Name: Riding the second wave of AI transformation with ING
Uploaded: 2024-07-01T08:58:34.732Z
Duration: 742.8
Description: Touching base with FinTech Futures at Money20/20 Europe, Bahadir Yilmaz, chief analytics officer at ING, discusses how the Dutch banking giant is executing its AI strategy, with insights into how the new face of innovation is being freshly cultivated within the bounds of fast-emerging regulatory safeguards.

Bahadir Yilmaz_ING_1.mp4

Bahadir Yilmaz_ING.vtt

ENGLISH

{{00:00:08}}
Hello, everybody, and
welcome to FinTech Futures.

{{00:00:11}}
We are here at day two of
Money 2020 in Amsterdam.

{{00:00:16}}
And right now I'm very happy to be
joined by Bahadir Yilmaz of ING.

{{00:00:22}}
Bahadir, how are you?

{{00:00:23}}
I'm doing well.

{{00:00:24}}
Thank you.

{{00:00:25}}


{{00:00:25}}
I'm very good.

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

{{00:00:26}}
I'm very happy to be here.

{{00:00:28}}
We've got, the the all the activity
around us, all these people that

{{00:00:33}}
I hope the camera is picking up.

{{00:00:36}}
But it's really fantastic
to be here with you.

{{00:00:39}}
We've caught you at a really interesting
moment because you are on the cusp of

{{00:00:44}}
going on stage this afternoon, right
over there on the big stage at the end

{{00:00:47}}
of the hall, which is really exciting.

{{00:00:49}}
Can you, before we, go into what that
session will entail and more about

{{00:00:56}}
what ING's plans are with technology
and more specifically AI, maybe you

{{00:01:00}}
can give our viewers just a little
bit of a snapshot of who you are

{{00:01:03}}
and what your position is, please?

{{00:01:04}}
Yeah, of course.

{{00:01:05}}
This is Bahadir Yilmaz, and I'm
the Chief Analytics Officer at ING.

{{00:01:08}}
I've been working at ING for the
last 30 years, and since last

{{00:01:12}}
year, I'm holding this position.

{{00:01:13}}
I'm leading all the AI development
at the bank and like we and my team

{{00:01:18}}
are working on digitalizing banking
processes using AI and machine learning.

{{00:01:22}}
That's fantastic.

{{00:01:24}}
Yes, on that note, AI and machine
learning is two very leading topics

{{00:01:28}}
here at the conference this week.

{{00:01:30}}
It's very difficult to get away from it.

{{00:01:32}}
And I believe that is one of the
foundational stones of your session later.

{{00:01:38}}
Can you tell us a little bit ahead of
time what that session is going to entail?

{{00:01:43}}
Yeah, later today I will join a
panel discussion on the digital

{{00:01:47}}
transformation trends and like the
enabling technologies related to that.

{{00:01:50}}
And of course like my focus will be
on generative AI and like I'm going

{{00:01:54}}
to talk about the impact of this new
technology in banking transformation.

{{00:01:58}}
Fantastic.

{{00:01:59}}
Very exciting, and I hope to join you
there in the audience, watching you, if

{{00:02:05}}
my schedule allows, I really do hope.

{{00:02:08}}
But that's a very interesting
question, and let's build on that.

{{00:02:11}}
ING, obviously, is a very
well established bank.

{{00:02:15}}
It's got a lot of processes,
a lot of technology at play.

{{00:02:19}}
AI, although nothing new, I think
its innovations and capabilities

{{00:02:24}}
are really being harnessed.

{{00:02:26}}
Now, in a way, to achieve new
heights, which is really exciting.

{{00:02:30}}
How is ING's approach to using AI?

{{00:02:33}}
What is that approach like?

{{00:02:35}}
I think, we are right now seeing the
second wave of AI transformation.

{{00:02:41}}
The first wave started in mid
2010s, with the introduction of

{{00:02:45}}
big data and machine learning.

{{00:02:47}}
And some Fintechs, some neobanks,
but also ING grasped the power of it.

{{00:02:53}}
And so we built all of our processes,
organizations, platforms to be able

{{00:02:57}}
to cater for that like technology and
also using that in our transformation.

{{00:03:02}}
But the real thing happened a year and
a half ago, like with the introduction

{{00:03:06}}
of ChetchiPT, everybody started to feel
it, everybody started to touch to it.

{{00:03:10}}
So like that, introduced a new game,
like where it is much more tangible

{{00:03:15}}
and it is much more impactful.

{{00:03:17}}
So we think we are ready for that
transformation with our people

{{00:03:20}}
capabilities, with our technological
capabilities, and it's going to be really

{{00:03:23}}
transformational in the upcoming decade.

{{00:03:25}}
Oh, absolutely.

{{00:03:26}}
And I, I think it's really exciting
to see a bank like ING test its,

{{00:03:32}}
technology and really test the
limits of what it's able to do.

