is no.

{{00:04:31}}
Having that answer quickly allows
the customer to shop around and

{{00:04:34}}
that ultimately increases economic
opportunity for borrowers.

{{00:04:38}}
David Penn: That makes a lot of sense.

{{00:04:38}}
You mentioned some of the different
levers that institutions have to do,

{{00:04:41}}
lenders rather, have to deal with.

{{00:04:43}}
One of the ones obviously
is managing risk.

{{00:04:45}}
And one of the things that I know we've
been hearing some folks talk about when

{{00:04:47}}
it comes to AI is the ability of that
technology to help them manage risk.

{{00:04:51}}
Can you talk about that a little bit?

{{00:04:53}}
David Snitkof: Absolutely.

{{00:04:53}}
Lending is a funny business because
you give the customer your money and

{{00:04:57}}
you hope to get paid back over time.

{{00:04:58}}
It's a business of asymmetric risk
and decision making under uncertainty.

{{00:05:04}}
Asymmetric meaning you lose a
lot more money on a bad borrower

{{00:05:07}}
than you make on a good borrower.

{{00:05:08}}
And uncertainty meaning there's not,
it's not possible to know all the

{{00:05:13}}
information at the time of underwriting
that you would want to know to completely

{{00:05:16}}
forecast what a borrower is going to do.

{{00:05:19}}
So we see a lot of lenders adopting
cash flow underwriting to really try to

{{00:05:23}}
understand in a nuanced way What are the
unique financial dynamics of a borrower?

{{00:05:29}}
We see a lot of customers using
us to process their bank data.

{{00:05:33}}
When we look at bank statements or
digital bank data, we tag all the

{{00:05:37}}
transactions to try to see, is a
transaction revenue, expense, proceeds

{{00:05:42}}
from a lender, payments to a lender,
non sufficient funds, about 50 other

{{00:05:45}}
things, to really help understand, what
are the unique financial dynamics of

{{00:05:50}}
this borrower, so a lender is not making
a one size fits all credit assessment.

{{00:05:55}}
They're really trying to understand.

{{00:05:56}}
The revenue, expenses, the debt capacity,
the ability to pay of each borrower

{{00:06:01}}
individually so they can make the
right credit decision at the margin.

{{00:06:05}}
David Penn: I'm thinking
maybe a way of wrapping up.

{{00:06:07}}
You mentioned the analysis of bank data.

{{00:06:08}}
And I'm wondering if that's one of the
ways that some of the lenders who are

{{00:06:11}}
really starting to catch on to how to
really maximize some of the technology,

{{00:06:14}}
some of the things that they're doing.

{{00:06:16}}
One, is that true?

{{00:06:16}}
And if that is true, are there other
examples of that kind of thing from

{{00:06:19}}
some of the lenders that you really
think have figured it out perhaps?

{{00:06:22}}
David Snitkof: Yeah, so it's a great
question because yesterday the demo we

{{00:06:24}}
gave at Finovate was called...

{{00:06:26}}
David Snitkof: The best SMB lender
ever, which was the idea of what are

{{00:06:29}}
the absolute best lenders doing today?

{{00:06:32}}
How are they using the full suite of
capabilities that we offer to have

{{00:06:35}}
the best possible lending experience?

{{00:06:37}}
And what are they doing?

{{00:06:39}}
They're using tools to give borrowers
the choice to provide their data

{{00:06:42}}
either in digital or document form
to make sure they have no drop

{{00:06:46}}
off in their acquisition funnel.

{{00:06:48}}
They're using fraud prevention tools
like what we offer to make sure that if

{{00:06:52}}
there's any tampering with documents,
which does happen very frequently,

{{00:06:56}}
That they're able to identify fraud
before they lose money, before a

{{00:06:59}}
dollar of capital goes out the door.

{{00:07:01}}
They're using deep transaction
enrichment to understand what

{{00:07:06}}
bank transactions represent.

{{00:07:07}}
Getting data off the documents,
really understanding the unique

{{00:07:10}}
financial dynamics of the borrower.

{{00:07:12}}
And then using all this data
in concert to develop cash

{{00:07:16}}
flow aware credit risk models.

{{00:07:18}}
So that they can make, as we just talked
about, the best possible credit decision.

{{00:07:23}}
At the margin and then all throughout
what they really want to do is offer

{{00:07:27}}
an experience where a borrower can
get the right financial product

{{00:07:30}}
at the right time at the right
price and with the right structure.

{{00:07:34}}
And ultimately, I don't think borrowers
care if that's coming from a bank, a

{{00:07:39}}
credit union, a non bank FinTech company.

{{00:07:42}}
They just want to get the right products
on the right terms at the right time.

{{00:07:44}}
David Penn: Yeah, that
makes a lot of sense.

{{00:07:45}}
It's a really interesting time for
lending, really challenging time for

{{00:07:48}}
lending just an economic standpoint.

{{00:07:49}}
But then thinking of all these
technologies that are coming to there,

{{00:07:52}}
it's really great to have companies
like Ocrolus that are out there helping

{{00:07:54}}
bridge that gap between some of the
challenges in the economic environment

{{00:07:57}}
and some of the opportunities in
the technological environment.

