WEBVTT

1
00:00:00.040 --> 00:00:04.240
I'm Manveen Rana here at Davos
2024, and I'm joined by Vlad

2
00:00:04.240 --> 00:00:08.560
Lukic, who is the Global Lead on
Tech and Digital Advantage.

3
00:00:08.560 --> 00:00:09.480
Vlad, welcome.

4
00:00:10.160 --> 00:00:11.720
Thanks for joining us.

5
00:00:12.720 --> 00:00:15.120
Everybody here is talking about
AI this year.

6
00:00:16.280 --> 00:00:18.680
As you know, we have for many
years, but this year in

7
00:00:18.680 --> 00:00:21.360
particular, it feels like
everything is changing. For you,

8
00:00:21.680 --> 00:00:24.880
how much are you seeing
companies and leaders taking on

9
00:00:24.880 --> 00:00:25.600
that change?

10
00:00:25.600 --> 00:00:28.840
And for you, are there surprises
about how well or how badly

11
00:00:28.840 --> 00:00:29.600
they're doing?

12
00:00:30.240 --> 00:00:31.280
Yeah, it's hard to like,

13
00:00:31.400 --> 00:00:32.680
it's too soon to judge, right.

14
00:00:32.680 --> 00:00:35.880
I like to say that we are really
in the early stages of AI.

15
00:00:35.880 --> 00:00:36.760
It's been developed.

16
00:00:36.760 --> 00:00:38.800
It has been developed for the
last 50 plus years.

17
00:00:38.800 --> 00:00:39.840
So it's continuation.

18
00:00:39.840 --> 00:00:42.640
But now we're entering the
exponential phase of deploying

19
00:00:42.640 --> 00:00:44.480
it right, and deploying it at
scale.

20
00:00:44.760 --> 00:00:46.000
Last year we talked about GenAI

21
00:00:46.000 --> 00:00:49.360
and there was a magic moment of
like, whoa, is this real?

22
00:00:49.360 --> 00:00:52.560
Can it really be doing the
things that it seems to be

23
00:00:52.560 --> 00:00:52.920
doing?

24
00:00:53.120 --> 00:00:54.760
What about hallucinations, etc.

25
00:00:55.040 --> 00:00:58.160
This year it has shifted and a
lot of companies have tested it.

26
00:00:58.160 --> 00:01:01.120
They said, OK, it actually works
when deployed in the right way,

27
00:01:01.120 --> 00:01:04.080
but how do I scale this up, how
do I take it to the everyone in

28
00:01:04.080 --> 00:01:05.160
the organization, etc.

29
00:01:05.160 --> 00:01:09.560
As a result, it is in the
budgets for '24 for 90% plus

30
00:01:09.560 --> 00:01:10.400
companies.

31
00:01:10.400 --> 00:01:11.520
So it's happening.

32
00:01:12.160 --> 00:01:14.800
Now, only 10% of them or so are
actually deploying it

33
00:01:14.800 --> 00:01:15.400
efficiently.

34
00:01:15.400 --> 00:01:18.080
So they're trying to figure out
how do I scale this up, how do I

35
00:01:18.080 --> 00:01:20.560
get enough employees to learn
how to use it, how to put the

36
00:01:20.560 --> 00:01:21.880
right guardrails in place, etc.

37
00:01:21.880 --> 00:01:24.640
So it's just a different flavor
of the conversation as it's

38
00:01:24.640 --> 00:01:25.200
progressing.

39
00:01:25.440 --> 00:01:28.560
And are you seeing the leaders
who are having to have those

40
00:01:28.560 --> 00:01:30.960
conversations, having to work
out what to do?

41
00:01:31.320 --> 00:01:32.720
Are they making common mistakes?

42
00:01:32.720 --> 00:01:33.680
Are they getting this right?

43
00:01:34.560 --> 00:01:35.560
I'm sure yes.

44
00:01:35.560 --> 00:01:38.040
There's both common mistakes,
but also things that they're

45
00:01:38.040 --> 00:01:38.800
doing well, right?

