Nu Videocast IR | Artificial Intelligence at Nu

00:27:16
https://www.youtube.com/watch?v=07AlA4qrv0w

Ringkasan

TLDRI denne video præsenterer York Fredman, investor relations officer hos New Bank, CTO Victor, der deler sin indsigt i AI's rolle i banksektoren. De diskuterer, hvordan New Bank har integreret AI i kreditvurdering og kundeservice og viser hvordan avanceret teknologi kan forbedre effektivitet og kundeinteraktion. Victor fremhæver også betydningen af Open Finance, hvor deling af data kan berige beslutningsprocesser. AI præsenteres som en vital mulighed snarere end en blot en omkostning, med potentiale til at give værdi for både kunder og banken.

Takeaways

  • 🤖 AI er en transformativ kraft for banker.
  • 📊 Kreditvurdering kan forbedres gennem dyb læring.
  • 💬 AI forbedrer kundeservice gennem hurtigere respons.
  • 🔍 Open Finance muliggør bedre dataudnyttelse.
  • 📈 AI ses som en eksistentiel mulighed for værdi.
  • ⚖️ Der er risici ved AI, som skal håndteres omhyggeligt.
  • 💼 AI kan skabe nye jobmuligheder i stedet for at fjerne dem.
  • ⏱️ Hurtigere beslutningstagning er gavnligt for kundeoplevelsen.
  • 🌐 Kunder skal kunne stole på, at deres data deles ansvarligt.
  • 🔄 Automatisk betalingsstyring er en del af fremtidens banktjenester.

Garis waktu

  • 00:00:00 - 00:05:00

    Videoen præsenterer en samtale om AI-muligheder og risici i NewBank, ledet af investorrelationsofficer York Fredman og CTO Victor. De diskuterer, hvordan AI har været en central del af NewBanks strategi siden begyndelsen og dens anvendelse i områder som kreditvurdering og svindeldetektion.

  • 00:05:00 - 00:10:00

    Victor forklarer, hvordan NewBank anvender AI til at forbedre kundeservice og interne processer, og han præsenterer nogle KPI'er. Der er set signifikante forbedringer i kundeservicesystemer, samt hurtigere og mere præcise processer til intern softwareudvikling og analyse.

  • 00:10:00 - 00:15:00

    Samtalen omhandler også de potentielle risici ved brugen af AI, især i en reguleret finanssektor. Victor understreger vigtigheden af at have kontrolmekanismer på plads for at sikre, at AI-systemer fungerer ordentligt, og at der ikke er negative konsekvenser for kunderne.

  • 00:15:00 - 00:20:00

    Yderligere diskuteres, hvordan AI kan forbedre serviceniveauet på kundeservicesagen. Det forventes, at AI vil generere hurtigere svar og sikre, at kunderne får en bedre oplevelse ved at reducere behovet for manuelle interaktioner.

  • 00:20:00 - 00:27:16

    Victor afslutter med at tale om vigtigheden af at integrere AI i NewBanks forretningsmodeller, herunder i forhold til kreditvurderingsprocesser og øget effektivitet gennem data fra åbne finansieringsmodeller. Fokus for NewBank er at skabe værdi for kunderne gennem deres innovative AI-tilgang.

Tampilkan lebih banyak

Peta Pikiran

Video Tanya Jawab

  • Hvad er formålet med videoen?

    At diskutere betydningen af AI for New Banks og dens strategier.

  • Hvilke områder ser New Bank som de mest påvirkede af AI?

    Kreditvurdering og kundeservice.

  • Hvordan håndterer New Bank de risici, der er forbundet med AI?

    Ved at implementere strenge overvågnings- og valideringsmekanismer.

  • Hvad er 'Open Finance'?

    En platform, der tillader finansielle institutioner at dele kundedata med deres samtykke.

  • Hvad er potentialet for AI i banksektoren ifølge Victor?

    At revolutionere operations og forbedre kundeoplevelser markant.