{{00:03:37}}
It's very exciting.

{{00:03:38}}
One of the topics that I've really
witnessed hit throughout these,

{{00:03:44}}
throughout the sessions of the conference.

{{00:03:46}}
is the use of data within these AI
models and how it, how these models

{{00:03:51}}
are trained, what they're able to
reference, and essentially where

{{00:03:55}}
they get their information from.

{{00:03:56}}
In terms of ING, what does your data
strategy look like to ensure that you The,

{{00:04:04}}
AI models or large language models do not
incur any bias in their decision making?

{{00:04:10}}
I think there are two breeds of models
there that we have to evaluate separately.

{{00:04:13}}
First part is like home brew models
and like second ones are the ones

{{00:04:16}}
that we are relying on somebody
else's models, like foundation models.

{{00:04:20}}
And the approach is, similar.

{{00:04:23}}
So for homeroom models, like we have
some regulations that are setting the

{{00:04:27}}
boundaries like GDPR, but we also have
ethical boundaries that made us like

{{00:04:31}}
put together some sort of procedures.

{{00:04:34}}
to evaluate and measure the level of bias.

{{00:04:36}}
These are a bit our internal measures.

{{00:04:39}}
And we are constantly improving, like,
how we are doing this certain procedures.

{{00:04:44}}
But on the other hand we are using more
and more models, foundational models

{{00:04:47}}
coming from Big Tech and other vendors.

{{00:04:49}}
There, our principle is, do not
believe on the declaration, but test

{{00:04:54}}
it yourself and see whether it's
enough, according to bank standards.

{{00:05:00}}
And developing those new
standards and new measures is

{{00:05:04}}
part of our challenge right now.

{{00:05:05}}
We are really constantly spending a lot of
time, a lot of research on how to measure

{{00:05:09}}
bias, how to measure hallucination, in
not only on our own models, but the models

{{00:05:14}}
that we are picking up from the market.

{{00:05:16}}
Two fold approach.

{{00:05:17}}
I really like that approach.

{{00:05:18}}
I think it's, from the
sounds of it very, cautious.

{{00:05:22}}
I think that, ING is very precise.

{{00:05:26}}
and what it would like to achieve with AI.

{{00:05:28}}
And I, really like the sound of, what
you're saying in terms of testing these

{{00:05:32}}
models and really seeing that they
are doing what they're intended to do.

{{00:05:38}}
Working in, finance and in
banking is, one of the most

{{00:05:42}}
regulated industries in the world.

{{00:05:44}}
And I, heard on a panel yesterday
that you could never just give AI

{{00:05:49}}
free reign across a bank because
the regulator would never allow it.

{{00:05:53}}
Now if we look at AI specifically within
the regulatory field of of Europe, we see

{{00:05:59}}
that the European Commission is really
doubling down on its regulation of AI.

{{00:06:04}}
This is best noted in the assent
of the EU AI Act earlier this year

{{00:06:08}}
which is still coming into motion
and being realized and complied with.

{{00:06:13}}
What are then From ING's
perspective, the regulatory hurdles

{{00:06:18}}
in embracing the technology?

{{00:06:21}}
We see Europe's leadership in
regulating AI as, not as a challenge,

{{00:06:25}}
but as a competitive advantage.

{{00:06:28}}
We already have tons of experience
in managing models, governing models,

{{00:06:33}}
measuring models, making sure that
like models are fit for purpose.

{{00:06:36}}
So as a financial industry, we
already have a head start compared to

{{00:06:40}}
other industries, but also compared
to smaller players in the market.

{{00:06:43}}
So that means like our organizations are
more future proof in making sure that

{{00:06:48}}
these models are used in a proper way.

{{00:06:51}}
So I think governing it and like having
a regulation that is governing it within

{{00:06:55}}
Europe is gonna make us safer, all of us.

{{00:06:59}}
And also it's gonna give some banks,
especially bigger banks like with a

{{00:07:02}}
longer history, a slight competitive
advantage in adopting them.

{{00:07:06}}
I think this is what something
Europe did very well.

{{00:07:09}}
Yes I, would agree with you though.

{{00:07:11}}
I actually reported on the
arrival of the regulation.

{{00:07:15}}
I really think that it's a, trailbla and
I think that those European institutions,

{{00:07:19}}
including ING, that, that choose to
comply with it and actively take out

{{00:07:23}}
those measures to be compliant, will be
the ones that lead the way in, in how

{{00:07:29}}
it's deployed within financial services.

{{00:07:32}}
One of the other interesting conversations
that I've been witness to whilst I've

{{00:07:36}}
been here at Money2020 Europe this
week is that the implementation of

{{00:07:42}}
not just AI but technology at large
across the bank must start almost it

{{00:07:48}}
must be innate, it must start from
the top down and reach everybody.