46
00:01:38.800 --> 00:01:41.800
And the key thing for me is that
the ones that are doing it right

47
00:01:41.800 --> 00:01:43.440
are genuinely geeking out about
it.

48
00:01:43.440 --> 00:01:45.120
They're figuring out what it is.

49
00:01:45.120 --> 00:01:48.680
They're spending time to get
their hands dirty, to learn what

50
00:01:48.680 --> 00:01:49.440
it is, right?

51
00:01:49.480 --> 00:01:49.720
Yeah.

52
00:01:50.080 --> 00:01:51.480
And to stay close to it, right.

53
00:01:51.480 --> 00:01:54.920
The ones that are not doing well
are the ones who outsource it,

54
00:01:54.920 --> 00:01:58.120
who say three layers in the
organization, you go figure out

55
00:01:58.120 --> 00:02:00.960
this AI thing, or they think
about it as a project.

56
00:02:00.960 --> 00:02:03.600
So we'll just do it for six
months, then we're done, right?

57
00:02:03.600 --> 00:02:07.280
So it needs focus, it needs the
dedication, and it needs folks

58
00:02:07.280 --> 00:02:09.360
to stay close to it to try
things.

59
00:02:09.640 --> 00:02:12.280
And then the last thing, the
ones that do it well, they don't

60
00:02:12.280 --> 00:02:14.280
think about deploying it in a
little segment.

61
00:02:14.280 --> 00:02:17.800
They're thinking of rethinking
the whole workflow end-to-end,

62
00:02:17.800 --> 00:02:18.160
right.

63
00:02:18.160 --> 00:02:21.440
And the ones that think about
just how do I, we have a number

64
00:02:21.440 --> 00:02:24.600
of examples where you deploy
some of the AI, it extracts a

65
00:02:24.600 --> 00:02:27.440
lot of efficiencies in a
specific workflow, but that

66
00:02:27.440 --> 00:02:29.000
workflow is not a bottleneck.

67
00:02:29.000 --> 00:02:30.680
It's not a bottleneck for the
full process.

68
00:02:30.680 --> 00:02:33.000
So therefore there is no value
in doing it.

69
00:02:33.440 --> 00:02:36.840
So Vlad, you described how we're
now hitting that moment of

70
00:02:36.840 --> 00:02:39.600
exponential change where
everything is changing

71
00:02:39.600 --> 00:02:40.720
incredibly quickly.

72
00:02:41.480 --> 00:02:44.480
Are there aspects that you see
coming that you think people

73
00:02:44.480 --> 00:02:46.200
aren't really focusing enough
on?

74
00:02:46.960 --> 00:02:49.960
There is...yes and no.

75
00:02:50.880 --> 00:02:52.360
No because a lot of it is
transparent.

76
00:02:52.360 --> 00:02:54.000
We're seeing what's coming out,
right.

77
00:02:54.280 --> 00:02:57.360
But I do think there are a few
interesting aspects on the pure

78
00:02:57.360 --> 00:02:58.120
technical side.

79
00:02:58.360 --> 00:03:01.280
One is that a lot of software
companies are going to be

80
00:03:01.280 --> 00:03:03.400
seamlessly embedding a lot of
the GenAI

81
00:03:03.400 --> 00:03:07.040
and more broadly, AI features
into their offerings.

82
00:03:07.040 --> 00:03:10.000
So, a lot of the things that we
are even building with our

83
00:03:10.000 --> 00:03:13.160
clients, I think six to eight
months will be obsolete because

84
00:03:13.160 --> 00:03:16.080
it will be part of, it will be a
feature somewhere else.

85
00:03:16.280 --> 00:03:19.880
So managing expectations around
that, and getting ready for

86
00:03:19.880 --> 00:03:23.440
that, and focusing more on
deploying is going to be one of

87
00:03:23.440 --> 00:03:26.240
the new capabilities companies
need to build.

88
00:03:26.240 --> 00:03:28.360
Some of them will be
disappointed because they will

89
00:03:28.360 --> 00:03:30.840
say, oh, I just spent all this
money building it, that is OK

90
00:03:30.840 --> 00:03:33.400
because you build the muscle of
what does it take to test it,

91
00:03:33.400 --> 00:03:34.680
what does it take to deploy it?

