1
00:00:04,194 --> 00:00:07,434
Good afternoon, everyone present here.

2
00:00:07,814 --> 00:00:13,334
It's been a great day and an absolute pleasure to be here.

3
00:00:14,634 --> 00:00:21,994
The topic of my talk today is everyone has AI, who has the advantage?

4
00:00:22,754 --> 00:00:31,714
I'll be talking about what the orgs that are doing, that are leading in AI with their implementation and adoption,

5
00:00:32,034 --> 00:00:33,494
what are they doing differently?

6
00:00:35,814 --> 00:00:40,554
so before that, who am I and what's my POV?

7
00:00:40,554 --> 00:00:48,794
So I'm Dr. Sid Pednekar. I am an AI technology and data exec in asset management.

8
00:00:48,974 --> 00:00:56,114
I've spent most of my tenure in hedge funds and investment in academic research.

9
00:00:56,714 --> 00:01:04,734
What I've done and focused on is lead strategy and build specialized teams towards three main verticals,

10
00:01:04,734 --> 00:01:10,754
optimizing efficiency, driving PNL, and reducing costs.

11
00:01:11,514 --> 00:01:18,714
So I'll be following the discussion along with two hypothetical companies,

12
00:01:19,074 --> 00:01:22,454
Company A and Company B, right at the slider below.

13
00:01:23,054 --> 00:01:33,334
These are hypothetical, you know, with made-up workflows, but we'll be following their journey along the rest of this presentation.

14
00:01:35,234 --> 00:01:38,174
So, everyone has AI. Who has the advantage?

15
00:01:40,294 --> 00:01:44,314
Before we get started there, first let's see what the lay of the land is.

16
00:01:44,814 --> 00:01:54,214
And recent studies have shown that only about 37% of enterprises that were surveyed are

17
00:01:54,214 --> 00:01:57,854
showing a positive contribution to their EBIT.

18
00:01:58,394 --> 00:02:03,994
And only about 6% of those would be qualified as high performers.

19
00:02:03,994 --> 00:02:07,234
So that's about 5% contribution to their EBIT.

20
00:02:07,234 --> 00:02:16,254
Other studies earlier in the year have shown that about 66% are saying they're having a productivity increase,

21
00:02:16,254 --> 00:02:22,314
but only about 40% and 20% are actually seeing cost reductions and revenue growth.

22
00:02:23,114 --> 00:02:30,134
So as we know, access of AI is widespread, yet the impact remains uneven.

23
00:02:34,174 --> 00:02:43,554
So, the question I would ask, or how I think about this, is really in terms of three verticals.

24
00:02:44,274 --> 00:02:51,254
And the way I think about these three verticals should apply to organizations and individuals

25
00:02:51,254 --> 00:02:59,234
alike, and it should apply to functions, regardless of where you are in that organization.

26
00:03:00,254 --> 00:03:06,514
So the three main vertical pillars that I think of are, number one, design the right problem.

27
00:03:07,614 --> 00:03:10,734
Number two, build shared capabilities.

28
00:03:11,314 --> 00:03:13,934
And number three, compound the advantage.

29
00:03:15,094 --> 00:03:23,514
Now before we dive into each of these pillars, there is an idea that I think goes through all these verticals.

30
00:03:23,694 --> 00:03:29,394
And I think the best leading firms are taking advantage of that and capitalizing on it.

31
00:03:30,594 --> 00:03:36,114
As we know, throughout time, the only constant in this world has been change.

32
00:03:36,814 --> 00:03:43,554
What's different in the AI and superintelligence era is that intelligence has become cheap.

33
00:03:44,434 --> 00:03:51,994
Regardless of what your unit of intelligence is, whether it be software, analytics, recommendations, reports,

34
00:03:52,874 --> 00:03:55,394
AI is creating an abundance in all of it.

35
00:03:55,394 --> 00:04:01,874
And as we know, more answers don't necessarily create more value.

