Home
Entrar
Registrar
Loading...
How To Drive Ultimate Value From AI By Embedding It In Core Business Applications - Video học tiếng Anh
Praticar Escuta
Ouvir
/
Video
/
Business Insider
/
How To Drive Ultimate Value From AI By Embedding It In Core Business Applications
How To Drive Ultimate Value From AI By Embedding It In Core Business Applications
Selecionar modo de aprendizagem:
Ver legendas
Escolher palavra
Reescrever palavra
Highlight:
4000 IELTS Words
3000 Oxford Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Legendas (547)
0:00
Hello and welcome back to Beyond
0:03
Resilience, our series of conversations
0:06
exploring what is next for enterprise
0:09
success and adaptability.
0:16
Our guest today brings the perfect
0:19
perspective to navigate these changes.
0:22
From my 25 years of experience
0:25
implementing AI, I have learned that
0:28
staying ahead requires three key
0:31
elements.
0:33
The first one is technical expertise.
0:37
Then a human centric focus and finally
0:41
an entrepreneurial mindset.
0:44
Stefan Deus from SAP embodies exactly
0:48
that mix.
0:50
Stefan, welcome to the show.
0:52
>> Yeah, thank you. You know, excited to be
0:54
here today.
0:56
>> Could you tell more to our audience
0:59
about you and what gets you excited? I
1:01
>> mean, I'm Stefan uh Dutch uh I'm the
1:04
president of the global business at SAP.
1:07
So in essence that means you know the
1:08
global go to market of our application
1:11
portfolio our data data platform and our
1:14
AI capabilities is where I hold a global
1:17
responsibility for and you know today
1:19
we'll deep dive in in the business how
1:21
it comes to life you know what are
1:23
practical examples and also what are
1:24
some of the innovation that we are
1:26
seeing what gets me excited is is
1:29
applying technology to deliver business
1:31
value to our client base because
1:33
ultimately you know tech for the sake of
1:35
tech is not So exciting, but the moment
1:38
it has, you know, true impact, you know,
1:41
for our customers and they see that
1:42
impact back in their business results,
1:45
that gets me excited.
1:46
>> Yes. I love this. And I think, yeah,
1:48
business imperatives and business impact
1:51
should be driving
1:53
uh any action with technology. And
1:55
that's that's exactly what you said.
1:57
>> Yeah. No, 100%. And look, you know, that
1:59
that's, you know, with technology, we're
2:02
often, you know, talking about, you
2:04
know, AI and new innovation, but it
2:06
should never be for the sake of
2:08
innovation. It's all about, okay, but
2:10
how does that actually help you solve a
2:12
particular business challenge that an
2:14
organization is facing? And then if you
2:16
solve that challenge, you know, can we
2:18
see that back in the P&L or the balance
2:20
sheet in terms of, you know, true value
2:22
delivered. So yeah, that is what it is
2:25
all about.
2:25
>> Completely makes sense. You have an
2:29
entrepreneurial journey and you've been
2:32
at different startups and this is quite
2:35
remarkable because you've been driving
2:37
revenue, you've been driving rapid
2:39
scaling and such transformations that
2:42
you've driven uh require incredible
2:46
resilience and you know we talk a lot
2:48
about resilience in in this show.
2:50
>> Yeah. How has that experience influenced
2:53
you uh in the way you help your
2:56
customers today at SAP?