Lihat lebih banyak ringkasan video

Dapatkan akses instan ke ringkasan video YouTube gratis yang didukung oleh AI!
Teks
en
Gulir Otomatis:
  • 00:00:05
    hello and welcome to the first edition
  • 00:00:07
    of our new video cast today's session is
  • 00:00:10
    powered by the team and I am York
  • 00:00:13
    fredman new bank's investor relations
  • 00:00:15
    officer starting today I will begin a
  • 00:00:18
    series of conversations with new bank's
  • 00:00:20
    leaders to talk about hot topics for our
  • 00:00:22
    investor community and it is my pleasure
  • 00:00:25
    to have here VOR leier who is our chief
  • 00:00:28
    technology officer V is a longtime no
  • 00:00:31
    Banker one of the first Engineers to
  • 00:00:33
    join the company and he's here to talk
  • 00:00:35
    about our vision for AR diffusion TS
  • 00:00:38
    thank you Victor thank you so much y
  • 00:00:40
    it's a pleasure to be here pleasure to
  • 00:00:41
    talk to you about AI our first question
  • 00:00:45
    is about you know a topic that has
  • 00:00:48
    gained great relevance in
  • 00:00:50
    2023 and should continue to grow in
  • 00:00:53
    interest and applications so how does
  • 00:00:55
    new see AI in terms of opportunities and
  • 00:00:59
    risks that's a great question and a
  • 00:01:02
    question that for us is not a 2023
  • 00:01:04
    question it's a question that we've
  • 00:01:06
    started thinking about since the very
  • 00:01:08
    early days of newbank we came from an
  • 00:01:12
    idea that we needed to really accelerate
  • 00:01:14
    our learnings as we started from scratch
  • 00:01:17
    you know a tech company in financial
  • 00:01:18
    services we didn't have decades to catch
  • 00:01:21
    up with incumbents on the abilities to
  • 00:01:23
    detect fraud to underwrite credit so we
  • 00:01:26
    saw the ability to create a system that
  • 00:01:29
    was Cloud native that had a lot of rigor
  • 00:01:32
    on how we defined and created our data
  • 00:01:35
    infrastructure that was highly scalable
  • 00:01:37
    that leverage machine learning to make
  • 00:01:39
    the best decisions possible so we could
  • 00:01:41
    really accelerate that Loop of learning
  • 00:01:44
    thought that was the only way that we
  • 00:01:45
    can compete in this market and that's
  • 00:01:47
    what we did so we've applied machine
  • 00:01:49
    learning techniques at scale in several
  • 00:01:52
    parts of our business so in credit
  • 00:01:54
    underwriting and fraud and operations
  • 00:01:56
    and for us what we see now is just
  • 00:01:58
    another chapter in that Journey whereas
  • 00:02:01
    in the past we used a lot of the
  • 00:02:02
    supervised learning types of techniques
  • 00:02:05
    we see with llms and the massive impact
  • 00:02:08
    they've had this year that we can
  • 00:02:10
    actually apply new types of Technologies
  • 00:02:14
    to a broad set of problems in our
  • 00:02:15
    business and we're early adopters and
  • 00:02:17
    we're ready to really leverage it in a
  • 00:02:19
    in a way that we believe is going to be
  • 00:02:21
    transformational for our customers and
  • 00:02:23
    for Value creation for new bank as well
  • 00:02:26
    Victor what is the area of Fu in which
  • 00:02:28
    we see the fastest largest effects of AI
  • 00:02:33
    implementation so first I'll I'll start
  • 00:02:36
    by defining a bit of how we see AI right
  • 00:02:39
    so artificial intelligence we see it as
  • 00:02:42
    automated systems that leverage some
  • 00:02:45
    sort of machine learning to make
  • 00:02:46
    intelligent Intelligent Decisions right
  • 00:02:48
    to be able to make decisions that create
  • 00:02:50
    a lot of value and in that we also see
  • 00:02:54
    uh the subset that's more recent right
  • 00:02:57
    now which is generative AI so so when we
  • 00:03:00
    see think about AI as a whole immense
  • 00:03:03
    amount of value creation comes from
  • 00:03:05
    credit underwriting and we see our
  • 00:03:06
    differentiation our ability to
  • 00:03:07
    underwrite at scale so quickly and
  • 00:03:09
    iterating our models in a way that is
  • 00:03:12
    really impactful in our ability to grow
  • 00:03:15
    safely in a way we can detect fraudsters
  • 00:03:18
    in a way that we can route chats route
  • 00:03:22
    uh calls that we can understand what the
  • 00:03:24
    customer needs when they need it so this
  • 00:03:27
    is something that we have a lot of AI
  • 00:03:28
    systems in the backgound
  • 00:03:30
    but now when we see look at llms and
  • 00:03:33
    generative AI we basically believe that
  • 00:03:37
    there's a a new set of possibilities
  • 00:03:41
    right that we can can use and the first
  • 00:03:44
    early adopter pie really will be in the
  • 00:03:47
    operation side that's where we're
  • 00:03:49
    talking to customers it's natural
  • 00:03:51
    language there's a lot of back and forth