36
00:04:02,554 --> 00:04:04,434
So what is scarce?

37
00:04:04,794 --> 00:04:07,594
What is the scarce resource in an abundant world?

38
00:04:09,774 --> 00:04:13,474
What are companies taking advantage of, the leaders at least?

39
00:04:14,674 --> 00:04:15,694
It's judgment.

40
00:04:17,194 --> 00:04:18,414
What is judgment?

41
00:04:19,574 --> 00:04:23,394
Judgment is choosing what matters and how to act

42
00:04:23,394 --> 00:04:25,294
when there's no single obvious answer.

43
00:04:25,394 --> 00:04:29,954
It's the creativity that lets you explore what possibilities are worth exploring.

44
00:04:30,294 --> 00:04:35,054
It's the taste that tells you what good looks like in your given context.

45
00:04:35,274 --> 00:04:40,774
It's the decisions where you can weigh what trade-offs are worth making.

46
00:04:42,814 --> 00:04:43,994
is ubiquitous.

47
00:04:44,474 --> 00:04:48,014
It exists throughout your organization on a daily basis.

48
00:04:48,454 --> 00:04:54,134
Your C-suite, the C-suite is asking questions, taking advantage of judgment all the time.

49
00:04:54,134 --> 00:05:00,014
So are the software engineers, the AI leaders, the portfolio managers, and the data scientists.

50
00:05:00,634 --> 00:05:11,574
So the question that they're focused on is, given judgment is a finite asset, a limited capacity,

51
00:05:11,954 --> 00:05:16,654
the main question becomes, where is our best judgment worth spending?

52
00:05:17,394 --> 00:05:19,354
I'd like for you to hold on to that thought.

53
00:05:19,994 --> 00:05:22,234
With that, we'll dive into our first pillar.

54
00:05:22,574 --> 00:05:23,934
Design the right problem.

55
00:05:24,474 --> 00:05:37,974
So questions, a question that the leading organizations are asking all the time is they're obsessed over what is the right problem to solve, right?

56
00:05:38,154 --> 00:05:39,554
It's tempting to start with a tool.

57
00:05:39,774 --> 00:05:44,314
Today we're surrounded with better models, harnesses.

58
00:05:44,454 --> 00:05:45,954
We're surrounded with all these AI tools.

59
00:05:46,094 --> 00:05:47,374
We're surrounded with co-pilots.

60
00:05:48,594 --> 00:05:51,514
But that isn't the way about, is it?

61
00:05:51,614 --> 00:05:53,914
At least that's not the way the leaders are approaching it.

62
00:05:54,134 --> 00:05:57,714
Now, let's go back to our hypothetical two companies.

63
00:05:57,934 --> 00:05:59,534
Let's say they're from asset management.

64
00:05:59,714 --> 00:06:03,894
Let's say we're tracking the workflow of two research analysts

65
00:06:03,894 --> 00:06:07,394
in these hypothetical organizations.

66
00:06:08,314 --> 00:06:13,314
Well, let's say organization B starts with the tool, the co-pilot.

67
00:06:13,634 --> 00:06:17,154
They're asking, where can we add AI to the analyst's work?

68
00:06:17,734 --> 00:06:21,614
They're looking at the current existing process to say,

69
00:06:21,614 --> 00:06:27,694
earning transcript comes in, I discovered a summarization workflow, let me pass all of

70
00:06:27,694 --> 00:06:34,814
this in through my fire hose, analysts get earning summaries faster, win-win, right?

71
00:06:35,534 --> 00:06:39,374
Well, not so much, at least not compared to the guys who are leading.

72
00:06:39,894 --> 00:06:45,914
And such a workflow of being a faster horse introduces one key nuance.

73
00:06:45,914 --> 00:06:53,414
That is, you are losing context as you have information, I guess, reduction as part of your summarization.

74
00:06:53,854 --> 00:06:58,194
So what is a hypothetical organization A doing in this?