2:59
>> No, absolutely. You look the you know
3:00
the the first you know 10 years of my
3:02
career I was in finance and supply chain
3:04
and then the past decade in tech and
3:07
especially the past seven years sort of
3:09
going from scale up or startup actually
3:12
you know 50 people couple of clients to
3:14
then scale up with more than 3,000
3:16
people was just an you know interesting
3:19
experience and then coming into an
3:20
organization like SAP which obviously
3:23
you know well established uh you know
3:25
very loyal customer base um you know an
3:28
amazing product portfol folio good
3:30
people uh was just interesting to see
3:32
the difference between you know uh
3:34
startup scaleup phase to you know
3:36
company like SAP but I I would have to
3:38
say I find that the way we operate and
3:42
the way we think there's a lot of things
3:44
from the startup scaleup scene that I'm
3:46
seeing and the reason is all about the
3:49
innovation right there's massive
3:50
innovation around data there's massive
3:52
innovation around AI agentic things like
3:55
that so how do you then apply a startup
3:57
mindset to then solving you know
4:00
problems but do that you know at pace do
4:02
that jointly with customers uh and make
4:05
sure that you know we continue then to
4:07
deliver so I find um it is different but
4:11
also actually there's a lot of
4:13
commonalities in terms of how we think
4:15
and you know how we how we drive success
4:17
for our customers
4:18
>> the mindset the mindset of of a startup
4:21
within a big company
4:23
>> for sure
4:23
>> to for innovation and solving problems
4:25
that clients have
4:27
>> yeah and ultimately look at The startup
4:28
it's simple, right? You have an idea and
4:30
you say, "I want to solve this problem
4:32
for you." And you say, "Hey, can I help
4:34
you solve this problem?" So, we solve
4:35
this problem for you. You like it. You
4:37
tell your neighbor, you know, and and we
4:39
try to, you know, go around, you know,
4:42
and tell some other people. But
4:44
ultimately what you often see in
4:46
startups, it is that customer obsession,
4:49
right? That is an obsession of solving a
4:51
problem. And then the moment you do
4:52
that, well, obviously, you know, startup
4:55
becomes scale up, scale up becomes a big
4:56
enterprise. But I think now what we are
4:59
seeing given the speed of innovation the
5:02
only way for a big company right to to
5:05
keep actually operating like they do
5:08
today is to you know get into that
5:11
mindset of you know caring deeply about
5:15
you know customers caring deeply about
5:17
solving their challenges and then making
5:19
sure that you know you apply the latest
5:21
technology to do so and we'll talk more
5:23
about that but I think I love it
5:25
>> you know that's that's how both worlds
5:26
are coming together in my mind
5:28
>> a craft mind shift
5:30
uh at scale.
5:31
>> Yeah. Yeah.
5:33
>> Iteration iterations with clients trying
5:36
and making sure you you get to to fit
5:38
their needs at the end of the day.
5:39
>> Yeah. 100%.
5:40
>> Uh with the flexibility and the mindset
5:42
of an entrepreneur. Love it.
5:44
>> Uh you shared a bit about SAP business
5:47
just now. Um
5:49
>> and that being a pathway to resilience
5:52
and adaptability.
5:54
So, it's a big topic, you know, here in
5:56
our conversations on on this show. Um,
5:59
and you like to use a phrase about it
6:01
that I will always remember. You talk
6:04
about best of sweets rather than best of
6:07
breed. Can you tell us more about this?
6:09
>> Yeah. Yeah, absolutely. And ultimately,
6:11
you know, let me break down the business
6:14
because the way we envision the business
6:17
is a flywheel
6:19
of apps, data, and AI coming together.
6:22
Now, why a flywheel? You know, in
6:24
physics, a flywheel is all about
6:26
individual components that are coming
6:28
together to create new energy
6:29
>> and we believe that holds true in
6:31
enterprise software as well. Now, it
6:33
starts with the application layer. So,
6:36
the apps that we provide uh run the end
6:39
to end business processes of an
6:41
organization. So, think about you know
6:43
there's the core ERP but then there's
6:44
finance, there's supply chain, there's
6:47
procurement, there's you know customer
6:49
experience, there's HR and so forth. So
6:51
every single core business process is
6:54
run in an application. If I just take
6:56
supply chain, if you would double click
6:57
on that, there is planning, there is
7:00
manufacturing, there is logistics, there
7:02
is warehousing, there is asset
7:04
management and so forth. So SAP has the
7:07
broadest portfolio of apps, you know, of
7:10
of all software companies in the world.
7:12
But these apps do not only run end to
7:15
end business processes. They also
7:17
generate a lot of extremely valuable
7:19
data. Now that data is what we combine
7:24
with non SAP data in what we call the
7:27
business data cloud that does a couple
7:29
of things. The first thing it builds a
7:32
reliable and trusted source to do
7:34
anything AI related.
7:36
>> Mhm.