  • 00:03:53
    it's something that requires a lot of
  • 00:03:55
    precision and something that's very time
  • 00:03:57
    sensitive so being able to use llms in a
  • 00:04:00
    very smart way to improve that
  • 00:04:03
    interaction with customers we think it's
  • 00:04:05
    transformational but it's just not just
  • 00:04:07
    the interface of the customer we also
  • 00:04:09
    see the co-pilot concept permeating
  • 00:04:12
    every aspect of our productivity so it
  • 00:04:14
    starts with agents agents having a
  • 00:04:17
    co-pilot a new bank co-pilot that can
  • 00:04:19
    help answer a ticket faster or parse the
  • 00:04:23
    information better or categorize a type
  • 00:04:25
    of back office flow in a you know a more
  • 00:04:28
    precise way or in a faster way and a
  • 00:04:31
    more assertive way and then we also go
  • 00:04:33
    to productivity touching U software
  • 00:04:36
    development and that's something that
  • 00:04:38
    we're also experimenting with can we use
  • 00:04:40
    copilot to write softer that's more
  • 00:04:42
    assertive that's better can we use it
  • 00:04:43
    also as a sanity check for us to make
  • 00:04:45
    sure that there are no bugs there no
  • 00:04:47
    issues right we also see it on the
  • 00:04:49
    analytics side and and and and and how
  • 00:04:52
    we basically can parse information A
  • 00:04:54
    Better Way come to better decisions vtor
  • 00:04:57
    can you share some kpis of how no has
  • 00:05:00
    been using AI in client facing and
  • 00:05:03
    non-client facing activities and how you
  • 00:05:05
    expect these kpis to evolve over time
  • 00:05:09
    yeah so for example on clients facing
  • 00:05:12
    usages of AI we've had deep learning
  • 00:05:15
    models interacting with customers for a
  • 00:05:17
    while now and it's something that we
  • 00:05:20
    have been working on even before the
  • 00:05:22
    explosion of geni uh last year but as we
  • 00:05:26
    now use gen into these models we see
  • 00:05:31
    sometimes doubling tripling of
  • 00:05:34
    deflections or self-surface rates from
  • 00:05:37
    customers so it means that the customer
  • 00:05:39
    is basically just by talking to the AI
  • 00:05:42
    they can Self Service in a rate that's a
  • 00:05:44
    step change from where we we're getting
  • 00:05:46
    before so it's a it's a really
  • 00:05:49
    significant change and when we answer
  • 00:05:51
    the customer and they're satisfied with
  • 00:05:52
    it and we do it quickly you know it's a
  • 00:05:55
    it's a magical type of experience and
  • 00:05:58
    then when we look internally at when we
  • 00:06:00
    look at things like code generation test
  • 00:06:02
    generation we're seeing also a great
  • 00:06:04
    speed up in our ability to generate that
  • 00:06:06
    code and that those types of things
  • 00:06:09
    they're not 5% improvements here 3%
  • 00:06:11
    improvements there these are step change
  • 00:06:14
    improvements where deploy it in back
  • 00:06:16
    office processes for example the ability
  • 00:06:18
    to get a customer report around
  • 00:06:20
    something that they deem to be a a
  • 00:06:23
    transaction they don't don't recognize
  • 00:06:25
    or a fraudulent Behavior we can much
  • 00:06:27
    more assertively again step change in
  • 00:06:30
    Precision dictate is this what type of
  • 00:06:33
    reason is this what type ofo it is and
  • 00:06:36
    with that we can really drastically
  • 00:06:38
    speed up our Asser our ability to
  • 00:06:40
    service our customers and to solve the
  • 00:06:43
    problem so we're seeing really big step
  • 00:06:45
    changes everywhere we deploy them and
  • 00:06:47
    naturally these comes with small tests
  • 00:06:49
    and then at okay we see different things
  • 00:06:51
    but it's always something that uh it's
  • 00:06:53
    really exciting it's not something
  • 00:06:55
    around the edge it's not a marginal
  • 00:06:56
    thing it it does feel like a platform
  • 00:06:58
    shift but just to address one point that
  • 00:07:02
    I think we can't you know talk about AI
  • 00:07:04
    without talking about risks and some of
  • 00:07:06
    the risks that might exist so I think to
  • 00:07:09
    get the obvious out first right we're a
  • 00:07:11
    regulated entity that means that we need
  • 00:07:14
    and must abide by all regulatory
  • 00:07:16
    framework that's you know on on top of
  • 00:07:18
    us and we do that uh with a lot of rigor
  • 00:07:22
    and a lot of precision not just for AI
  • 00:07:23
    for everything that we operate in uh but
  • 00:07:26
    then AI I think introduces new types of
  • 00:07:28
    risks right so there there's the risk of
  • 00:07:31
    just what is this new system that humans
  • 00:07:34
    it's not just us but like humans don't
  • 00:07:36
    fully understand how it works and how to
  • 00:07:39
    explain it and it's for us it's all