92
00:03:34.680 --> 00:03:37.640
Well, how do you change the
workflows around it, so it's not

93
00:03:37.640 --> 00:03:38.320
wasted, right.

94
00:03:38.320 --> 00:03:39.000
So it's useful.

95
00:03:39.000 --> 00:03:41.360
It's the muscle that is being
built so that you learn how to

96
00:03:41.360 --> 00:03:41.760
deploy it.

97
00:03:42.320 --> 00:03:44.200
So that's kind of one thing that
is going to be a surprise.

98
00:03:44.200 --> 00:03:47.120
The second one is that the
companies that will unlock the

99
00:03:47.120 --> 00:03:50.360
value will be the ones that are
rethinking workflows end-to-end

100
00:03:50.360 --> 00:03:51.280
and reshaping them.

101
00:03:51.560 --> 00:03:54.760
But then be surgical about where
do they put some of the AI

102
00:03:54.760 --> 00:03:55.800
features within it.

103
00:03:56.000 --> 00:03:58.880
Some of it they will develop,
some of it will be what they

104
00:03:58.880 --> 00:04:01.720
buy, but it will be around the
reengineering of the whole

105
00:04:01.720 --> 00:04:02.160
workflow.

106
00:04:02.160 --> 00:04:05.360
So those are not necessarily
purely technical insights, but

107
00:04:05.360 --> 00:04:08.760
those are the things that are
usually blind spots for a lot of

108
00:04:08.760 --> 00:04:09.360
my clients.

109
00:04:09.440 --> 00:04:09.840
Yeah.

110
00:04:10.160 --> 00:04:12.400
And you talked earlier about the
ones who are getting it right,

111
00:04:12.400 --> 00:04:14.640
the ones who are doing it well
by being quite geeky and getting

112
00:04:14.640 --> 00:04:15.120
very involved.

113
00:04:15.200 --> 00:04:17.280
Just give us a sense of what
that looks like.

114
00:04:17.280 --> 00:04:19.080
Give us an example of where that
works really well.

115
00:04:19.520 --> 00:04:23.160
The Executive Team knows enough
about the technology so that the

116
00:04:23.160 --> 00:04:26.520
conversation they're having, you
and I, if you were on that

117
00:04:26.520 --> 00:04:29.960
Executive Team, we're talking
about where is the friction in

118
00:04:29.960 --> 00:04:32.840
the process, where is the
friction in the customer

119
00:04:32.840 --> 00:04:33.480
experience?

120
00:04:33.800 --> 00:04:34.200
Great.

121
00:04:34.200 --> 00:04:35.200
So this is the bottom line.

122
00:04:35.200 --> 00:04:38.280
How do we now, given that, let's
think now how do we deploy the

123
00:04:38.280 --> 00:04:41.360
technology. When it's not, when
they're not geeking out on it,

124
00:04:41.360 --> 00:04:42.720
when they don't know enough,

125
00:04:42.960 --> 00:04:46.280
the conversations are very much
of the likes of, is this real,

126
00:04:46.280 --> 00:04:48.640
are you sure, how much is it
going to cost?

127
00:04:48.640 --> 00:04:50.160
Should we buy, should we build
it?

128
00:04:50.440 --> 00:04:53.080
And so they don't spend any time
thinking about really where the

129
00:04:53.080 --> 00:04:55.360
value creation will happen or
where the frictions are.

130
00:04:55.360 --> 00:04:59.760
So that's a simple aggregated
difference between those that

131
00:04:59.760 --> 00:05:04.160
are leaning in versus not, and
then the ones to take it one

132
00:05:04.160 --> 00:05:08.480
click further, the ones that
engage into the deployment of

133
00:05:08.480 --> 00:05:09.880
this, they test it.

134
00:05:09.880 --> 00:05:12.800
They're the ones that in the
Executive Team, they want to not

135
00:05:12.800 --> 00:05:15.920
just see the demo, but they want
to actually see what will their

136
00:05:15.920 --> 00:05:16.560
Sales Rep see

137
00:05:16.560 --> 00:05:19.720
if they're using this tool, they
will have a point of view how to

138
00:05:19.720 --> 00:05:20.200
refine it.