7:36
>> And second is often times why does AI
7:39
fail or why do transformations fail
7:42
because companies haven't solved for the
7:44
data problem. So I think you know that
7:46
is extremely important that there has to
7:48
be you know a data foundation that then
7:51
supports organizations when it comes to
7:53
AI. Now what is extremely important is
7:56
the notion of what we call semantically
7:59
rich data
8:00
>> because it's very easy to pull data from
8:02
different systems but you lose the
8:04
semantics. So give you an example simple
8:06
a purchase order or a sales order. If
8:09
you would have to pull that from SAP
8:10
systems you now need to pull data from
8:13
you know 15 different tables but then
8:15
you need to recreate the logic that that
8:18
is a sales order purchase order or you
8:20
know anything else. So often times
8:23
companies spend an enormous amount of
8:25
time and effort in pulling data from
8:27
different applications trying to
8:29
recreate the semantics but now the data
8:31
is disconnected from the source
8:33
application. the data is disconnected
8:35
from the end to-end business process and
8:38
again it it takes an enormous amount of
8:40
time on getting getting it ready for AI.
8:43
>> So apps data then comes AI. Now what I
8:47
found very interesting you might have
8:49
seen that uh MIT they uh released this
8:54
uh article and said hey 95% of Gen AI
8:57
pilots fail.
8:58
>> Everybody talks about it.
8:59
>> There's a couple of things to that that
9:00
are interesting. I think a lot of
9:02
companies are trying to treat AI like a
9:07
separate layer somewhere in the
9:08
technology stack. So it's disconnected
9:10
from your end to end business processes,
9:12
it might be disconnected from your data
9:14
strategy, it's something you know
9:16
somewhere but the moment it doesn't make
9:19
it back to the end to end business
9:20
process context is very very difficult
9:23
to drive value. So one of the strategies
9:25
from SAP from from early on when it
9:28
comes to AI is to embed AI back in the
9:32
core application. Now think about this
9:35
in an you know finance you want to get
9:38
to autonomous closing and you want AI to
9:41
help you book acrals. So that is an
9:43
activity that needs to make it back to
9:45
the core application. In supply chain,
9:48
you want to create AI to predict future
9:51
demand so that you can then synchronize
9:53
you know manufacturing and things like
9:55
that has to make it back to the core
9:57
application. Procurement uh think about
10:00
you know how do I get to autonomous
10:02
sourcing has to make it back to the core
10:04
application. So we do not believe that
10:07
there is some wrapper somewhere out
10:09
there. It has to come back to the to the
10:11
core application and that's why we are
10:13
saying there's an app layer there's a
10:15
data layer and then there's embedded AI
10:17
in the application and agents that sort
10:19
of come together to deliver a much
10:22
better sort of enterprise technology
10:24
experience to our customers but also
10:27
very much focused on you know solving
10:29
some of the business challenges. So the
10:30
insight that is created at AI level
10:32
comes down back to the app level where
10:35
people are working currently and people
10:38
can see this insight and and work with
10:40
this insight in mind and yes
10:42
>> and and make better decisions.
10:44
>> Yes. And those three components of the
10:46
flywheel always have to work together.
10:47
Maybe give you an example working
10:49
capital. So let's say you're the CFO uh
10:52
of a company
10:54
>> and um I work in your team and you would
10:57
ask me you know Monday morning Stefan
11:00
you know how's our cash conversion cycle
11:02
doing and at that moment in time I
11:05
should be able to use software to
11:07
immediately tell you hey cash conversion
11:09
cycle is going up so your follow-up
11:11
question hey Stephan that's not great
11:14
why is it going up so I should now at my
11:16
fingertips have information is this
11:18
inventory is this payables Is this
11:20
receivables? Now let's say it's
11:22
receivables. These I would say are
11:24
insights that you today might be able to
11:26
get from some reports. But now the
11:28
question is what do I do with the
11:30
insights? So let's say it's receivables.
11:32
Your follow-up question then is okay
11:33
Stefan where? So let's say in North
11:36
America I have a segment of customers
11:38
that is paying me late or is at the risk
11:42
of actually not paying me at all. Then
11:44
your question is okay Stefan will this
11:47
continue? What do I do about it? So now
11:49
it comes interesting. Now you need
11:51
historical payment information. You
11:53
might want to pull some data, interest
11:55
rates, GDP growth, um you know some
11:59
credit ratings from external agencies.
12:01
Then you need to apply an algorithm
12:04
>> to then predict cash conversion cycle
12:06
going forward. Now assume we do all of
12:09
that and the cash conversion cycle is
12:10
predicted to go up. Then your question
12:12
is okay what do I do about it?
