WEBVTT

1
00:00:00.040 --> 00:00:05.601
A whole class of content that maybe
marketers thought AI wasn't going to get to

2
00:00:05.601 --> 00:00:09.443
So the upstream, the big brand campaigns,
the TV ads, you know, - yeah, yeah,

3
00:00:09.443 --> 00:00:12.203
we'll use Gen AI for adaptation and
bottom of funnel.

4
00:00:12.203 --> 00:00:15.645
But top of funnel will stick with our current providers.

5
00:00:15.645 --> 00:00:16.965
And we'll be done in the traditional way.

6
00:00:17.205 --> 00:00:18.405
No it won't.

7
00:00:18.405 --> 00:00:21.246
Everything is going to be done with AI,
top funnel, bottom of funnel.

8
00:00:21.446 --> 00:00:22.727
And I think video is changing that.

9
00:00:24.087 --> 00:00:25.808
We're joined by Will from Pencil.

10
00:00:25.808 --> 00:00:26.928
Thanks very much for being here.

11
00:00:26.928 --> 00:00:27.929
Thank you for having me.

12
00:00:28.089 --> 00:00:30.649
Will,
so many marketing leaders are still

13
00:00:30.649 --> 00:00:32.250
struggling to scale GenAI.

14
00:00:33.210 --> 00:00:36.051
What are the biggest problems you're
seeing that are holding them back?

15
00:00:36.051 --> 00:00:39.972
So I think one of the most interesting
things is how the problems have changed

16
00:00:39.972 --> 00:00:43.494
and how quickly the problems marketing
leaders are facing is changing.

17
00:00:43.974 --> 00:00:47.975
Maybe a year ago,
this was very much a technology risk thing.

18
00:00:47.975 --> 00:00:49.135
Was the tech good enough?

19
00:00:49.135 --> 00:00:50.576
Was the creative outputs of GenAI

20
00:00:50.576 --> 00:00:53.697
good enough,
were the legal and governance boundaries

21
00:00:53.697 --> 00:00:56.738
of agreements, and that kind of thing,
good enough?

22
00:00:57.818 --> 00:01:01.780
A year on some of those problems,
many of those problems are actually

23
00:01:01.780 --> 00:01:06.421
solved or considered solved by marketing
leaders and it becomes more operational,

24
00:01:06.421 --> 00:01:08.942
it becomes more human,
more emotional even.

25
00:01:09.342 --> 00:01:12.463
And how do you build a modernized
creative workflow,

26
00:01:12.463 --> 00:01:14.304
especially at enterprise level?

27
00:01:14.944 --> 00:01:17.785
So there's the technical answer and then
there's the human answer.

28
00:01:17.785 --> 00:01:18.585
Give us both.

29
00:01:18.745 --> 00:01:22.386
Yeah,
I think the human answer is you build it

30
00:01:22.386 --> 00:01:24.187
with clarity of vision.

31
00:01:24.267 --> 00:01:26.348
So know what you're trying to achieve
from the beginning.

32
00:01:26.588 --> 00:01:27.948
You build it iteratively.

33
00:01:27.948 --> 00:01:29.469
So you start with the easy bits.

34
00:01:29.469 --> 00:01:31.669
You start with the bits that people are
excited about.

35
00:01:32.189 --> 00:01:37.311
You start with the obvious bits.
You go in knowing that you may need to

36
00:01:37.311 --> 00:01:40.272
start again,
that you may need to kill things.

37
00:01:40.272 --> 00:01:44.273
Very experimental,
but not blindly experimental.

38
00:01:44.273 --> 00:01:47.234
Experimenting towards the North Star,
towards the goal.

39
00:01:47.755 --> 00:01:52.196
I think the technical answer to how you
build a modern creative workflow, well,

40
00:01:52.196 --> 00:01:56.557
we obviously feel strongly that with
GenAI, it's better for a client to adopt,

41
00:01:56.557 --> 00:01:59.998
a single technology that's very customizable.

42
00:01:59.998 --> 00:02:03.520
A platform layer that aggregates all of
the best models,

43
00:02:03.520 --> 00:02:06.921
integrates with all the media channels,
but is very customizable to their organization.

44
00:02:06.921 --> 00:02:10.562
And then provide that to their creative
partners, their agencies,

45
00:02:10.562 --> 00:02:15.043
their internal teams to use, rather than
waiting to see what bubbles up across the

46
00:02:15.043 --> 00:02:17.284
industry and what folks prefer.

47
00:02:17.884 --> 00:02:19.765
Because that introduces risk, right?

48
00:02:19.885 --> 00:02:24.006
Where does the value accrue to,
to the partner or to the client?

49
00:02:24.006 --> 00:02:27.727
What are the legal guardrails, the safety,
how much thinking has gone into it?