75
00:06:58,774 --> 00:07:00,814
Well, they're obsessed on the right problem to ask.

76
00:07:01,494 --> 00:07:06,194
They could be asking, why are our material signals reaching us too late?

77
00:07:06,534 --> 00:07:11,274
Where do we focus the attention, the AI, the bandwidth, the capability on?

78
00:07:11,474 --> 00:07:13,614
Where is our bottleneck, right?

79
00:07:13,614 --> 00:07:17,054
And so it becomes this redesigning of the workflow,

80
00:07:17,054 --> 00:07:20,314
or at least a re-evaluation on what to focus on.

81
00:07:20,554 --> 00:07:21,674
That's what they're doing different.

82
00:07:22,254 --> 00:07:25,194
Now, this consideration isn't particular to finance.

83
00:07:25,194 --> 00:07:27,454
It's across the realm and function.

84
00:07:28,054 --> 00:07:32,394
You're asking the same question on what to focus on,

85
00:07:32,494 --> 00:07:35,374
whether it be research capacity, decision speed,

86
00:07:35,854 --> 00:07:39,054
quality of your data, latency, reliability.

87
00:07:40,294 --> 00:07:42,194
It traverses all of that.

88
00:07:43,314 --> 00:07:44,714
studies show this.

89
00:07:44,974 --> 00:07:48,514
They show leading studies, on average,

90
00:07:49,114 --> 00:07:53,294
spend their attention on fewer higher quality decisions

91
00:07:53,294 --> 00:07:54,994
than the other cohort.

92
00:07:55,234 --> 00:07:57,094
So leaders are being more selective

93
00:07:57,094 --> 00:07:59,294
in where they focus their attention

94
00:07:59,294 --> 00:08:01,234
and what they focus their attention.

95
00:08:01,514 --> 00:08:02,434
So it's great.

96
00:08:02,594 --> 00:08:05,554
The leading companies have an idea

97
00:08:05,554 --> 00:08:07,834
of what the right problems are.

98
00:08:08,014 --> 00:08:08,834
Where do they go from there?

99
00:08:09,594 --> 00:08:15,314
Well, they also concentrate their investments in the highest value workflows.

100
00:08:16,794 --> 00:08:21,154
I get it, and I've been culpable of something similar,

101
00:08:21,414 --> 00:08:25,274
where AI is almost creating too many opportunities.

102
00:08:26,224 --> 00:08:28,314
And experimentation can be a good thing.

103
00:08:29,204 --> 00:08:34,014
But one needs to separate experimentation and the resulting innovation theater,

104
00:08:34,014 --> 00:08:38,854
or worse, the productivity theater, which actual value generation.

105
00:08:40,074 --> 00:08:44,134
Going back to our two hypothetical companies, what does this mean?

106
00:08:44,274 --> 00:08:48,834
Company B is looking at how do we launch the most number of AI pilots?

107
00:08:49,194 --> 00:08:53,674
And what that ends up looking like is a document summarization bot here,

108
00:08:53,854 --> 00:08:58,134
a co-pilot there, a data retrieval bot here.

109
00:08:58,434 --> 00:09:00,014
It's scattered all over the place.

110
00:09:00,014 --> 00:09:04,414
whereas what company A is focused on in their investments

111
00:09:04,414 --> 00:09:09,314
is potentially use cases like end-to-end workflows

112
00:09:09,314 --> 00:09:12,254
on the research pipeline.

113
00:09:12,674 --> 00:09:17,894
They're focused on aligning engineering, analytics, AI, capacity,

114
00:09:18,314 --> 00:09:21,334
and judgment towards a singular focus.

115
00:09:21,754 --> 00:09:28,094
And research has shown that the leaders spend more than 80%

116
00:09:28,094 --> 00:09:32,334
of their investments on these high-value use cases,

117
00:09:33,434 --> 00:09:35,694
transforming key business functions

118
00:09:35,694 --> 00:09:37,354
and creating new offerings.