12:14
Assurance, some factoring, whatever it
12:17
is. And that is how the flywheel should
12:19
come together right where apps working
12:22
capital app comes together with data SAP
12:26
data non SAP data then apply AI to make
12:29
a prediction and based on that
12:31
prediction we want to act and we want to
12:33
act now so I think you know that is how
12:35
that flywheel comes together in an
12:37
endto-end business context
12:38
>> what do you see as other trends more
12:42
specifically in AI uh for the clients
12:44
that you you're supporting
12:46
>> yeah so first and foremost I I mean the
12:48
embedded AI that makes it back to the
12:51
core application uh is something that
12:53
clients really really appreciate because
12:55
they also understand that this sort of
12:57
wrapper on top uh I mean by MIT research
13:01
right is is not really working. So I
13:03
think embedded is one.
13:04
>> Second is there is a lot of excitement
13:08
around agents.
13:09
>> Yes. Because agents can not only you
13:13
know do a whole series of activities in
13:16
an autonomous way but they can connect
13:19
one business process with another one.
13:21
So they can also you know break down
13:24
silos. So give an example there's an
13:27
agent on the commercial side that can
13:29
help us predict which deals we will
13:32
close and which deals will not or what
13:34
we should do to actually get those deals
13:37
closed so that we can increase in sales.
13:39
But if we increase sales, we need to
13:41
make sure that we can actually deliver.
13:43
So human is a manufacturing company. We
13:45
need to have manufacturing capacity. We
13:48
need logistics and warehousing capacity.
13:50
We need you know raw materials or
13:53
components uh readily available. We need
13:55
to make sure that our suppliers can
13:57
deliver those and so forth. So if you
13:59
just think about an entire value chain
14:01
from getting a product in the hands of a
14:03
customer to sourcing the components that
14:07
has to be orchestrated by a series of
14:10
agents that can help you know
14:13
organizations get to you know better
14:16
decisions and improve business results.
14:18
So what we are seeing is an increase in
14:22
the need for agents. uh customers want
14:24
to work together with us to get there
14:26
but they also understand that this has
14:28
to be uh sort of across you know
14:31
business processes and I think the
14:33
excitement around AI when it comes to
14:35
you know productivity uh that's sort of
14:38
fading right yes we are all using the
14:40
different LLMs and yes you know we all
14:43
get to information you know faster but
14:46
an LLM per se or you know the sort of
14:50
productivity AI use cases are not
14:52
necessarily necessarily impacting
14:53
topline. They're not necessarily
14:55
impacting you know working capital of an
14:58
organization and they're also uh not
15:00
necessarily impacting the cost profile
15:01
of an organization. So now the key
15:03
question is and that's you know where we
15:06
are in the forefront of you know working
15:08
together with our customers to deliver
15:10
aic experience across different business
15:12
processes that then ultimately leads to
15:16
you know proper you know business
15:18
results in in a pen of a company.
15:20
>> I love it. So it's about um uh
15:24
collapsing all those silos that we've
15:26
seen in companies from I mean typically
15:28
traditionally from for centuries
15:31
and having those agents working across
15:34
those different functions
15:36
um they don't know silos they are agents
15:39
and% the other big trend in AI UIUX is
15:44
changing fundamentally
15:46
so you know a lot of organizations they
15:48
have you know thousand 2,000 different
15:50
applications and users you know logging
15:53
into the applications and doing all
15:55
kinds of work but that is not how we
15:59
interact with software in the future so
16:01
we have a capability that we call dual
16:04
at SAP so see that as a new UIUX the
16:07
super orchestrator it's it's a you know
16:09
check GPT perplexity like sort of
16:12
interface
16:13
>> so you ask questions but you also give
16:16
instructions but now you as an as a You
16:21
don't have to log into five different
16:22
applications to do something that is
16:26
being orchestrated by Juul. So the way
16:28
we start thinking about interacting with
16:31
software becomes different. Stefan, I've
16:33
seen so many companies struggle with
16:36
legacy software with on premises systems
16:40
and the pain is real. I mean slow
16:42
processes, high cost, constant
16:44
breakdowns. How can SAP solve for that?
16:48
>> Yeah, absolutely. So obviously you know
16:50
a lot of our customer base you know
16:52
understands that in order to consume
16:54
innovation at pace and at scale you have
16:58
to you know bring your core applications
17:01
and your data capabilities in the cloud.