50
00:02:27.727 --> 00:02:29.168
Clients want that visibility.

51
00:02:29.368 --> 00:02:35.290
So I think how you build a new workflow
starts with such a foundational thing as

52
00:02:35.290 --> 00:02:38.731
the structure of the platform that you
choose to work with.

53
00:02:39.211 --> 00:02:43.373
What's the next frontier for AI powered
creative workflows?

54
00:02:43.653 --> 00:02:45.853
Yes,
I'm going to say what everyone says at

55
00:02:45.853 --> 00:02:48.574
the moment, which is video and agents.

56
00:02:49.014 --> 00:02:53.776
And I guess what I'd like to try and do
is go a little bit deeper maybe than what

57
00:02:53.776 --> 00:02:56.457
you read in the media about video and
agents.

58
00:02:56.777 --> 00:03:03.619
So I think what's interesting about video
first, that's obvious, is two things.

59
00:03:03.619 --> 00:03:06.140
One is unlocking.

60
00:03:06.460 --> 00:03:09.701
The other thing that video is doing,
I think, is bringing craft,

61
00:03:09.701 --> 00:03:13.422
narrative and storytelling into the GenAI
conversation in a way that text

62
00:03:13.422 --> 00:03:16.023
generation and image generation didn't
bring it in.

63
00:03:16.383 --> 00:03:20.385
And I think that really opens the door to
creativity in a way that text and image

64
00:03:20.385 --> 00:03:21.905
didn't. Video brings that back.

65
00:03:22.105 --> 00:03:26.707
So I've never seen more creative leaders
more excited about GenAI today when

66
00:03:26.707 --> 00:03:31.188
Google has announced VEO2 and VEO3 and
where video is starting to become the

67
00:03:31.188 --> 00:03:31.548
focus.

68
00:03:31.868 --> 00:03:33.229
So that's video.

69
00:03:33.309 --> 00:03:34.349
I think that's very exciting.

70
00:03:34.749 --> 00:03:38.951
Agents, you know,
I'm the co-founder of Pencil.

71
00:03:38.951 --> 00:03:39.791
We have a mission.

72
00:03:40.471 --> 00:03:43.952
We believe that AI should be a tool for
creative people and not a replacement.

73
00:03:44.272 --> 00:03:47.113
So agents actually for me was a bit,
oh God, right?

74
00:03:47.113 --> 00:03:49.714
It's AI using tools, not AI being a tool.

75
00:03:50.915 --> 00:03:55.036
So we've had to think very carefully
about how we want to make agents real in

76
00:03:55.036 --> 00:03:55.676
our product.

77
00:03:57.117 --> 00:04:01.118
I think that it's important for us to
think about agents not as plumbing, right?

78
00:04:01.118 --> 00:04:05.479
It's not like I have a team of ten people
and now I have ten agents that replace

79
00:04:05.479 --> 00:04:06.159
those people.

80
00:04:07.040 --> 00:04:10.001
You know, an agent for brief writing,
an agent for strategy,

81
00:04:10.001 --> 00:04:12.922
an an agent for copywriting,
an agent for media management.

82
00:04:13.162 --> 00:04:16.603
Like yes, those agents might exist,
but it's not a people replacement thing.

83
00:04:16.963 --> 00:04:20.684
So the way we think about agents is,
the people that are doing those jobs do

84
00:04:20.684 --> 00:04:23.285
those jobs today,
but they work with an agent who has

85
00:04:23.285 --> 00:04:25.486
access to data at each stage in the
workflow.

86
00:04:25.926 --> 00:04:30.087
The way that I think we'd like marketers
and brands to think about agents isn't

87
00:04:30.087 --> 00:04:33.888
plumbing together agents that do things
and then setting them off to run

88
00:04:33.888 --> 00:04:34.889
autonomously. More,

89
00:04:35.049 --> 00:04:38.490
I'm working with teams that I already
have who are each working with an

90
00:04:38.490 --> 00:04:39.370
appropriate agent.

91
00:04:39.650 --> 00:04:44.332
And that person is now super powered and
has capacity they didn't have before.

92
00:04:44.612 --> 00:04:47.093
So it's how you attack this idea of
agents.

93
00:04:47.093 --> 00:04:50.854
So it'll enhance and augment the work
that's being done by people rather than

94
00:04:50.854 --> 00:04:51.494
replacing it.

95
00:04:51.614 --> 00:04:51.974
Absolutely.

96
00:04:51.974 --> 00:04:53.135
That's a very hopeful message.

97
00:04:53.135 --> 00:04:54.895
Will, thanks very much for joining us here

98
00:04:55.015 --> 00:04:55.415
In Cannes.

99
00:04:55.616 --> 00:04:55.896
Thank you.