119
00:09:38,654 --> 00:09:39,414
Perfect.

120
00:09:39,634 --> 00:09:43,634
We know they're focusing on finding the right problem,

121
00:09:43,634 --> 00:09:46,894
and they're also spending deep investment

122
00:09:46,894 --> 00:09:47,974
in select workflows.

123
00:09:48,494 --> 00:09:50,634
Well, they're also doing something different.

124
00:09:51,514 --> 00:09:53,614
They're reimagining how work is done.

125
00:09:55,294 --> 00:09:57,594
I think this is key.

126
00:09:58,094 --> 00:10:01,054
And it goes back to what we discussed previously, which is judgment.

127
00:10:02,694 --> 00:10:07,194
They're redefining and reorganizing functions, roles, capacities

128
00:10:07,194 --> 00:10:09,074
around this key element of judgment.

129
00:10:11,034 --> 00:10:13,774
What does that mean for our hypothetical company?

130
00:10:14,054 --> 00:10:16,914
Well, company B might be focused on, as we discussed,

131
00:10:17,634 --> 00:10:21,554
I'm getting my earnings transcript, I'm getting my research reports.

132
00:10:21,674 --> 00:10:24,334
How do I summarize it? Maybe I make that faster.

133
00:10:24,534 --> 00:10:26,674
Maybe I update my financial models faster.

134
00:10:26,674 --> 00:10:29,254
Maybe I send my recommendation to my PM faster.

135
00:10:30,234 --> 00:10:34,614
Company A is focused on an entirely different ballgame.

136
00:10:35,454 --> 00:10:39,194
Hypothetically, they could be asking the analyst to define the investment thesis,

137
00:10:39,374 --> 00:10:44,994
and they're watching and gathering evidence that could refute that thesis or support it.

138
00:10:45,414 --> 00:10:50,034
They'd be monitoring continuously, they'd be investigating exceptions,

139
00:10:50,914 --> 00:10:58,134
and they could be deep diving on select examples

140
00:10:58,134 --> 00:11:01,354
that could lead to actual generation of value.

141
00:11:01,854 --> 00:11:04,334
So the winners don't just automate, they redesign.

142
00:11:05,794 --> 00:11:08,874
studies show that the cohort of high performers

143
00:11:08,874 --> 00:11:12,334
are 75% more likely, or in practice,

144
00:11:12,334 --> 00:11:14,194
they actually do redesign their workflow

145
00:11:14,194 --> 00:11:17,014
versus 25% of the cohort that doesn't.

146
00:11:18,814 --> 00:11:22,354
So where does this leave us with our two hypothetical companies

147
00:11:22,454 --> 00:11:44,034
and designing the right problem? Company B adds AI to existing workflows, they're completing individual projects and they're moving to the next initiative, whereas Company A is rethinking how work is done. They're redesigning processes around AI and they're obsessing over the right problem. That's the difference in how organizations put AI to work.

148
00:11:45,634 --> 00:11:47,674
Great, so they've picked the right problem.

149
00:11:47,914 --> 00:11:49,514
What happens next?

150
00:11:50,634 --> 00:11:54,374
Well, we all know in order to actually enable that capacity,

151
00:11:54,374 --> 00:11:58,054
you need to build share capabilities across the org.

152
00:11:59,374 --> 00:12:04,254
And what's a trend that's seemingly emerging for some time now

153
00:12:04,254 --> 00:12:08,514
is that the lines between technology and business are converging.

154
00:12:08,814 --> 00:12:10,314
And we're seeing this everywhere.

155
00:12:10,514 --> 00:12:13,494
We're seeing engineers that are moving closer to business outcomes.

156
00:12:13,494 --> 00:12:17,274
We're seeing business professionals that are building tools and automations.

157
00:12:17,914 --> 00:12:22,914
And as you can see and imagine, you can have forward deployed engineers,

158
00:12:23,234 --> 00:12:25,134
you can have forward deployed product managers,

159
00:12:25,294 --> 00:12:29,814
but you could also have the inverse, which is forward deployed portfolio managers.