17:04
>> Uh so there is you know long list of
17:07
clients that understood that you know
17:10
that went early and are now the ones
17:13
that can actually consume you know a lot
17:15
of the AI innovation. Now obviously
17:17
there's still a group of clients that is
17:19
still on prem and still not in the
17:21
cloud. So obviously the way uh to get
17:24
there is to move to the cloud and in my
17:27
mind it's like okay you know let's get
17:28
the core foundation to the cloud and
17:30
then you know there's a lot a lot of
17:32
innovation uh that we can work on
17:35
together.
17:35
>> Stefan all this is very exciting but big
17:39
question comes to my mind which is what
17:41
is next? What's next to business suite?
17:44
What's in your mind in your vision? We
17:46
believe that the app layer will be
17:50
commoditized.
17:51
>> Yes. Now very very similar to actually
17:54
the infrastructure layer while critical
17:57
uh got commoditized by the cloud we
17:59
believe the app layer will be
18:00
commoditized by AI that means
18:04
>> especially by agents
18:05
>> I mean ultimately from an end user
18:07
perspective yes we would still need apps
18:10
to orchestrate the workflow and the end
18:11
to end process but the value sits in the
18:15
data and in the AI layer. So the future
18:17
of the business suite will be we commit
18:20
to have the best possible embedded AI
18:22
experience you know in every single
18:24
application. Um you know we have around
18:27
400 use cases already embedded in the
18:30
core application.
18:32
>> Uh we absolutely want to be a winner in
18:34
the agent space. Uh we also believe that
18:38
that is something that we will do more
18:39
and more with partnerships with other AI
18:42
companies. uh and then we absolutely
18:44
envision that that front end user you
18:48
know experience uh will be conducted
18:51
more and more on duel uh so we believe
18:54
it is extremely exciting times to be at
18:56
the company uh we're also you know
18:59
getting amazing feedback from clients
19:00
that are working with us you know hand
19:02
in hand to truly embark on this business
19:06
v journey and then ultimately I you know
19:08
we started this conversation it is not
19:11
technology for the sake of techn
19:12
technology. It's not an agent for the
19:14
sake of an agent. It is AI innovation to
19:18
deliver a better outcome for the
19:20
businesses that we serve. And I think
19:22
that is really uh the future and how we
19:25
envision that is yes we are doing all
19:27
those things but ultimately we do that
19:30
because we want to deliver you know
19:32
better topline you know reduce cost and
19:34
improve working capital for our customer
19:36
base.
19:37
>> It totally makes sense. Thank you
19:40
Stefan. Moving beyond resilience means
19:44
looking toward the future.
19:47
And as we close, what is the one message
19:50
you would like to leave the audience
19:52
with about the possibilities of business
19:55
street for their own businesses? Yeah,
19:58
the main message is it is absolutely
20:01
possible and what we are finding from
20:04
obviously you know dealing with you know
20:06
four 400,000 plus customers that we have
20:09
is the customers that have this curious
20:13
mindset that you know throws you know
20:15
these big hairy problems that say hey
20:18
you know I want to reimagine you know
20:20
how I conduct my business they're
20:22
absolutely you know seeing the benefits
20:24
and are you know full on the journey at
20:26
the same time there's obviously also a
20:28
group of customers that is you know
20:30
slower when it comes to adoption of new
20:33
technology maybe more conservative
20:36
um but but I do believe that we are
20:38
going to see a divide between companies
20:41
that say hey I'm going for this you know
20:44
I need to innovate I need to change the
20:46
way I conduct business and I have a very
20:49
very strong point of view that they will
20:50
definitely be winners in the marketplace
20:52
so I think the key message that I would
20:54
have is hey innovation is
20:57
AI is happening.
20:58
>> Mhm.
20:59
>> You know, be part of it.
21:01
>> I love this. That's really a powerful
21:04
message and and a message that I believe
21:08
every leader should carry with them as
21:10
they plan for for the future of their
21:12
business.
21:13
Stefan, thank you for joining us. uh
21:17
your examples made it clear how SAP
21:20
turns values like innovation and agility
21:23
into real practical outcomes for
21:26
businesses.
21:28
>> We appreciate it and um you know great
21:30
being here and great catching up with
21:32
you.
21:32
>> I also want to thank you for joining us
21:34
for these conversations. I hope you have
21:37
learned as much as I have and that you
21:40
live with a clear vision of adaptability
21:42
built not just on resilience but on
21:45
growth. Thank you.