  • 00:07:42
    about having a very rigorous approach to
  • 00:07:45
    evaluate the input the output of these
  • 00:07:48
    models and to have guard rails and
  • 00:07:50
    mechanisms to make sure that it's not
  • 00:07:52
    behaving in a way that can be value
  • 00:07:54
    destructive and that's a series of tests
  • 00:07:57
    it's a Ser series of monitoring
  • 00:07:58
    mechanisms it's also part of manual
  • 00:08:01
    validation so that's something that's
  • 00:08:04
    really really important for us for us is
  • 00:08:06
    to even in an environment in which fully
  • 00:08:08
    explainability is not there we can
  • 00:08:10
    ensure safety by having other mechanisms
  • 00:08:14
    in place to ensure that safety the other
  • 00:08:17
    point that I think is also very
  • 00:08:18
    important to to highlight here is when
  • 00:08:20
    we talk about AI there's a bit of oh the
  • 00:08:23
    machines are going to take over the
  • 00:08:25
    world and what happens to humans
  • 00:08:27
    and my belief
  • 00:08:30
    is every technology shift that Humanity
  • 00:08:34
    has faced down the line met meant the
  • 00:08:37
    creation of new jobs meant the a change
  • 00:08:40
    of new types of work it me meant the
  • 00:08:42
    elevation of humans to be able to have
  • 00:08:46
    even more leverage to impact more people
  • 00:08:49
    to create more value and I truly believe
  • 00:08:52
    especially coming from a very human
  • 00:08:54
    Centric company because we're human from
  • 00:08:56
    with how we treat our customers with how
  • 00:08:58
    we treat our uh our team we believe that
  • 00:09:02
    having that thoughtful approach on how
  • 00:09:03
    we use AI that puts human as you know
  • 00:09:07
    the technology service of humans that we
  • 00:09:10
    can Empower our teams that we can
  • 00:09:13
    service our customers that people can
  • 00:09:15
    get more exciting work get more leverage
  • 00:09:18
    right so that value the human as a human
  • 00:09:21
    The Human Experience the human work for
  • 00:09:23
    us is not contradictory to valueing AI
  • 00:09:27
    but it's a part of that wheel and
  • 00:09:30
    something that we we're very very
  • 00:09:32
    focused on and going to continue
  • 00:09:33
    investing in this direction as well you
  • 00:09:36
    touched upon very in interesting uh
  • 00:09:39
    different levels you touched upon uh
  • 00:09:43
    credit under writing customer
  • 00:09:46
    interaction is there evidence that the
  • 00:09:48
    level of service for instance can
  • 00:09:51
    improve with the use of AI absolutely
  • 00:09:54
    absolutely so we see that AI provides a
  • 00:10:00
    few things that are quite powerful right
  • 00:10:03
    AI is going to be able to one answer
  • 00:10:05
    things immediately not requiring a queue
  • 00:10:08
    of people to actually get to your ticket
  • 00:10:11
    so a lot of the times for customers that
  • 00:10:13
    ability to get some input some feedback
  • 00:10:16
    in a timely fashion that's really really
  • 00:10:19
    valuable right so just First Response
  • 00:10:22
    times is something that can drastically
  • 00:10:23
    change but on top of that we're working
  • 00:10:26
    to get to very high levels of
  • 00:10:28
    conservativeness and so we can get to
  • 00:10:30
    self-resolution to the customer to
  • 00:10:32
    actually get in one contact be able to
  • 00:10:36
    fully answer and deal with their issue
  • 00:10:39
    in a way that with a human would take a
  • 00:10:41
    lot more interactions just naturally
  • 00:10:42
    just because of the nature of how it is
  • 00:10:44
    to communicate with the human so we
  • 00:10:46
    think that uh AI can provide us uh a
  • 00:10:49
    speed that we can create it in a way in
  • 00:10:51
    a thoughtful way that will not uh
  • 00:10:53
    decrease quality and that we can create
  • 00:10:56
    mechanisms that we can iterate tests in
  • 00:10:58
    a safe way in a safe envir environment
  • 00:11:00
    and part of that safety and part of the
  • 00:11:02
    the way we think about it is if it can
  • 00:11:04
    create scape valves if the customer is
  • 00:11:07
    not we don't want anyone to be trapped
  • 00:11:09
    talking to an AI right we want people to
  • 00:11:11
    use the AI as as long as it's providing
  • 00:11:14
    more value than the alternative waiting
  • 00:11:16
    for someone to deal with that
  • 00:11:18
    interaction so that combo really is a a
  • 00:11:21
    way that we can only we only aim to get
  • 00:11:23
    the upside right so that's how we're
  • 00:11:25
    thinking about that the deploying that
  • 00:11:27
    value to to to our customers without
  • 00:11:29
    even talking about uh you know
  • 00:11:31
    efficiency which is one of the pillars
  • 00:11:33
    of the company things that I would be
  • 00:11:35
    doing uh in dollar terms you are now
  • 00:11:39