{{00:07:59}}
David Snitkof: Yeah, absolutely.

{{00:08:00}}
And we work with about 400, 450
lenders today across Consumers,

{{00:08:05}}
small business mortgage and more.

{{00:08:06}}
One of the most fun parts of
my job is just getting to talk

{{00:08:09}}
to all these companies and some
have been lenders for decades.

{{00:08:12}}
Some are huge banks and some are
literally startups who started lending

{{00:08:15}}
six months ago and it's just great to
help them think about how do they just

{{00:08:19}}
more more effectively automate and
digitize their entire lending workflow.

{{00:08:24}}
David Penn: Fascinating stuff.

{{00:08:25}}
Thank you very much for spending
some time with us today.

{{00:08:26}}
Thank you.

{{00:08:27}}
All right.

{{00:08:27}}
And thank you.

{{00:08:29}}
David Snitkof: Thank you.

David Penn: Hi, I'm David Penn,
Research Analyst with Finovate.

We're  here at FinovateFall
2023 at Times Square, New York.

He is the Senior Vice President
of Growth with Ocrolus.

I thought before we began, maybe you
could tell us just a little bit about

So Ocrolus is an AI driven document
automation platform that powers

So we work with consumer lenders,
small business lenders, mortgage

lenders, and others to really automate
their document driven workflows.

David Penn: Okay excellent, Now we're
talking quite a bit about AI and machine

learning here at, Finovate, just like
we are in the entire tech industry.

What would you say are some of the
biggest misconceptions that lenders

have when it comes to bringing AI and
machine learning, these sorts of enabling

David Snitkof: The interesting
thing is in lending, people have

been using machine learning in
one form or another for decades.

In some way, a lot of stuff's
been here for a long time.

In another way, there are some new
capabilities, but I think one of

the misconceptions is that there
has to be a lot of magic that

there has to be something that's
totally new or totally outside the

So every single bank, every single
credit union, even every non bank

lender is thinking about how can
they use AI to their advantage or

And I think what a lot of
people don't realize is you

You don't need to have AI
experts in your own company.

Although it helps, you don't need to
do incredible innovation internally.

You can partner with companies
that have already been building AI.

Ocrolus has been building document
automation AI for years and years now.

We have processed literally
hundreds of millions of pages

So if you want to get started with
an AI based lending solution, you

Sometimes you can get even
farther by partnering.

David Penn: You mentioned a couple
of major challenges certainly within

this space and beyond the idea of
having to take a big bite to start.

This idea that "oh my gosh, I'm going
to have to do this all on my own."

Are there particular challenges that
some of the folks that you've worked

And how are some of the ways that
AI and machine learning have helped

David Snitkof: So 2023 has been a
very challenging year for lenders.

Higher cost capital, increased
delinquency, fraud has been up

And if you really think about
what is the business of a lender?

Now, there's many different types
of lenders, many different lending

products, but ultimately, business
model wise, they're all the same.

They spend money on operational
expenses and cost of capital and

cost of acquisition, and they lose
money on credit and fraud losses.

So you have to affect one of
those levers to a really matter.

And because of increased cost of
capital and increased delinquency, what

we've seen a lot of lenders focusing
on this year is operational expense.

How can they have a much
more efficient operation?

How can they use less human
work in doing their processing?

How can they make decisions
faster, more efficiently?

So what the best lenders coming to us
have been able to do is say, if I have a

document driven lending process, how can
we make sure that we only need to use our

own humans when it's absolutely necessary,
when it's an exception workflow?

So what we've been seeing people doing
a lot is automating a lending process,

digitizing a lending process, so that they
can do it in a very cost effective way.

Only using human intervention
where it's absolutely necessary.

David Penn: Now you mentioned
things like automation in terms

I'm thinking in terms of
enhancing the customer experience.

How do some of these technologies
work their way through so

that the customer feels the
improvements that are being made?

David Snitkof: So I think the
bar for customer experience

Throughout people's entire lives,
not even in financial services.

They are used to services where they
get service very quickly where it's all

digital, where it's all mobile, where it's
very easily acceptable and the switching

costs are very, low, especially in the U.

It's not very hard to switch between
providers, so people will demand a

customer experience where they get
quick decisions in a transparent way

and where they get the product that's
right for them and on the right terms.

And so where I think AI comes into being
is where people can get recommendations

for the right product at the right time,
but really more importantly, where a

lender, any financial services provider
is able to deliver a process that is

So they get an answer back
quickly, even if the answer is no.

Having that answer quickly allows
the customer to shop around and

that ultimately increases economic
opportunity for borrowers.

You mentioned some of the different
levers that institutions have to do,

One of the ones obviously
is managing risk.

And one of the things that I know we've
been hearing some folks talk about when

it comes to AI is the ability of that
technology to help them manage risk.

David Snitkof, SVP of Growth, Ocrolus, talks about the role automation plays in lending and the new innovations that lenders are taking on.