160
00:12:30,054 --> 00:12:35,354
And it's really this combination of getting the right people in the room

161
00:12:35,354 --> 00:12:41,634
by breaking down organization silos that the leaders are really leaning on.

162
00:12:42,694 --> 00:12:51,034
studies have shown that more and more of the large enterprises are viewing their technology arm as a business growth function

163
00:12:51,034 --> 00:12:54,634
and that's creating an edge for these companies

164
00:12:57,194 --> 00:13:05,154
so the other sort of vertical inside building shared capabilities that is probably a personal favorite of mine

165
00:13:05,154 --> 00:13:09,294
it's standardizing the core and personalizing the edges

166
00:13:10,194 --> 00:13:13,574
And this is something that isn't new to the technology world.

167
00:13:13,754 --> 00:13:14,894
We've built infrastructure.

168
00:13:14,894 --> 00:13:17,854
We've built shared modules and system design.

169
00:13:18,034 --> 00:13:24,114
But what's interesting here today is that's taking a very different shape with the capabilities out there

170
00:13:24,114 --> 00:13:26,334
and with the org and the breaking down of silos.

171
00:13:26,694 --> 00:13:31,294
And this could look very different depending on where you sit.

172
00:13:31,944 --> 00:13:34,034
You could be building shared cores

173
00:13:34,034 --> 00:13:37,294
for a team that looks very different

174
00:13:37,294 --> 00:13:39,394
with analysts and clients.

175
00:13:39,394 --> 00:13:42,274
You could be doing that with a bunch of analysts,

176
00:13:42,294 --> 00:13:44,834
or you could be doing that for software engineers

177
00:13:44,834 --> 00:13:46,134
or data scientists.

178
00:13:47,074 --> 00:13:50,334
With the data scientist example in particular,

179
00:13:50,654 --> 00:13:52,234
that could look very interesting

180
00:13:52,234 --> 00:13:56,454
where instead of having one-off workflows

181
00:13:56,454 --> 00:13:58,834
that is really common in the industry,

182
00:13:58,834 --> 00:14:00,754
you could build a shared core

183
00:14:00,754 --> 00:14:03,834
of the data sets, their semantic layers,

184
00:14:04,134 --> 00:14:06,154
the trade-offs of what worked, what didn't,

185
00:14:06,954 --> 00:14:09,554
what some data sets do good of,

186
00:14:09,654 --> 00:14:12,154
where are the gaps, and so on and so forth,

187
00:14:12,254 --> 00:14:14,514
such that each new personnel

188
00:14:14,514 --> 00:14:17,774
becomes almost a plug-and-play module

189
00:14:17,774 --> 00:14:20,674
in the system where they can then stand

190
00:14:20,674 --> 00:14:22,774
on the shoulder of their predecessors.

191
00:14:23,714 --> 00:14:26,494
So what should the next person not have to rebuild?

192
00:14:28,414 --> 00:14:36,074
now what's also interesting is um you know the enterprise wasn't really built for agents um agents

193
00:14:36,074 --> 00:14:42,714
they they work across workflows and applications they they take context they uh deliberate they

194
00:14:42,714 --> 00:14:50,054
take actions whereas really the enterprises are focused um previously on a single human

195
00:14:50,054 --> 00:14:53,394
or a collection of them interacting with a single application.

196
00:14:53,854 --> 00:14:57,874
So the leaders are cognizant that something needs to change

197
00:14:57,874 --> 00:15:01,054
and that they need to modernize their infrastructure.

198
00:15:01,054 --> 00:15:05,014
And they're very self-aware of where they are in their journey.

199
00:15:06,174 --> 00:15:10,554
And so studies show that 71% of the orgs

200
00:15:10,554 --> 00:15:13,614
are modernizing their core infrastructure to support AI.