    doing in cents I guess so we the way we
  • 00:11:44
    think about using Ai and know I think
  • 00:11:46
    even the way we make revenue and cost
  • 00:11:49
    decisions is there there are one
  • 00:11:52
    decision which is ultimately uh what's
  • 00:11:55
    most value ACC creative to the customer
  • 00:11:57
    and we're going to see that what the
  • 00:11:59
    customer cares about is for their issues
  • 00:12:01
    to be resolved quickly uh in a way that
  • 00:12:04
    they don't have a lot of friction you
  • 00:12:06
    don't have a lot of points of contact
  • 00:12:07
    and when we think about having to wait
  • 00:12:09
    on a queue for a human having to get the
  • 00:12:11
    wrong answer for some reason and so on
  • 00:12:14
    uh a lot of the things that have are
  • 00:12:17
    related to costs are actually just
  • 00:12:19
    inefficiencies of the system so if I can
  • 00:12:22
    plug in technology to remove those
  • 00:12:23
    inefficiencies I would get a better
  • 00:12:25
    service and by removing the
  • 00:12:28
    inefficiencies I remove some of these
  • 00:12:29
    costs and I get a much more engaged
  • 00:12:32
    customer which is ultimately what we
  • 00:12:34
    care and a more engaged customer is a
  • 00:12:35
    customer that will do more business with
  • 00:12:37
    us come back with us we come back to to
  • 00:12:40
    engage with us in a daily basis and then
  • 00:12:42
    we can have a much much deeper
  • 00:12:43
    connection with them and yes we aim to
  • 00:12:46
    be a price leader because we think in
  • 00:12:48
    financial services being a price leader
  • 00:12:50
    is crucial but we don't do it at the
  • 00:12:52
    cost of quality or we don't do it at the
  • 00:12:54
    cost of the experience we think it's
  • 00:12:55
    actually uh the right path is to have
  • 00:12:58
    the two to have the lower cost with a
  • 00:13:00
    better experience conversely we are
  • 00:13:03
    actually increasing primary banking
  • 00:13:05
    accounts which is one of the main levels
  • 00:13:08
    for our value creation right while
  • 00:13:10
    decreasing cost to serve per customer so
  • 00:13:13
    that is exactly the formula that we
  • 00:13:14
    believe is a winning formula let's I
  • 00:13:17
    know touch upon now um the credit
  • 00:13:21
    underwriting improvements that AI might
  • 00:13:24
    provide customer scoring and credit
  • 00:13:27
    modeling are some of the potential
  • 00:13:29
    successful applications of AI as well so
  • 00:13:32
    how do we see this evolving so that is a
  • 00:13:35
    really interesting piece uh we might be
  • 00:13:39
    one of the largest uh machine learning
  • 00:13:41
    systems that does credit underw writing
  • 00:13:44
    just out of the sheer scale scale of the
  • 00:13:46
    number of customers we have number of
  • 00:13:48
    decisions we have right every day we
  • 00:13:49
    make a decision on do I increase your
  • 00:13:51
    limit or not so it's a hundreds of
  • 00:13:53
    millions billions of decisions that we
  • 00:13:55
    make uh and what we saw is that with a
  • 00:13:59
    combo of having the right data ensuring
  • 00:14:02
    data quality ensuring uh the processing
  • 00:14:04
    of the data in an efficient way in a
  • 00:14:06
    scalable way that we can deploy models
  • 00:14:09
    machine learning models using uh
  • 00:14:12
    supervised learning techniques or
  • 00:14:14
    regression classifications the more
  • 00:14:16
    standard approaches we show that we can
  • 00:14:20
    do that in a way that we can iterate
  • 00:14:22
    quickly we can underwrite with quality
  • 00:14:25
    understand what we're doing and do so in
  • 00:14:27
    a way that in 10 years can catch up and
  • 00:14:30
    often times surpass the ability to
  • 00:14:32
    underwrite of O other
  • 00:14:34
    competitors what we see now with deep
  • 00:14:37
    learning and generative AI is a big
  • 00:14:39
    opportunity to wait those traditional
  • 00:14:43
    techniques they brought a lot of
  • 00:14:45
    Leverage for us can we also apply these
  • 00:14:48
    new technologies can we also use deep
  • 00:14:51
    learning with unstructured data to a to
  • 00:14:55
    get a credit scoring to make credit
  • 00:14:57
    decisions as well
  • 00:14:59
    and that has a component around
  • 00:15:01
    rethinking how we plug our data the
  • 00:15:04
    types of models we create and how we
  • 00:15:06
    validate the models and how we also uh
  • 00:15:09
    explain those models it would be
  • 00:15:11
    basically models against models to
  • 00:15:13
    understand the types of decisions we're
  • 00:15:14
    making because just of the very nature
  • 00:15:16
    of deep learning so that's a journey
  • 00:15:18
    that we're just beginning and it's a
  • 00:15:20
    journey that we believe is can be a step
  • 00:15:22
    change when we see in adtech we see that
  • 00:15:26
    deep learning meant often times a step