{{00:07:52}}
How does ING then foster a culture
of innovation within itself?

{{00:07:58}}
We have a much more internally
focused innovation approach

{{00:08:01}}
for the last 2 3 years.

{{00:08:02}}
We are thinking innovation
does not necessarily have to

{{00:08:05}}
be external or new products.

{{00:08:07}}
It can also mean making our internal
processes much more efficient, effective,

{{00:08:11}}
and it can make our services much
more suitable to customers needs.

{{00:08:15}}
That's our innovation approach.

{{00:08:17}}
We are not necessarily looking after
new partnerships and FinTechs, but

{{00:08:20}}
we are really looking at technology
and we are trying to see how it

{{00:08:23}}
can make our processes better.

{{00:08:25}}
That's why we are not approaching
it from a technology point of

{{00:08:28}}
view, we are approaching it from a
process and product point of view.

{{00:08:31}}
We are seeing our products and processes
and like we are seeing where do we

{{00:08:35}}
really need help from this technology
and that's why we are applying these

{{00:08:38}}
technologies to a few targeted domains.

{{00:08:42}}
We would like to do these like
few priorities better rather than

{{00:08:46}}
trying this technology everywhere.

{{00:08:48}}
Where are those priorities at the moment?

{{00:08:50}}
Yeah, I think I would
separate it into two groups.

{{00:08:52}}
So like first group is like what
we do within the walls of the bank.

{{00:08:56}}
We want to make the banking
machine a bit more efficient.

{{00:09:00}}
There we use it in know
your customer processes.

{{00:09:03}}
We are going to use it in
contact centers and we are also

{{00:09:05}}
going to use it for coding.

{{00:09:07}}
So these are our biggest groups in
terms of a number of people that

{{00:09:11}}
are working on these processes.

{{00:09:12}}
And also in client facing areas, we're
going to use it in contact centers.

{{00:09:16}}
And we're also gonna use it in marketing.

{{00:09:21}}
We, our idea is, all of our clients are
receiving the same email from us, like

{{00:09:25}}
maybe we should, to a degree, based
on their consent, personalize that

{{00:09:29}}
message that we are giving to them.

{{00:09:30}}
This is gonna help us in satisfying
our customers more and more.

{{00:09:34}}
And that's, a really fantastic point
that you've just touched on there

{{00:09:39}}
again, is personalization within
banking, another leading topic

{{00:09:42}}
here at Money2020 Europe this week.

{{00:09:46}}
And it's almost shocking how innately
tied the topic of AI and how one pursuit

{{00:09:54}}
can almost not exist without the other.

{{00:09:57}}
In that frame of lens, then we had Deer.

{{00:10:00}}
Where can we expect, after all of those
sort of tests and initial entry points are

{{00:10:06}}
done, within ING, across our marketing,
across our communication channels,

{{00:10:10}}
where else can we then expect To see AI
arrive within ING, I see big gap at the

{{00:10:16}}
product layer, and I will explain why.

{{00:10:17}}
So we see a big dominance from
the infrastructure players like a

{{00:10:21}}
media a MD, and like they really,
pumped up their valuations, but like

{{00:10:26}}
they're also dominating the market.

{{00:10:27}}
On top of that one, we see a
dominance from the hyperscalers.

{{00:10:30}}
So like Google, Microsoft, IBM, these
are a bit like dominating with their

{{00:10:33}}
platforms, but what we are not really
seeing yet is like the product layer.

{{00:10:37}}
So like we don't see like
products that are like commonly

{{00:10:39}}
used across the industry.

{{00:10:41}}
Not only in banking, but also in the
other industries, so that product

{{00:10:45}}
layer is going to come and like we're
going to see more and more products

{{00:10:48}}
that are specifically coming either
from FinTechs or other startups that

{{00:10:52}}
are going to be industry standards.

{{00:10:54}}
I think we saw this revolution with
some of the software as a service

{{00:10:58}}
vendors 10 years ago, so that type of
revolutionary wave is going to come,

{{00:11:01}}
I think, in the next four quarters.

{{00:11:03}}
That's what we are not really seeing yet.

{{00:11:05}}
I expect that to evolve further.

{{00:11:07}}
Quite soon though, really, in
the next four quarters, yeah.

{{00:11:10}}
Next time when we meet again in Money 2020
next year, if you can tell me all about

{{00:11:15}}
the benefits of that advancement, and I
really look forward to that conversation.

{{00:11:20}}
But I understand now that I can't keep
you for too long because I understand you

{{00:11:24}}
have to go and prepare for your session.

{{00:11:27}}
How do you feel about going up on
that stage over there later on?

{{00:11:30}}
I'm very excited.

{{00:11:31}}
I think it's a very trendy topic.