139
00:05:20.200 --> 00:05:23.160
So they're getting involved into
the development of some of these

140
00:05:23.160 --> 00:05:24.960
tools and having a point of view
on it.

141
00:05:24.960 --> 00:05:27.160
And I suppose as they do that,
they see more opportunities.

142
00:05:27.160 --> 00:05:29.440
Then they say great, where are
you getting the data for this

143
00:05:29.440 --> 00:05:29.840
from that?

144
00:05:29.840 --> 00:05:31.240
Are you guys sure that's
accurate data?

145
00:05:31.240 --> 00:05:35.000
Did we clear all the privacy
policies, etc. So they can

146
00:05:35.000 --> 00:05:39.280
actually start steering you as a
developer towards answering a

147
00:05:39.280 --> 00:05:43.000
lot of the critical questions
around the design of the

148
00:05:43.000 --> 00:05:43.920
solution etc.

149
00:05:44.600 --> 00:05:46.720
But as you say, a lot of
companies aren't leaning in,

150
00:05:46.720 --> 00:05:49.240
they're sort of hoping somebody
will tell them about the change

151
00:05:49.240 --> 00:05:49.800
that's coming.

152
00:05:49.920 --> 00:05:50.400
Correct.

153
00:05:50.440 --> 00:05:55.120
Are there some industries really
that you think don't see how

154
00:05:55.120 --> 00:05:58.760
they're entirely about to change
at the moment.

155
00:05:58.760 --> 00:06:01.160
They haven't sort of seen, they
haven't started to adapt.

156
00:06:01.200 --> 00:06:04.720
Yeah, so fascinating insight for
me over the last 12 months was

157
00:06:04.720 --> 00:06:08.280
that there is no one sector, one
industry that is ahead than the

158
00:06:08.280 --> 00:06:08.680
others.

159
00:06:08.680 --> 00:06:11.680
What I'm finding is that in
every single one of them you

160
00:06:11.680 --> 00:06:15.080
have few folks that are so far
ahead, like leaning in and truly

161
00:06:15.080 --> 00:06:18.480
both disrupting themselves but
figuring out how to deploy this,

162
00:06:18.480 --> 00:06:20.880
and then some that are just
observing, right.

163
00:06:21.160 --> 00:06:22.560
And so it was lesser on that

164
00:06:22.640 --> 00:06:25.160
it's clustered in few
industries, it's across the

165
00:06:25.160 --> 00:06:25.480
board.

166
00:06:25.480 --> 00:06:28.840
So just last week I had
back-to-back meetings, one with

167
00:06:28.840 --> 00:06:32.560
an agricultural manufacturer of
seeds, the other one was with

168
00:06:32.560 --> 00:06:33.960
agricultural equipment.

169
00:06:33.960 --> 00:06:37.320
Then it was a telecom company,
then it was an insurance

170
00:06:37.320 --> 00:06:40.920
company, then it was a bank,
then it was a manufacturer of

171
00:06:40.920 --> 00:06:44.000
like building materials.
Back-to-back-to-back with

172
00:06:44.000 --> 00:06:46.680
companies that are doing really
cool stuff.

173
00:06:46.680 --> 00:06:49.800
So it's not just one sector,
it's actually proliferated

174
00:06:49.800 --> 00:06:52.120
across, which is very
interesting to me.

175
00:06:52.360 --> 00:06:54.520
Well, that's a vision of the
future that's coming.

176
00:06:55.000 --> 00:06:55.560
It's coming our way.

177
00:06:55.560 --> 00:06:56.680
It's coming fast.

178
00:06:56.960 --> 00:06:58.600
Thanks very much for joining us,
Vlad.

179
00:06:58.600 --> 00:07:01.960
That's Vlad Lukic, who is the
Global Leader of Tech and

180
00:07:01.960 --> 00:07:03.080
Digital Advantage.

181
00:07:03.400 --> 00:07:04.400
And I'm Manveen Rana.