WEBVTT

00:00:00.000 --> 00:00:01.334
Marco, Fedde, welcome.

00:00:01.334 --> 00:00:04.295
Marco, AI is everywhere at the moment.

00:00:04.295 --> 00:00:06.131
So, what, and I suppose how,

00:00:06.131 --> 00:00:08.717
are companies prioritizing right now?

00:00:08.717 --> 00:00:10.010
I think there are three areas

00:00:10.010 --> 00:00:12.387
that companies are prioritizing right now

00:00:12.387 --> 00:00:13.346
that are different from one year ago.

00:00:13.346 --> 00:00:16.224
Number one is companies realize
that data is everything.

00:00:16.224 --> 00:00:19.644
Data context is very important
for the technology to work.

00:00:19.644 --> 00:00:22.480
Number two is that it's
very important to make sure

00:00:22.480 --> 00:00:26.067
that the agentic technology is
embedded in the flow of work.

00:00:26.067 --> 00:00:29.154
So then, it is totally
linked to business processes

00:00:29.154 --> 00:00:31.573
and it is not living in a different layer.

00:00:31.573 --> 00:00:34.868
And number three, safety, the trust,

00:00:34.868 --> 00:00:36.995
the compliance is very, very important

00:00:36.995 --> 00:00:40.123
to make sure that we can
rely on the AI technology.

00:00:40.123 --> 00:00:42.375
I imagine the imperative for companies now

00:00:42.375 --> 00:00:45.170
is to move from experimentation to impact-

00:00:45.170 --> 00:00:46.880
that has to be.

00:00:46.880 --> 00:00:47.923
How do they do that?

00:00:47.923 --> 00:00:49.800
What are the key moments?

00:00:49.800 --> 00:00:51.593
I think there are probably two key moments

00:00:51.593 --> 00:00:53.053
when you have a successful pilot,

00:00:53.053 --> 00:00:55.347
and you want to expand it into production.

00:00:55.347 --> 00:00:58.850
Number one is that it
implies corporate changes

00:00:58.850 --> 00:01:00.519
and organizational changes

00:01:00.519 --> 00:01:02.854
because you are inverting
your processes as a company,

00:01:02.854 --> 00:01:05.565
and you will now have human
and agents working together.

00:01:05.565 --> 00:01:06.858
So you need C-level alignment,

00:01:06.858 --> 00:01:10.904
and you need proper support,
top down-not only bottom up.

00:01:10.904 --> 00:01:13.907
And then, the second point is that

00:01:13.907 --> 00:01:17.786
more pilots moving into
production will mean more agents,

00:01:17.786 --> 00:01:20.539
and more humans working
together in a different way,

00:01:20.539 --> 00:01:21.915
so the orchestration layer

00:01:21.915 --> 00:01:24.459
to govern those relationships
is very important.

00:01:24.459 --> 00:01:27.129
Fedde, when companies are
trying to scale agentic AI,

00:01:27.129 --> 00:01:28.672
what are some of the common areas,

00:01:28.672 --> 00:01:30.716
the points in which they get a bit stuck?

00:01:30.716 --> 00:01:33.093
Yeah, most of the time, in my experience,

00:01:33.093 --> 00:01:36.221
it's not technology these days,
it's three different things.

00:01:36.221 --> 00:01:39.892
One, they underestimate
how important it is

00:01:39.892 --> 00:01:43.395
not to approach agentic
AI as a tool change

00:01:43.395 --> 00:01:45.230
but really as an operating model change,

00:01:45.230 --> 00:01:47.775
as a way to rethink how work gets done.

00:01:47.775 --> 00:01:50.944
Secondly, they don't
update the operating model

00:01:50.944 --> 00:01:52.279
to get it ready for autonomy.

00:01:52.279 --> 00:01:54.531
And so they don't have
a rich context layer,

00:01:54.531 --> 00:01:56.700
clean API environment, real-time data,

00:01:56.700 --> 00:01:58.577
modular processes, and so on.

00:01:58.577 --> 00:02:00.704
Third, they don't think
about the incentives.

00:02:00.704 --> 00:02:04.583
For instance, they
won't change the rewards

00:02:04.583 --> 00:02:06.168
for their commercial teams to

00:02:06.168 --> 00:02:09.171
be more about outcomes than activities.

00:02:09.463 --> 00:02:13.133
And, finally, what does an
AI-native company look like?

00:02:13.133 --> 00:02:15.886
Well, in this industry, we are in the
Mobile World Congress today.

00:02:15.886 --> 00:02:17.680
I think in these regulated industries,

00:02:17.680 --> 00:02:21.058
it is very important to
have deep industry context

00:02:21.058 --> 00:02:22.810
to make sure that agentic technology

00:02:22.810 --> 00:02:24.395
can drive the right impacts.

00:02:24.395 --> 00:02:25.688
Second thing I will talk about

00:02:25.688 --> 00:02:28.399
orchestration of different
agents working together.

00:02:28.399 --> 00:02:30.442
It's very important to
have a proper framework

00:02:30.442 --> 00:02:32.987
and proper technology and
tools supporting this.

00:02:32.987 --> 00:02:34.572
And last but not least,

00:02:34.572 --> 00:02:36.532
the agents need to be able to

00:02:36.532 --> 00:02:38.367
reason in an autonomous way,

00:02:38.367 --> 00:02:40.411
to make sure that the companies can switch

00:02:40.411 --> 00:02:43.414
from this reactive approach
to proactive approach,

00:02:43.414 --> 00:02:45.040
to make sure that you
can solve an incident

00:02:45.040 --> 00:02:46.375
before it happens,

00:02:46.375 --> 00:02:48.544
or you can present an offer to a customer

00:02:48.544 --> 00:02:50.171
before he asks for it.

00:02:50.171 --> 00:02:52.715
These are great, maybe
let me add a couple.

00:02:52.715 --> 00:02:55.593
One, they had moved fully from pilots

00:02:55.593 --> 00:02:58.095
to building an AI backbone,

00:02:58.095 --> 00:03:01.724
from data to common framework for agents

00:03:01.724 --> 00:03:06.062
to that context and orchestration
layer we're talking about.

00:03:06.062 --> 00:03:06.896
And then, secondly,

00:03:06.896 --> 00:03:09.649
they also moved from setting
productivity targets,

00:03:09.649 --> 00:03:10.524
taking up cost,

00:03:10.524 --> 00:03:12.568
to also thinking about how
they grow the top line.

00:03:12.568 --> 00:03:15.446
And so they get . . .
penetrate the revenue ends.

00:03:16.447 --> 00:03:18.157
Fedde, Marco, a pleasure as always.

00:03:18.157 --> 00:03:18.991
Thank you so much for your time.

00:03:18.991 --> 00:03:19.825
Thank you.
Thank you.