WEBVTT

1
00:00:00.291 --> 00:00:02.752
Rob, good to see you. Vikas,
nice to be with you.

2
00:00:03.628 --> 00:00:08.383
Rob, you've described AI as the
opportunity of a lifetime for

3
00:00:08.383 --> 00:00:09.259
businesses.

4
00:00:09.300 --> 00:00:11.845
What do you think makes the
opportunity so compelling?

5
00:00:12.595 --> 00:00:15.265
I don't think technology has
ever been this accessible.

6
00:00:15.557 --> 00:00:19.269
I look at the work we're doing
with clients right now, and it's

7
00:00:19.269 --> 00:00:22.272
not years, it's not weeks, and
sometimes it's days.

8
00:00:22.814 --> 00:00:26.192
And this is an opportunity to
completely change how you do

9
00:00:26.192 --> 00:00:26.693
business.

10
00:00:27.277 --> 00:00:29.279
It's the productivity that it
can bring.

11
00:00:29.863 --> 00:00:32.365
It's how you can augment your
developers.

12
00:00:33.241 --> 00:00:37.746
AI is going to touch every
business, every government, in

13
00:00:37.746 --> 00:00:41.124
every part, which is why now is
the moment.

14
00:00:41.499 --> 00:00:43.752
It is the opportunity of a
lifetime to take advantage of.

15
00:00:45.170 --> 00:00:49.340
And Rob, we're here at Davos and
there's a lot of dialogue around

16
00:00:49.340 --> 00:00:50.633
both the opportunity.

17
00:00:50.633 --> 00:00:53.136
There's so many stories about
the impact of AI.

18
00:00:53.470 --> 00:00:56.639
There are also a lot of stories
about the worries and

19
00:00:56.639 --> 00:01:00.143
challenges. So put yourself in
the shoes of a CEO or a CIO.

20
00:01:00.143 --> 00:01:02.979
If you're considering an AI
transformation, what are the

21
00:01:02.979 --> 00:01:04.981
factors that you should be
considering?

22
00:01:05.940 --> 00:01:08.026
Let's go big picture for a
moment.

23
00:01:08.068 --> 00:01:12.072
There's a there's an old macro
investing equation that says GDP

24
00:01:12.072 --> 00:01:15.867
growth comes from population
growth plus productivity growth

25
00:01:15.867 --> 00:01:16.910
plus debt growth.

26
00:01:17.994 --> 00:01:20.330
Well, population is not growing
in many countries around the

27
00:01:20.330 --> 00:01:21.831
world if you look at the next 20
years.

28
00:01:22.123 --> 00:01:25.960
So that means the only way we're
going to grow as a global

29
00:01:25.960 --> 00:01:28.296
economy is by driving
productivity.

30
00:01:29.130 --> 00:01:31.883
AI is going to be that catalyst
to drive productivity.

31
00:01:32.425 --> 00:01:33.343
How is it going to do it?

32
00:01:33.551 --> 00:01:38.640
It's going to make everybody
more effective, faster at doing

33
00:01:38.640 --> 00:01:39.474
their job.

34
00:01:39.974 --> 00:01:44.145
We've seen three major use cases
for AI so far: customer service,

35
00:01:44.145 --> 00:01:45.939
every company has customers.

36
00:01:45.939 --> 00:01:48.733
They're trying to figure out how
do I serve them better, how do I

37
00:01:48.733 --> 00:01:50.235
make it a more intimate
experience?

38
00:01:50.318 --> 00:01:51.569
You can do that with AI.

39
00:01:52.153 --> 00:01:53.738
Second is digital labor.

40
00:01:53.988 --> 00:01:58.743
How do I augment my employees to
make them 50% more productive in

41
00:01:58.743 --> 00:02:03.081
doing repetitive tasks. Again,
accessible to every customer.

42
00:02:03.581 --> 00:02:04.749
Now I say third is code.

43
00:02:05.291 --> 00:02:07.293
Most companies have software
developers.

44
00:02:07.710 --> 00:02:09.045
They're building products.

45
00:02:09.045 --> 00:02:10.463
They're building technology.