{{00:11:33}}
So bankings are not necessarily the
most attractive part of the industry.

{{00:11:40}}
But I think with the new wave of
digitalization and the new wave of AI

{{00:11:44}}
we're going to be much more attractive.

{{00:11:45}}
Fantastic.

{{00:11:46}}
Good luck to you.

{{00:11:47}}
Thank you so much for
this interview, Verhadir.

{{00:11:49}}
It's been a pleasure to speak to
you and to get ING's perspective.

{{00:11:53}}
Thank you to everybody who tuned in
online to listen to this conversation.

{{00:11:59}}
I hope you enjoyed our
conversation just as much as I did.

{{00:12:02}}
And if you would like to find out any
more about ING, AI, Money2020 Europe.

{{00:12:08}}
or technology advancements across
banking in general, please head

{{00:12:12}}
over to our website, Fintechfutures.

{{00:12:14}}
com.

{{00:12:15}}
Until then, see you later.

{{00:12:17}}
Thank you.

Hello, everybody, and
welcome to FinTech Futures.

We are here at day two of
Money 2020 in Amsterdam.

And right now I'm very happy to be
joined by Bahadir Yilmaz of ING.

We've got, the the all the activity
around us, all these people that

But it's really fantastic
to be here with you.

We've caught you at a really interesting
moment because you are on the cusp of

going on stage this afternoon, right
over there on the big stage at the end

Can you, before we, go into what that
session will entail and more about

what ING's plans are with technology
and more specifically AI, maybe you

can give our viewers just a little
bit of a snapshot of who you are

This is Bahadir Yilmaz, and I'm
the Chief Analytics Officer at ING.

I've been working at ING for the
last 30 years, and since last

I'm leading all the AI development
at the bank and like we and my team

are working on digitalizing banking
processes using AI and machine learning.

Yes, on that note, AI and machine
learning is two very leading topics

And I believe that is one of the
foundational stones of your session later.

Can you tell us a little bit ahead of
time what that session is going to entail?

Yeah, later today I will join a
panel discussion on the digital

transformation trends and like the
enabling technologies related to that.

And of course like my focus will be
on generative AI and like I'm going

to talk about the impact of this new
technology in banking transformation.

Very exciting, and I hope to join you
there in the audience, watching you, if

But that's a very interesting
question, and let's build on that.

ING, obviously, is a very
well established bank.

It's got a lot of processes,
a lot of technology at play.

AI, although nothing new, I think
its innovations and capabilities

Now, in a way, to achieve new
heights, which is really exciting.

I think, we are right now seeing the
second wave of AI transformation.

The first wave started in mid
2010s, with the introduction of

And some Fintechs, some neobanks,
but also ING grasped the power of it.

And so we built all of our processes,
organizations, platforms to be able

to cater for that like technology and
also using that in our transformation.

But the real thing happened a year and
a half ago, like with the introduction

of ChetchiPT, everybody started to feel
it, everybody started to touch to it.

So like that, introduced a new game,
like where it is much more tangible

So we think we are ready for that
transformation with our people

capabilities, with our technological
capabilities, and it's going to be really

And I, I think it's really exciting
to see a bank like ING test its,

technology and really test the
limits of what it's able to do.

One of the topics that I've really
witnessed hit throughout these,

throughout the sessions of the conference.

is the use of data within these AI
models and how it, how these models

are trained, what they're able to
reference, and essentially where

In terms of ING, what does your data
strategy look like to ensure that you The,

AI models or large language models do not
incur any bias in their decision making?

I think there are two breeds of models
there that we have to evaluate separately.

First part is like home brew models
and like second ones are the ones

that we are relying on somebody
else's models, like foundation models.

So for homeroom models, like we have
some regulations that are setting the

boundaries like GDPR, but we also have
ethical boundaries that made us like

to evaluate and measure the level of bias.

And we are constantly improving, like,
how we are doing this certain procedures.

But on the other hand we are using more
and more models, foundational models

There, our principle is, do not
believe on the declaration, but test

it yourself and see whether it's
enough, according to bank standards.

And developing those new
standards and new measures is

We are really constantly spending a lot of
time, a lot of research on how to measure

bias, how to measure hallucination, in
not only on our own models, but the models

I think it's, from the
sounds of it very, cautious.

and what it would like to achieve with AI.

And I, really like the sound of, what
you're saying in terms of testing these

models and really seeing that they
are doing what they're intended to do.

Working in, finance and in
banking is, one of the most

And I, heard on a panel yesterday
that you could never just give AI

FinTech Futures

Expert Insight

Touching base with FinTech Futures at Money20/20 Europe, Bahadir Yilmaz, chief analytics officer at ING, discusses how the Dutch banking giant is executing its AI strategy, with insights into how the new face of innovation is being freshly cultivated within the bounds of fast-emerging regulatory safeguards.