  • 00:15:28
    change in ability to convert customers
  • 00:15:31
    so that increased Precision that we see
  • 00:15:33
    saw in deep learning and advertising we
  • 00:15:35
    think that perhaps that also exists in
  • 00:15:37
    credit but the co the implications of a
  • 00:15:40
    wrong decision in credit are much higher
  • 00:15:43
    than a wrong decision around what ad
  • 00:15:45
    you're showing a customer so we take
  • 00:15:49
    absolute the the seriousness of
  • 00:15:52
    deploying a new credit model it's
  • 00:15:55
    absolutely something that we take very
  • 00:15:57
    seriously so we are our early adopters
  • 00:15:59
    we experiment with things but we fully
  • 00:16:03
    appreciate the seriousness of making any
  • 00:16:05
    one decision that's wrong with credit so
  • 00:16:06
    we're taking that in a very thoughtful
  • 00:16:08
    experimenting in a very you know careful
  • 00:16:11
    way uh deliberate way but if you're
  • 00:16:14
    going to ask me is there that jump I
  • 00:16:16
    think it's it's possible Right but I
  • 00:16:18
    think we need more time to iterate and
  • 00:16:20
    see that happening in in in production
  • 00:16:23
    in reality and it's going to take a
  • 00:16:24
    little while and the bar is set higher
  • 00:16:26
    because we already are overall better
  • 00:16:30
    than most of our peers in terms of asset
  • 00:16:34
    quality on an income level basis on a
  • 00:16:36
    like for like basis so I think when you
  • 00:16:39
    mention about uh you know uh improving
  • 00:16:43
    further I you are considering probably
  • 00:16:46
    other benchmarks elsewhere yes so we we
  • 00:16:49
    are and I think it's just not just about
  • 00:16:51
    being better on a you know cost or risk
  • 00:16:55
    basis per income it's about
  • 00:16:57
    understanding your decision decisions
  • 00:16:58
    it's about being able to say this is the
  • 00:17:01
    strategy to the optimizes value creation
  • 00:17:04
    for my customer base often times
  • 00:17:07
    optimizing value creation doesn't
  • 00:17:08
    necessarily mean reducing risk right or
  • 00:17:10
    getting to zero risk it's about being
  • 00:17:13
    intentional it's understanding and
  • 00:17:15
    saying this is what I want to happen I
  • 00:17:17
    have the tools that allow me to do that
  • 00:17:19
    and then seeing it play out so we have
  • 00:17:21
    been able to do that with more
  • 00:17:23
    traditional machine learning and we if
  • 00:17:26
    and when we deploy deep learning
  • 00:17:28
    techniques in credit it it has to be
  • 00:17:31
    like that or better so vctor let's
  • 00:17:35
    switch gears and talk about open finance
  • 00:17:38
    for those that watching us uh in other
  • 00:17:40
    countries open finance is a platform in
  • 00:17:42
    Brazil through which financial
  • 00:17:44
    institutions are allowed to share
  • 00:17:46
    customers financial data under their
  • 00:17:49
    consent SII implementation requires
  • 00:17:53
    quality data in high volumes those two
  • 00:17:56
    go hand in hand can give me a few
  • 00:17:59
    examples of how we can leverage open
  • 00:18:01
    financing Ai and the other way around
  • 00:18:05
    that is a phenomenal question so one of
  • 00:18:07
    the things that I mentioned earlier
  • 00:18:09
    was our need to accelerate our learnings
  • 00:18:12
    because we were starting you know
  • 00:18:14
    recently versus having decades uh as a
  • 00:18:17
    company part of the issue there is we
  • 00:18:19
    didn't have that history of the data and
  • 00:18:22
    what the data feels like for the you
  • 00:18:24
    know the customer elsewhere the past
  • 00:18:26
    decade the past two decades their family
  • 00:18:29
    and so on so that richness of data was
  • 00:18:31
    something that was very hard to get to
  • 00:18:33
    and we had to infer through other
  • 00:18:34
    mechanisms and and and find ways to
  • 00:18:36
    mitigate that absence with open
  • 00:18:39
    finance we now get more access if the
  • 00:18:42
    customer grants us that access the
  • 00:18:44
    customer trusts us to that which should
  • 00:18:46
    we believe it's one of the most big the
  • 00:18:49
    biggest signs of that we have a strong
  • 00:18:50
    connection with the customer is there
  • 00:18:52
    willingness to share that information
  • 00:18:53
    with us because it shows hey I trust you
  • 00:18:55
    with my data I trust you with my salary
  • 00:18:57
    this is a you know a relationship that
  • 00:18:59
    it's going to be value Creator let's do
  • 00:19:01
    it so we really value and take it very
  • 00:19:02
    seriously that decision by the way we
  • 00:19:05
    are Market leaders right now in consents
  • 00:19:08
    13.5 million consents right yes yes yes
  • 00:19:11
    that's right that's right so that now
  • 00:19:14
    that we have the data then the data can
  • 00:19:17