46
00:02:10.713 --> 00:02:13.341
If you can make them 30 to 40%
more productive,

47
00:02:14.217 --> 00:02:16.970
it completely changes how fast
you can release products and

48
00:02:16.970 --> 00:02:18.012
engage your customers.

49
00:02:18.388 --> 00:02:21.141
So those are the use cases that
I'm seeing in the real world

50
00:02:21.141 --> 00:02:21.391
today.

51
00:02:21.808 --> 00:02:24.769
And a lot of the talk around AI
is around generative AI.

52
00:02:25.186 --> 00:02:28.064
At BCG, we like to talk about
the left brain and the right

53
00:02:28.064 --> 00:02:30.817
brain, so both the generative,
as well as the predictive

54
00:02:30.817 --> 00:02:31.568
elements of AI.

55
00:02:31.568 --> 00:02:34.279
How do you feel those will
balance each other out, or how

56
00:02:34.279 --> 00:02:36.739
do you think those are
complementary to each other?

57
00:02:37.198 --> 00:02:40.869
I do think people mix these,
maybe intentionally, maybe

58
00:02:40.869 --> 00:02:44.330
unintentionally. Maybe it
matters, maybe it doesn't.

59
00:02:44.330 --> 00:02:46.875
But I would say the generative
AI is exactly what it sounds

60
00:02:46.875 --> 00:02:47.083
like.

61
00:02:47.584 --> 00:02:51.546
You're generating new content,
so it's incredibly effective for

62
00:02:51.546 --> 00:02:55.508
some of the use cases I talked
about for doing summarization of

63
00:02:55.508 --> 00:02:56.134
documents.

64
00:02:56.593 --> 00:03:00.889
But let's not forget, we spent
the better part of a decade

65
00:03:00.889 --> 00:03:05.185
before on core machine learning,
deep learning, predictive

66
00:03:05.185 --> 00:03:05.935
analytics.

67
00:03:05.935 --> 00:03:10.023
This is about how do I make a
prediction about what might

68
00:03:10.023 --> 00:03:10.523
happen.

69
00:03:11.149 --> 00:03:15.069
The existence of generative AI
has not reduced the need for

70
00:03:15.069 --> 00:03:15.778
prediction.

71
00:03:16.196 --> 00:03:18.698
If anything, I'd say it's made
it more important.

72
00:03:18.698 --> 00:03:22.202
So when we're working with
clients, we say maybe don't

73
00:03:22.202 --> 00:03:24.120
start with the technology yet.

74
00:03:24.162 --> 00:03:27.165
Let's start with the use case,
then let's figure out the right

75
00:03:27.165 --> 00:03:28.416
technology to bring to it.

76
00:03:28.583 --> 00:03:32.295
It's part of why what we've done
with watsonx is making all types

77
00:03:32.295 --> 00:03:33.463
of models available.

78
00:03:33.796 --> 00:03:37.383
Machine learning models, open
source models, some models built

79
00:03:37.383 --> 00:03:37.759
by IBM.

80
00:03:38.343 --> 00:03:40.386
We think there is a model for
every job.

81
00:03:40.845 --> 00:03:42.388
The question is, can you figure
out what that is?

82
00:03:42.972 --> 00:03:46.935
And similarly, what do you think
is the relationship between AI

83
00:03:46.935 --> 00:03:49.646
and cloud, hybrid cloud, quantum
computing.

84
00:03:50.271 --> 00:03:54.525
AI is probably perfect for this
era of what we call hybrid

85
00:03:54.525 --> 00:03:54.984
cloud.

86
00:03:55.360 --> 00:03:57.904
If you go back five years,
everybody was thinking how do I

87
00:03:57.904 --> 00:03:59.113
move everything to one cloud.

88
00:03:59.530 --> 00:04:01.991
I don't talk to any client now
that says that's their

89
00:04:01.991 --> 00:04:02.492
imperative.

90
00:04:02.533 --> 00:04:06.829
The imperative now is how am I
going to operate my business

91
00:04:06.829 --> 00:04:08.498
across multiple clouds?

