WEBVTT

1
00:00:00.000 --> 00:00:01.600
Hi Sesh, thanks for joining us.

2
00:00:01.600 --> 00:00:03.400
When it comes to GenAI,

3
00:00:03.640 --> 00:00:05.720
There are tech challenges
that business leaders face.

4
00:00:05.720 --> 00:00:06.720
What are they?

5
00:00:07.440 --> 00:00:09.320
I see many tech challenges.

6
00:00:09.320 --> 00:00:10.600
Let me just start with three of
them.

7
00:00:10.600 --> 00:00:12.360
And and there are others.

8
00:00:12.720 --> 00:00:15.680
I think the first is just data
quality and availability.

9
00:00:15.680 --> 00:00:16.240
A lot of GenAI

10
00:00:16.240 --> 00:00:19.720
is driven by underlying data
quality, availability, how you

11
00:00:19.720 --> 00:00:23.320
actually go to the right sources
of data, how you develop the

12
00:00:23.320 --> 00:00:26.200
right pipelines, how you
implement the right data

13
00:00:26.200 --> 00:00:26.840
governance.

14
00:00:27.440 --> 00:00:30.040
These are things that have to be
put in place to be able to

15
00:00:30.040 --> 00:00:30.720
extract insight.

16
00:00:30.720 --> 00:00:33.160
I think that's one, and that's a
real area of focus.

17
00:00:33.920 --> 00:00:38.200
I think a second thing is there
are so many different failure

18
00:00:38.200 --> 00:00:39.840
modes when you implement

19
00:00:39.840 --> 00:00:41.880
GenAI solutions.

20
00:00:42.440 --> 00:00:46.640
Hallucinations, prompt injection
and jailbreaking, degeneration

21
00:00:46.640 --> 00:00:50.200
of models, catastrophic
forgetting of models, privacy

22
00:00:50.200 --> 00:00:54.320
violations, these are all things
that have to be put in place.

23
00:00:54.320 --> 00:01:00.600
LLM bias, fair use of of LLMS,
all of these start to become

24
00:01:00.600 --> 00:01:02.800
challenges that you have to
overcome when you're deploying

25
00:01:02.800 --> 00:01:04.160
at-scale solutions.

26
00:01:04.160 --> 00:01:05.480
That's the second thing that I see.

27
00:01:06.000 --> 00:01:09.640
The third thing is just human
capital and scarcity of talent.

28
00:01:10.440 --> 00:01:13.840
We are clearly seeing most
organizations having to upskill

29
00:01:13.840 --> 00:01:17.440
or reskill their people to be
able to put these people against

30
00:01:17.440 --> 00:01:18.720
GenAI work.

31
00:01:18.720 --> 00:01:25.360
46% of people, based on the
survey that we did recently, in

32
00:01:25.360 --> 00:01:27.320
these organizations on average
need to be upscaled over the

33
00:01:27.320 --> 00:01:28.600
next three years.

34
00:01:29.280 --> 00:01:32.800
Then there are of course a
broader set of challenges like

35
00:01:32.800 --> 00:01:35.600
legacy integrations, we talked
about security.

36
00:01:36.600 --> 00:01:39.880
There is a lot that needs to be
done around the explainability

37
00:01:39.880 --> 00:01:43.080
or interpretability of the
results that these models give you

38
00:01:43.560 --> 00:01:46.240
So there are a host of other
things that need to be overcome.

39
00:01:46.440 --> 00:01:49.160
But the nice thing is we are
starting to find solutions.

40
00:01:49.160 --> 00:01:53.440
We're able to actually address
these issues and get solutions

41
00:01:53.440 --> 00:01:54.040
to scale.

42
00:01:54.320 --> 00:01:57.080
And are there opportunities,
tech-driven opportunities, that

43
00:01:57.080 --> 00:01:59.120
business leaders are
overlooking, I imagine?

44
00:02:00.840 --> 00:02:05.520
We did a recent survey and we
surveyed 1400 executives.

45
00:02:06.480 --> 00:02:09.400
The big thing that came out of
that is there's only one in ten

46
00:02:09.400 --> 00:02:11.160
executives that are actually pioneers.

47
00:02:11.440 --> 00:02:14.560
Nine out of ten are actually
waiting and watching, to the

48
00:02:14.560 --> 00:02:16.160
first question that you asked.

49
00:02:17.040 --> 00:02:19.800
And maybe the way I answer the
question is, what are the things

50
00:02:19.800 --> 00:02:21.240
that the pioneers are doing well?

51
00:02:21.240 --> 00:02:23.960
The first thing is they're
massively ambitious.

52
00:02:24.200 --> 00:02:26.080
So when they think about
investing in GenAI

53
00:02:26.080 --> 00:02:31.720
for top line and productivity,
their ambition is 1.3 to 1.4

54
00:02:31.720 --> 00:02:38.160
times higher than what we see
[in] folks who are not ready to invest.

55
00:02:38.880 --> 00:02:41.400
I think the second is around
upskilling their talent.

56
00:02:43.240 --> 00:02:47.040
There are close to, you know,
60% of these executives who

57
00:02:47.040 --> 00:02:50.400
think that their senior
leadership is not ready to

58
00:02:50.400 --> 00:02:51.640
actually capitalize

59
00:02:51.640 --> 00:02:52.440
on GenAI solutions.

60
00:02:53.160 --> 00:02:54.920
So upscaling is a massive opportunity.

61
00:02:55.560 --> 00:02:58.720
A third one is really
understanding the cost base of

62
00:02:58.720 --> 00:02:59.360
these GenAI

63
00:02:59.360 --> 00:03:02.200
solutions, what is the cost to
actually implement the

64
00:03:02.200 --> 00:03:05.000
solutions, direct costs and
indirect costs and really

65
00:03:05.000 --> 00:03:08.680
understanding the detail of what
these costs are going to be over time.

66
00:03:08.720 --> 00:03:09.680
That's the third one.

67
00:03:10.280 --> 00:03:13.760
I think the fourth is around
partnerships and ecosystems.

68
00:03:14.480 --> 00:03:17.600
Many of these pioneers are
investing in really deep

69
00:03:17.600 --> 00:03:20.280
partnerships, building out their ecosystems.

70
00:03:20.760 --> 00:03:23.080
And then the last one I would
say is when you develop these

71
00:03:23.280 --> 00:03:26.200
GenAI solutions, you have to do it
with responsible AI principles.

72
00:03:26.840 --> 00:03:31.400
We still see only, you know
about 17% of organizations have

73
00:03:31.400 --> 00:03:34.120
responsible AI on the CEO's agenda.

74
00:03:34.480 --> 00:03:37.240
So these are the things that the
winners are doing.

75
00:03:37.240 --> 00:03:41.080
I think the observers need to
learn from this and start to

76
00:03:41.080 --> 00:03:43.320
really focus on these five things.

77
00:03:43.440 --> 00:03:44.840
It's a really overwhelming issue.

78
00:03:45.320 --> 00:03:46.880
Sesh, thank you for breaking it
down for us.

79
00:03:47.080 --> 00:03:48.040
Thank you for having me.

80
00:03:48.040 --> 00:03:48.600
This is great.

81
00:03:49.040 --> 00:03:49.720
Good talking to you.