201
00:15:14,874 --> 00:15:22,194
and 6% to 10% or 23% of them are investing 6% to 10% of their annual revenue

202
00:15:22,194 --> 00:15:25,954
in core system modernization to take advantage of this.

203
00:15:26,234 --> 00:15:31,854
And a common complaint here or a common challenge here is data quality

204
00:15:31,854 --> 00:15:33,714
and the ability to reuse data.

205
00:15:35,694 --> 00:15:39,214
So, great. We have the right people in the room.

206
00:15:39,214 --> 00:15:43,634
We have the right knowledge there.

207
00:15:43,874 --> 00:15:46,014
What do we do next?

208
00:15:46,314 --> 00:15:50,794
And then comes the concept of, hey, how do we make this knowledge reusable and accessible?

209
00:15:51,154 --> 00:15:58,074
And that's where we have, depending on the adoption and where they are, you can have different types of team agents,

210
00:15:58,214 --> 00:16:01,694
where you can have expert agents being subject matter experts.

211
00:16:01,734 --> 00:16:08,214
You can have workflow agents that work across streams like Slack bots that are creating and opening tickets.

212
00:16:08,214 --> 00:16:10,534
they're closing them, and what have you.

213
00:16:10,654 --> 00:16:13,114
And then you can have really chief of staff kind of agents

214
00:16:13,114 --> 00:16:16,914
that keep a tab on what the state of the organization is

215
00:16:16,914 --> 00:16:17,914
and what needs to be done.

216
00:16:19,294 --> 00:16:21,654
So, you know, going back to our workflows,

217
00:16:22,094 --> 00:16:24,334
A, again, is looking at this very differently

218
00:16:24,334 --> 00:16:25,834
in terms of shared capabilities.

219
00:16:26,594 --> 00:16:30,694
Now, the last pillar here that I want to draw your attention to is,

220
00:16:31,674 --> 00:16:34,674
well, when does this system really compound?

221
00:16:34,674 --> 00:16:37,474
What happens when it's deployed?

222
00:16:38,514 --> 00:16:43,954
Well, the leading companies, what they're doing is they're actually thinking about it slightly differently.

223
00:16:44,294 --> 00:16:48,834
What are failures that can give us an edge that we can learn from?

224
00:16:48,994 --> 00:16:51,734
If there's a success, how do we make it part of the system?

225
00:16:52,834 --> 00:16:57,454
And the way they think about this is, you know, the individual learned but did the system.

226
00:16:57,694 --> 00:17:02,194
Is there a way, perhaps, when they've encountered something where they need to make a correction,

227
00:17:02,194 --> 00:17:04,814
that that knowledge is fed back into the org,

228
00:17:05,014 --> 00:17:09,094
either through a new tool or context or what have you.

229
00:17:09,314 --> 00:17:14,994
And what's key here is preserving the evidence along with those learnings.

230
00:17:16,934 --> 00:17:22,894
So really the last vertical here is AI freeze capacity.

231
00:17:23,154 --> 00:17:23,934
Where does it go?

232
00:17:24,214 --> 00:17:27,994
Being cognizant that you need to redesign your system

233
00:17:27,994 --> 00:17:33,514
and really focus it on key metrics that you're evaluating,

234
00:17:33,854 --> 00:17:37,174
you're monitoring, and you're creating a flywheel off.

235
00:17:37,814 --> 00:17:41,134
So where does this leave us for companies A and B?

236
00:17:41,634 --> 00:17:45,394
You know, company A has their entire work redesigned.

237
00:17:45,634 --> 00:17:48,414
They're building capabilities at Compound,

238
00:17:48,414 --> 00:17:50,814
and they have shared capabilities built.

239
00:17:51,594 --> 00:17:54,034
So I want to leave you off with one final question.

240
00:17:54,594 --> 00:17:57,094
Everyone has AI. Who has the advantage?

241
00:17:57,994 --> 00:17:58,274
Thank you.