92
00:04:08.623 --> 00:04:11.251
Public, private, edge.

93
00:04:11.751 --> 00:04:15.546
And AI is all about how do you
federate access to data across

94
00:04:15.546 --> 00:04:16.714
those environments.

95
00:04:17.048 --> 00:04:19.342
You may have applications across
those environments.

96
00:04:19.342 --> 00:04:23.805
So to some extent, AI will be
accelerated by the existence of

97
00:04:23.805 --> 00:04:26.391
hybrid cloud, not just cloud
alone.

98
00:04:26.391 --> 00:04:28.559
Let me switch gears here for a
minute.

99
00:04:28.559 --> 00:04:31.771
Rob, you've been an AI
practitioner for a long time.

100
00:04:32.563 --> 00:04:36.067
What was the moment when you
first thought to yourself: Ha!

101
00:04:36.067 --> 00:04:37.527
this could be really big?

102
00:04:38.736 --> 00:04:41.906
I think it always comes to life
with a client, and the one that

103
00:04:41.906 --> 00:04:45.034
I saw, this was actually back to
your point on the early days,

104
00:04:45.034 --> 00:04:46.327
more machine learning, was

105
00:04:46.327 --> 00:04:48.830
Bradesco, one of the largest
banks in the world.

106
00:04:49.372 --> 00:04:53.209
When I saw that they were doing
200,000 customer inquiries a

107
00:04:53.209 --> 00:04:56.963
month without ever touching a
human, which means they could

108
00:04:56.963 --> 00:05:00.633
put all their people on the
harder tasks, I was like, this

109
00:05:00.633 --> 00:05:04.554
is going to completely change
everything because now they can

110
00:05:04.554 --> 00:05:06.472
get higher Net Promoter Scores.

111
00:05:07.849 --> 00:05:10.143
Their employees are doing the
kind of work that they want to

112
00:05:10.143 --> 00:05:11.602
do, not the work that they have
to do.

113
00:05:12.020 --> 00:05:13.396
So that was the aha moment for
me.

114
00:05:13.396 --> 00:05:16.274
And I take that case and now
it's even better with generative

115
00:05:16.274 --> 00:05:16.399
AI.

116
00:05:17.025 --> 00:05:19.027
So that was the aha moment for
me.

117
00:05:19.360 --> 00:05:21.321
And fast forward to today.

118
00:05:21.321 --> 00:05:24.240
And you know, IBM has been
announcing so much around

119
00:05:24.240 --> 00:05:26.826
watsonx, what's your proudest
customer moment,

120
00:05:26.826 --> 00:05:29.412
more recently? We announced the
partnership with Dun & Bradstreet

121
00:05:29.412 --> 00:05:30.997
they're a data company.

122
00:05:31.164 --> 00:05:33.291
They're trying to figure out how
do they make their data more

123
00:05:33.291 --> 00:05:33.666
accessible.

124
00:05:34.167 --> 00:05:37.545
There's probably no better
partnership for us than working

125
00:05:37.545 --> 00:05:40.798
with a data company to bring
their data to life with AI.

126
00:05:41.424 --> 00:05:44.385
They're also using watsonx to
transform their operations.

127
00:05:44.886 --> 00:05:47.805
We're also building a
go-to-market with them where we

128
00:05:47.805 --> 00:05:51.142
can feature Dun&Bradstreet
data as part of how we go to

129
00:05:51.142 --> 00:05:51.517
market.

130
00:05:51.517 --> 00:05:55.229
So I think it's an example of
the kind of partnerships that

131
00:05:55.229 --> 00:05:58.316
happen in this AI and hybrid
cloud era, which are

132
00:05:58.316 --> 00:05:59.108
multifaceted.

133
00:05:59.400 --> 00:06:02.153
It's not just a customer vendor,
it's how do we actually be

134
00:06:02.153 --> 00:06:02.945
partners together.

135
00:06:03.988 --> 00:06:06.741
Rob, thank you for a wonderful
and thought-provoking

136
00:06:06.741 --> 00:06:07.408
conversation.

137
00:06:07.408 --> 00:06:08.159
Thank you, Vikas.