AI at Scale — Investment, Adoption, and Leadership — Transcript

Full transcript

1. 0:00Now, let's look at how large
2. 0:02organizations are investing in AI and
3. 0:04why leadership determines whether that
4. 0:06investment actually pays off. Here's the
5. 0:08headline. The investment is not
6. 0:10hypothetical. It's already the largest
7. 0:12capital deployment in modern history. In
8. 0:152026 alone, the four biggest tech
9. 0:17companies will spend over 715 billion
10. 0:21dollars on AI infrastructure. To put
11. 0:24that in perspective, the Apollo program,
12. 0:26which put a man on the moon, cost 321
13. 0:29billion dollars over 13 years.
14. 0:32The entire US Interstate Highway System
15. 0:35cost 705
16. 0:37billion dollars over 36 years. AI is
17. 0:41matching that in 12 months. That capital
18. 0:44is buying capability that's improving
19. 0:46faster than almost any technology we've
20. 0:49measured.
21. 0:50>> [music]
22. 0:50>> AI agent capabilities are doubling
23. 0:53roughly every 7 months. This is faster
24. 0:56than Moore's law, which estimates a
25. 0:58doubling of density every 2 years for
26. 1:01computer chips.
27. 1:02>> [music]
28. 1:02>> In 2022, ChatGPT could complete a coding
29. 1:05task that took 30 seconds. [music] By
30. 1:072026, autonomous agents complete 12-hour
31. 1:10workflows.
32. 1:12Projections put us at a full work month
33. 1:14of autonomous task completion by 2029.
34. 1:17At the same time, the cost of running
35. 1:19these systems is collapsing, over 280
36. 1:21times cheaper in just 24 months. Put
37. 1:24those two trends together, capability
38. 1:26rising, cost falling. That combination
39. 1:29is why I keep telling people this isn't
40. 1:30optional to understand. It's not a
41. 1:32question of if this reaches your field,
42. 1:34it's when. Here's the part most people
43. 1:36miss. AI capability is improving
44. 1:39exponentially. Organizations absorb
45. 1:41change linearly through training,
46. 1:43restructuring, trust building, habit
47. 1:45change. Draw those two curves on the
48. 1:47same chart, and the gap between them
49. 1:49just keeps widening. That gap is where
50. 1:51most companies actually struggle. Not
51. 1:54because the technology doesn't work,
52. 1:56because the organization hasn't caught
53. 1:58up to what the technology can already
54. 2:00do. This is where I see organizations
55. 2:02split into two groups.
56. 2:03The first group adds AI on top of what
57. 2:05they already do, a dashboard or tool
58. 2:08bolted onto an existing process without
59. 2:10changing how the work actually flows.
60. 2:12The second group embeds AI into the
61. 2:14workflow itself, changing what action
62. 2:15happens, who's involved, and what
63. 2:17triggers a response.
64. 2:18>> [music]
65. 2:19>> Look at Anticimex, a pest control
66. 2:20company. They didn't just add sensors to
67. 2:23their existing inspection schedule. They
68. 2:25rebuilt the entire operating model.
69. 2:27Instead of a technician visiting a site
70. 2:30because the calendar said so, their
71. 2:31smart system uses [music] sensors,
72. 2:34traps, and cameras to monitor conditions
73. 2:36continuously and dispatches a visit only
74. 2:40when the data actually shows a need.
75. 2:43>> That's a shift from scheduled service to
76. 2:44condition-based service, and critically,
77. 2:46the AI augments their field experts
78. 2:48rather than replacing them.
79. 2:50A person still makes the call just with
80. 2:52far better information.
81. 2:53>> [music]
82. 2:53>> The result, 150,000 traps and sensors
83. 2:56deployed across 18 markets. That's not a
84. 2:58tool sitting on top of the old process.
85. 3:01That's a redesigned workflow.