    feed into into our models and a model is
  • 00:19:20
    only as good as the quality of the data
  • 00:19:21
    that you have and being able to tap into
  • 00:19:24
    this high volume of data this history uh
  • 00:19:27
    really enriches our models enriches our
  • 00:19:30
    ability to differentiate risk enriches
  • 00:19:31
    our ability to tailor the right product
  • 00:19:33
    to the right customer so that's a a huge
  • 00:19:36
    element of our open finance strategy
  • 00:19:39
    which is only part of the strategy the
  • 00:19:41
    other piece is open AI allows us to also
  • 00:19:44
    have side effects that connect to uh the
  • 00:19:47
    open finance examples I can initiate a
  • 00:19:49
    payment elsewhere and move make money
  • 00:19:52
    movements uh and we believe that because
  • 00:19:55
    of that we can start being the the main
  • 00:19:57
    interface with the cust customer they go
  • 00:19:59
    to new because the experience is
  • 00:20:00
    smoother because they trust us they feel
  • 00:20:02
    like it's a it's a better way to bank
  • 00:20:04
    and they can Bank on other Banks through
  • 00:20:08
    us so that's something that we want to
  • 00:20:09
    more and more push it's like come to us
  • 00:20:12
    and we will help you Bank through other
  • 00:20:14
    mechanisms but as I mentioned trust one
  • 00:20:17
    of the things that we're committed and
  • 00:20:18
    we're working on is we believe that
  • 00:20:22
    AI is has the potential of bringing
  • 00:20:25
    bringing the private Banker to the
  • 00:20:27
    everyone's pocket
  • 00:20:29
    if the mobile phone brought the bank to
  • 00:20:31
    everyone's pocket AI can bring the
  • 00:20:33
    private bank someone can help you assist
  • 00:20:35
    andone so Our intention is that this
  • 00:20:38
    semester we want to launch the first
  • 00:20:39
    version of this and part of the the the
  • 00:20:43
    goal of this AI is not to just offer
  • 00:20:47
    what's the right product for the
  • 00:20:49
    customer within new bank it's to go
  • 00:20:51
    beyond the the boundaries of new bank
  • 00:20:52
    and say what's out there what exists and
  • 00:20:55
    can get open finance data we can get
  • 00:20:58
    data that's publicly available and say
  • 00:20:59
    what's the best product for this
  • 00:21:02
    customer at this point in time and we
  • 00:21:04
    will offer that product even if it's not
  • 00:21:06
    our product because our goal here is to
  • 00:21:09
    establish a relationship of trust and we
  • 00:21:12
    believe that you know the long-term uh
  • 00:21:16
    shareholder value customer value all the
  • 00:21:18
    interests are aligned it's just a matter
  • 00:21:20
    of looking in a really longterm priz the
  • 00:21:23
    early days of our money platform right
  • 00:21:25
    yes internally how do you think since
  • 00:21:28
    you touched uh on revenues uh about
  • 00:21:31
    whether AI is a cost or a revenue
  • 00:21:35
    opportunity I think AI is is more of
  • 00:21:39
    like an existential opportunity it's a
  • 00:21:42
    transformational opportunity and it will
  • 00:21:43
    touch fundamentally rethinking uh our
  • 00:21:47
    revenue streams it will fundamentally
  • 00:21:49
    change our interface with users and will
  • 00:21:53
    fundamentally change our uh operations
  • 00:21:56
    and our ability to service ticket
  • 00:21:58
    tickets and make decisions and deploy
  • 00:22:00
    code and so on so I think it's a
  • 00:22:02
    platform shift that have implications
  • 00:22:04
    that touch every business in different
  • 00:22:07
    degrees it's the cultural aspect of
  • 00:22:09
    making sure that the customer is at the
  • 00:22:11
    Forefront that we think will
  • 00:22:12
    differentiate who uh the companies that
  • 00:22:15
    will win uh this platform shift and the
  • 00:22:17
    companies that will uh not be able to
  • 00:22:19
    Leverage The Tool as much listening to
  • 00:22:22
    you it Rems me about pixs which was
  • 00:22:25
    faced by many parts of the industry as
  • 00:22:30
    uh you know a cost only yes we always
  • 00:22:32
    face that as uh you know savings
  • 00:22:35
    opportunity first and then as a revenue
  • 00:22:39
    driver and I think it's proven right as
  • 00:22:41
    P financing scales every day right
  • 00:22:43
    that's that's the best uh comparison
  • 00:22:46
    that we we can give which is pix
  • 00:22:50
    objectively is good for customers if you
  • 00:22:53
    can make a transfer that's free 247
  • 00:22:59
    reliable that changes your relationship
  • 00:23:01
    with money it enables businesses that
  • 00:23:04
    couldn't exist it remove like reduces
  • 00:23:07
    the the the barrier of entry into the
  • 00:23:10
    the uh you know formal economy in a way
  • 00:23:13
    that no technolog is ever done it's a
  • 00:23:16
    transformational
  • 00:23:17
    technology and if you just look at it as
