WEBVTT

d48aa813-2fa2-45e6-a38f-71bfc2e7b7b7-0
00:00:00.240 --> 00:00:04.040
Ernesto, welcome. Ernesto, how
is the increasing use of 

79aa9638-89fe-4be9-9f92-85c4cfa78f05-0
00:00:04.120 --> 00:00:08.050
GenAI-generated chatbots and
search engines affecting

79aa9638-89fe-4be9-9f92-85c4cfa78f05-1
00:00:08.050 --> 00:00:09.360
consumer behavior?

b1ad98de-44aa-442d-b677-cbbe6d8a6db7-0
00:00:10.120 --> 00:00:11.840
It's affecting it in a dramatic
way.

214d6561-a4b3-4ce9-b276-438348629a80-0
00:00:13.280 --> 00:00:17.253
The example I like to use is how
now consumers are going on the

214d6561-a4b3-4ce9-b276-438348629a80-1
00:00:17.253 --> 00:00:20.792
chatbots, and they're getting
the answer to the question

214d6561-a4b3-4ce9-b276-438348629a80-2
00:00:20.792 --> 00:00:24.456
there. And they don't need to
leave that environment to do

214d6561-a4b3-4ce9-b276-438348629a80-3
00:00:24.456 --> 00:00:25.760
what they need to do.

87fc130f-ce1f-4e63-aa87-38abef4b9919-0
00:00:25.800 --> 00:00:28.780
Whether that's more
information—they can ask another

87fc130f-ce1f-4e63-aa87-38abef4b9919-1
00:00:28.780 --> 00:00:31.760
question—or whether that's
making a purchase, right?

9fc82960-5c27-4a41-9564-938f27707a62-0
00:00:31.760 --> 00:00:35.080
In future, they will be able to
do all of that on the chatbot.

880ccea5-4c30-47f3-b922-d5b236087ca7-0
00:00:35.400 --> 00:00:38.822
So one of the implications of
that is going to be for

880ccea5-4c30-47f3-b922-d5b236087ca7-1
00:00:38.822 --> 00:00:39.520
publishers,

8f54e97b-350a-4885-957a-91cc254334ec-0
00:00:39.840 --> 00:00:43.251
if these consumers are not going
on their website, how they're

8f54e97b-350a-4885-957a-91cc254334ec-1
00:00:43.251 --> 00:00:44.280
going to serve ads.

2d7a92c9-c8e4-4195-9593-d05e278d5f4e-0
00:00:44.640 --> 00:00:48.200
So that's going to dramatically
change the advertising world.

71f72ecd-f3b2-4222-a3cf-a43fa72ff30f-0
00:00:48.520 --> 00:00:50.000
Let's explore that a little bit
more—the

250532ae-b462-43e0-a046-3740a8c3ec3e-0
00:00:50.000 --> 00:00:53.596
impact that all this will have
on brands, particularly in the

250532ae-b462-43e0-a046-3740a8c3ec3e-1
00:00:53.596 --> 00:00:54.640
advertising space.

60ea9948-1c75-4ba6-b744-bc5d1c4cf91b-0
00:00:55.320 --> 00:00:58.440
So not all brands will be
impacted in the same way.

a5d1ff2c-1bd2-4d78-b375-9c2c95f26c5f-0
00:00:59.000 --> 00:01:02.805
We expect that the brands that
rely the most on search for

a5d1ff2c-1bd2-4d78-b375-9c2c95f26c5f-1
00:01:02.805 --> 00:01:07.062
their performance marketing will
be impacted the most. And not in

a5d1ff2c-1bd2-4d78-b375-9c2c95f26c5f-2
00:01:07.062 --> 00:01:10.932
the same way, of course, and
depends on what type of search

a5d1ff2c-1bd2-4d78-b375-9c2c95f26c5f-3
00:01:10.932 --> 00:01:12.480
the consumers are doing.

d37a767f-0dba-43f1-8411-6b705d798d92-0
00:01:12.880 --> 00:01:16.480
So we just published a report
where we talk about the journey

d37a767f-0dba-43f1-8411-6b705d798d92-1
00:01:16.480 --> 00:01:19.907
of a consumer, the different
phases, and we think that the

d37a767f-0dba-43f1-8411-6b705d798d92-2
00:01:19.907 --> 00:01:23.391
brands that rely the most on
what we call mid-funnel search

d37a767f-0dba-43f1-8411-6b705d798d92-3
00:01:23.391 --> 00:01:26.992
need to change how they allocate
their advertising dollars to

d37a767f-0dba-43f1-8411-6b705d798d92-4
00:01:26.992 --> 00:01:30.593
shift them more and more towards
chatbots and new advertising

d37a767f-0dba-43f1-8411-6b705d798d92-5
00:01:30.593 --> 00:01:31.000
models.