86. 3:02J.P. Morgan Chase is a good example of
87. 3:04what this looks like in enterprise
88. 3:05scale. Their AI assistant, Coach, didn't
89. 3:08stay a simple chatbot. It evolved into a
90. 3:10proactive partner servicing client
91. 3:12insights before an advisor even asks,
92. 3:14drawing on over 100 integrated internal
93. 3:17data systems to answer questions in
94. 3:18seconds. Today, over 165,000
95. 3:21employees use it daily, [music] and
96. 3:23advisors report roughly 50% productivity
97. 3:26improvement on certain tasks. What made
98. 3:28the difference wasn't the tool. It was
99. 3:30top-down commitment paired with
100. 3:31bottom-up innovation. Leadership
101. 3:33prioritized the redesign, while
102. 3:35employees built and shared over 100,000
103. 3:37custom AI assistants of their own.
104. 3:40Regardless of the specific technology,
105. 3:42embedded AI tends to do one of three
106. 3:44things inside a workflow, predict an
107. 3:46outcome, recognize a pattern, or
108. 3:48automate a process.
109. 3:49That's the functional core, whether
110. 3:51you're looking at a bank, a pest control
111. 3:52company, or your own workflow. One
112. 3:54honest caveat, this path isn't a
113. 3:56straight line up. Research shows AI
114. 3:59adoption often follows a J curve.
115. 4:01Productivity actually dips initially,
116. 4:02sometimes for a year or more, before it
117. 4:04climbs. Older, more established
118. 4:06organizations tend to see steeper
119. 4:08short-term losses since more of their
120. 4:10existing practices get disrupted. The
121. 4:12companies that recover fastest share one
122. 4:14trait. Leadership stays committed
123. 4:15through the dip instead of abandoning
124. 4:17the effort at the first sign of
125. 4:18friction. And it matters where you apply
126. 4:21AI. In one large study, generative AI
127. 4:24improved task performance by over 12%
128. 4:27when used on the right kind of task, but
129. 4:29performance actually got worse when
130. 4:31people applied it to tasks outside AI's
131. 4:33current capability.
132. 4:34The tool isn't universally helpful.
133. 4:36Knowing where it helps and where it
134. 4:38doesn't yet is part of the leadership
135. 4:39judgment required here. None of this
136. 4:41happens in a vacuum of pure logic.
137. 4:44People react to AI differently and
138. 4:46predictably so. Some people feel
139. 4:48inspired. They believe in AI's potential
140. 4:50and feel bought into using it.
141. 4:52Others feel oppressed, capable of
142. 4:53succeeding perhaps, but resentful of
143. 4:55being pushed into it. Some are
144. 4:57indifferent, low belief, low investment
145. 4:59either way. And some feel frustrated.
146. 5:01They want to succeed with it, but don't
147. 5:03yet feel equipped to. None of these
148. 5:05reactions is wrong. They're data.
149. 5:07A leader's job is to notice where their
150. 5:08people actually are, not assume everyone
151. 5:11starts from curiosity. So, what does
152. 5:13leadership actually look like through
153. 5:14this?
154. 5:15Five things in sequence. [music]
155. 5:17Learning, genuinely understanding what
156. 5:20AI means for your business, not just
157. 5:21hearing about it. Doing, because AI gets
158. 5:24built and adapted by you, not delivered
159. 5:27finished to you. Imagining, using it to
160. 5:29invent new operating models, not just
161. 5:31automate old ones. Acting, building
162. 5:34change resilience and new skills for
163. 5:36everyone, not just technical teams. And
164. 5:38caring, [music] directly addressing
165. 5:40people's fears about job loss and
166. 5:41identity, rather than pretending those
167. 5:43fears don't exist.
168. 5:45Skip the last one and you'll have the
169. 5:46best technology in your industry and a
170. 5:48workforce that quietly resisted the
171. 5:50whole way. The investment is real, the
172. 5:52capability is compounding, and the gap
173. 5:54between what AI can do and what
174. 5:55organizations actually absorb is where
175. 5:57leadership matters most. Before you move
176. 5:59on, take a moment honestly with
177. 6:01yourself. Where do you currently sit?
178. 6:03Inspired, frustrated, indifferent, or
179. 6:05oppressed when it comes to AI?
180. 6:08How much do you believe you could
181. 6:09succeed with it if you tried? And
182. 6:11separately, how much are you actually
183. 6:13bought in? Your answers today aren't
184. 6:14fixed, but naming them honestly is the
185. 6:17first step of the learning stage I just
186. 6:19described for you, not just for the
187. 6:21organizations we've talked about. Next,
188. 6:23we'll consider some ethical principles
189. 6:25to help guide responsible work with AI.