  • 00:23:20
    the first step oh the first step is oh
  • 00:23:24
    the the players that were're charging
  • 00:23:25
    for for wire transfers we never charge
  • 00:23:27
    for all right but there are players that
  • 00:23:30
    were charging for it the first thing
  • 00:23:31
    they're going to look at is like oh this
  • 00:23:33
    is a big cut in my Revenue line need to
  • 00:23:35
    fight against it or we're not going to
  • 00:23:37
    adopt it or this is not going to be good
  • 00:23:38
    and so and it's understandable why
  • 00:23:40
    that's the gut reaction but we we saw it
  • 00:23:44
    as it's good for customers great they'll
  • 00:23:47
    come back they need to love it they need
  • 00:23:49
    to love doing it with us and if they do
  • 00:23:51
    it with us then maybe we earned the
  • 00:23:55
    right to offer something else all value
  • 00:23:58
    propositions around it right and I think
  • 00:24:00
    pix financing is a great example of that
  • 00:24:03
    uh I think it's not the last example I
  • 00:24:05
    think there will be a lot of other
  • 00:24:06
    Innovations we can do on top of that and
  • 00:24:08
    I think open finance is a machine of
  • 00:24:10
    doing exactly
  • 00:24:12
    that how engaged in AI do you see new
  • 00:24:17
    Banks management team are they really
  • 00:24:20
    investing in it yes yes uh it is a daily
  • 00:24:26
    basis concern it's something that Davids
  • 00:24:28
    on top of all the time every new uh
  • 00:24:32
    announcement every new uh technology
  • 00:24:34
    coming out every new project you know
  • 00:24:37
    our M team is engaged trying to
  • 00:24:39
    understand how to leverage it trying to
  • 00:24:40
    really stress test our strategy obsessed
  • 00:24:43
    about our ability to get the best talent
  • 00:24:45
    to work on this uh making sure that we
  • 00:24:47
    have the appropriate uh funds to support
  • 00:24:51
    that growth speaking of our future
  • 00:24:53
    Vision now we already touched upon the
  • 00:24:57
    concept of the money platform new is no
  • 00:24:59
    for pushing the industry forward with
  • 00:25:02
    Innovative technology how do we stay
  • 00:25:04
    relevant and ahead in a topic that's
  • 00:25:06
    been so heavily studied are there any
  • 00:25:09
    less obvious applications that you think
  • 00:25:12
    someone might be
  • 00:25:13
    missing so I I don't think it's about
  • 00:25:17
    less obvious I think it's about creating
  • 00:25:21
    applications that feel obvious but
  • 00:25:24
    they're hard and they require a lot of
  • 00:25:27
    non sexy work and they require the right
  • 00:25:31
    cultural alignment and obsession with
  • 00:25:34
    the customer so when we talk about money
  • 00:25:37
    platform which you know in a very
  • 00:25:39
    simplified way it's about getting the
  • 00:25:41
    private Banker into everyone's pocket
  • 00:25:44
    it's about things like a self-driving
  • 00:25:45
    Bank the idea that you don't need to be
  • 00:25:50
    consciously aware of the actions that we
  • 00:25:52
    you're taking all the time to make sound
  • 00:25:55
    financial
  • 00:25:56
    decisions and and to be able to do that
  • 00:26:00
    it's a lot of Plumbing work it's about
  • 00:26:02
    making sure that auto payments work it's
  • 00:26:04
    about making sure that you have all the
  • 00:26:07
    payment mechanisms in place that you can
  • 00:26:10
    create the experiences that allow the
  • 00:26:12
    customer to make the right decisions at
  • 00:26:14
    the right time and each customer has
  • 00:26:16
    their own pattern of behavior and we
  • 00:26:18
    need to talk to a bunch of customers to
  • 00:26:20
    understand what's your mental model what
  • 00:26:21
    works for you what do you understand
  • 00:26:23
    this how do I position it how do I show
  • 00:26:25
    it so what we see is every time we build
  • 00:26:30
    a new feature after we deploy it it
  • 00:26:33
    becomes oh that's so obvious but it's
  • 00:26:36
    only obvious because it's obvious for
  • 00:26:38
    the customer there's a clear pain that
  • 00:26:40
    they had and after you do it well at
  • 00:26:44
    scale it becomes obvious for everyone so
  • 00:26:47
    when we think about money platform when
  • 00:26:48
    we think about self-driving Bank when we
  • 00:26:50
    think about how we're deploying open
  • 00:26:52
    finance at scale I believe it will feel
  • 00:26:54
    a lot like that you'll feel like oh it's
  • 00:26:57
    obvious after we do it that's so
  • 00:27:00
    inspiring can't wait to see what is next
  • 00:27:04
    so thank you so much for your time today
  • 00:27:06
    vctor and thank you for watching as well
  • 00:27:09
    thank you thank you so much thank you so
  • 00:27:14
    much
Tags
  • AI
  • New Bank
  • Kundenservice
  • Kreditvurdering
  • Open Finance
  • Teknologi
  • Risici
  • Muligheder
  • Kunstig Intelligens
  • Finans