5cbcd902-ebb9-43d2-89d2-c01eeac4b1b4-0
00:01:31.400 --> 00:01:35.094
What other ways can brands
prepare for the shift that's

5cbcd902-ebb9-43d2-89d2-c01eeac4b1b4-1
00:01:35.094 --> 00:01:36.480
coming down the line?

19506962-0395-4682-a187-8d526e9402c5-0
00:01:37.200 --> 00:01:41.082
It's early days, so I cannot
tell you that there is a one

19506962-0395-4682-a187-8d526e9402c5-1
00:01:41.082 --> 00:01:43.560
silver bullet to address this
shift.

78329a34-f372-4a52-b150-5c64e353faf7-0
00:01:43.840 --> 00:01:45.200
There are few things that can be
tried.

ef88d9d8-cbe0-4793-8b70-b827c9a6a652-0
00:01:45.480 --> 00:01:49.310
One is understand how the LLMs
are digesting information, so

ef88d9d8-cbe0-4793-8b70-b827c9a6a652-1
00:01:49.310 --> 00:01:53.267
that your products and services
are available to the LLMs, and

ef88d9d8-cbe0-4793-8b70-b827c9a6a652-2
00:01:53.267 --> 00:01:57.161
they're served in the right way
to the consumers when they're

ef88d9d8-cbe0-4793-8b70-b827c9a6a652-3
00:01:57.161 --> 00:01:58.480
doing their searches.

777c261f-9059-495e-9dd4-8a862b4c7ba0-0
00:01:58.960 --> 00:02:02.560
The second way is also to do
partnerships with these LLMs so

777c261f-9059-495e-9dd4-8a862b4c7ba0-1
00:02:02.560 --> 00:02:06.338
that you can make some of your
content available in an easy way

777c261f-9059-495e-9dd4-8a862b4c7ba0-2
00:02:06.338 --> 00:02:08.640
to the LLMs, but maybe not
everything.

9826d1d4-6adc-43bf-a255-bf6c69abd270-0
00:02:09.120 --> 00:02:12.560
So one of the things we learned
is that LLMs need a lot of data

9826d1d4-6adc-43bf-a255-bf6c69abd270-1
00:02:12.560 --> 00:02:14.120
like text, but also pictures.

496de66b-7447-4678-b245-64f763395b65-0
00:02:14.680 --> 00:02:17.643
And we are working with a
company that is providing the

496de66b-7447-4678-b245-64f763395b65-1
00:02:17.643 --> 00:02:21.030
text, but keeping the pictures,
so that the consumers are still

496de66b-7447-4678-b245-64f763395b65-2
00:02:21.030 --> 00:02:23.200
going on the website to see the
product.

7061c792-0656-40e4-8d82-607883a10854-0
00:02:23.520 --> 00:02:26.232
So figuring out ways to
collaborate so that you're

7061c792-0656-40e4-8d82-607883a10854-1
00:02:26.232 --> 00:02:28.200
giving something but not
everything.

8b7e4c5d-cb44-4f61-9466-98468b81859b-0
00:02:28.640 --> 00:02:32.105
Then, obviously, as these new
advertising formats are

8b7e4c5d-cb44-4f61-9466-98468b81859b-1
00:02:32.105 --> 00:02:36.276
available—and, obviously, there
are some chatbot that are trying

8b7e4c5d-cb44-4f61-9466-98468b81859b-2
00:02:36.276 --> 00:02:39.999
advertising—learn, test, see
what works, see what doesn't

8b7e4c5d-cb44-4f61-9466-98468b81859b-3
00:02:39.999 --> 00:02:40.320
work.

3ab2a165-4438-407e-8079-6ed61836d4f0-0
00:02:40.760 --> 00:02:44.655
And then, specifically for some
brands where the product line is

3ab2a165-4438-407e-8079-6ed61836d4f0-1
00:02:44.655 --> 00:02:48.311
complex so there is a lot of
products, I think experimenting

3ab2a165-4438-407e-8079-6ed61836d4f0-2
00:02:48.311 --> 00:02:51.967
with the proprietary LLM that is
engaging with consumers and

3ab2a165-4438-407e-8079-6ed61836d4f0-3
00:02:51.967 --> 00:02:55.383
explaining to them the products
and giving them a better

3ab2a165-4438-407e-8079-6ed61836d4f0-4
00:02:55.383 --> 00:02:58.320
experience, that could be
something also to try.

78c1c8fe-61ab-4a74-9ec9-ede20311707c-0
00:02:58.720 --> 00:02:59.800
Ernesto, a pleasure as always.

950a6d11-f31f-405f-8400-2bb9a79cfd64-0
00:02:59.800 --> 00:03:00.240
Thank you.

3e972d1c-c1ad-4267-a8f0-83c7dd454e0c-0
00:03:00.480 --> 00:03:00.920
Thank you.