Work IQ: The New Intelligence Layer of Microsoft 365

1
00:00:00,000 --> 00:00:02,520
Microsoft just made a quiet move that changes everything.

2
00:00:02,520 --> 00:00:05,040
They stopped treating your Microsoft 365 tenant

3
00:00:05,040 --> 00:00:07,000
as a storage system and started treating it

4
00:00:07,000 --> 00:00:08,120
as a reasoning system.

5
00:00:08,120 --> 00:00:10,960
It is no longer a place where you go to search for files.

6
00:00:10,960 --> 00:00:12,520
It is now a system that understands

7
00:00:12,520 --> 00:00:14,000
what those files actually mean.

8
00:00:14,000 --> 00:00:15,260
The difference isn't small.

9
00:00:15,260 --> 00:00:16,960
This shift restructures your governance,

10
00:00:16,960 --> 00:00:19,800
it exposes your culture and it forces your organization

11
00:00:19,800 --> 00:00:21,560
to confront decision-making in ways

12
00:00:21,560 --> 00:00:23,160
you have been avoiding for years.

13
00:00:23,160 --> 00:00:24,800
We call this work IQ.

14
00:00:24,800 --> 00:00:27,440
Most leaders do not realize this is already happening,

15
00:00:27,440 --> 00:00:29,800
but this episode diagnosis exactly what is breaking

16
00:00:29,800 --> 00:00:32,920
and why the old model cannot survive what comes next.

17
00:00:32,920 --> 00:00:34,720
The productivity measurement crisis.

18
00:00:34,720 --> 00:00:36,280
Your organization is busy.

19
00:00:36,280 --> 00:00:37,320
That is not the problem.

20
00:00:37,320 --> 00:00:38,840
The problem is that you have no idea

21
00:00:38,840 --> 00:00:41,560
whether being busy actually means being intelligent.

22
00:00:41,560 --> 00:00:43,840
Think about what you actually measure right now.

23
00:00:43,840 --> 00:00:45,600
You look at email volume, meeting hours

24
00:00:45,600 --> 00:00:47,800
and task completion rates because those are the metrics

25
00:00:47,800 --> 00:00:50,760
on your dashboards and the numbers your executive C.

26
00:00:50,760 --> 00:00:52,440
But those numbers have almost nothing to do

27
00:00:52,440 --> 00:00:55,040
with whether your organization is getting smarter or dumber.

28
00:00:55,040 --> 00:00:56,080
They measure activity.

29
00:00:56,080 --> 00:00:57,440
They do not measure value.

30
00:00:57,440 --> 00:00:59,360
This gap between being busy and being effective

31
00:00:59,360 --> 00:01:00,640
has become structural.

32
00:01:00,640 --> 00:01:03,600
It is baked into how most companies operate today.

33
00:01:03,600 --> 00:01:06,280
You can have a person sitting in back-to-back meetings all day

34
00:01:06,280 --> 00:01:08,200
responding to emails at 11 p.m.

35
00:01:08,200 --> 00:01:10,120
and closing every task on time.

36
00:01:10,120 --> 00:01:12,280
While your critical decisions are still getting made slower

37
00:01:12,280 --> 00:01:12,880
than ever.

38
00:01:12,880 --> 00:01:14,760
Context gets lost every time someone switches

39
00:01:14,760 --> 00:01:16,760
between outlook and teams and teams end up

40
00:01:16,760 --> 00:01:18,800
solving the same problem five different times

41
00:01:18,800 --> 00:01:21,080
because nobody knows a solution already exists.

42
00:01:21,080 --> 00:01:22,280
Three floors down.

43
00:01:22,280 --> 00:01:23,360
Here is the brutal fact.

44
00:01:23,360 --> 00:01:26,480
95% of enterprises see little to no measurable impact

45
00:01:26,480 --> 00:01:27,800
from AI right now.

46
00:01:27,800 --> 00:01:29,640
This is not because the AI is bad,

47
00:01:29,640 --> 00:01:31,800
but because leaders are still using activity metrics

48
00:01:31,800 --> 00:01:32,760
to measure intelligence.

49
00:01:32,760 --> 00:01:35,040
They deployed a co-pilot that costs money to run.

50
00:01:35,040 --> 00:01:36,600
They trained people to use it.

51
00:01:36,600 --> 00:01:39,800
And then they measured success by how many people opened the app.

52
00:01:39,800 --> 00:01:42,240
They did not measure whether decisions got made faster

53
00:01:42,240 --> 00:01:44,320
or if rework went down or if people finally

54
00:01:44,320 --> 00:01:45,480
stopped searching for information

55
00:01:45,480 --> 00:01:47,280
because the system understood what they needed

56
00:01:47,280 --> 00:01:48,800
before they even asked.

57
00:01:48,800 --> 00:01:51,760
Work IQ forces a reckoning with this broken system

58
00:01:51,760 --> 00:01:53,520
because work IQ does not hide.

59
00:01:53,520 --> 00:01:55,480
It makes what is actually happening visible.

60
00:01:55,480 --> 00:01:57,360
When work IQ starts reasoning over your emails,

61
00:01:57,360 --> 00:01:58,600
your meetings and your files,

62
00:01:58,600 --> 00:02:01,600
it builds a semantic model of how your organization actually works.

63
00:02:01,600 --> 00:02:03,920
Not how you say it works, but how it actually works.

64
00:02:03,920 --> 00:02:05,720
The moment an AI system understands

65
00:02:05,720 --> 00:02:08,000
which documents matter and who is talking to whom

66
00:02:08,000 --> 00:02:10,560
it becomes impossible to hide a broken operating model

67
00:02:10,560 --> 00:02:13,080
behind a dashboard of activity metrics.

68
00:02:13,080 --> 00:02:16,040
The real cost to your business is not the time people spend working.

69
00:02:16,040 --> 00:02:18,800
Time is infinite in a sense because people will always generate more of it.

70
00:02:18,800 --> 00:02:21,560
The real cost is context switching and decision friction.

71
00:02:21,560 --> 00:02:24,240
Value dies every time someone has to stop what they are doing

72
00:02:24,240 --> 00:02:26,480
to search for information they should already have

73
00:02:26,480 --> 00:02:28,280
or when a decision gets delayed

74
00:02:28,280 --> 00:02:31,000
because the right people lack the right context.

75
00:02:31,000 --> 00:02:33,680
Institutional knowledge gets locked in a single person's email

76
00:02:33,680 --> 00:02:35,360
instead of being available to the system

77
00:02:35,360 --> 00:02:37,760
and that friction is where your progress stops.

78
00:02:37,760 --> 00:02:39,680
Work IQ targets that friction directly.

79
00:02:39,680 --> 00:02:42,840
It does not just speed up your email or make your meetings shorter.

80
00:02:42,840 --> 00:02:45,440
It builds a persistent permission away understanding

81
00:02:45,440 --> 00:02:47,560
of what matters and who needs to know it.

82
00:02:47,560 --> 00:02:50,320
That is a completely different category of capability

83
00:02:50,320 --> 00:02:53,080
than just making people faster at the tasks they already do.

84
00:02:53,080 --> 00:02:54,840
But here's what most organizations miss.

85
00:02:54,840 --> 00:02:57,280
You cannot just plug work IQ into a broken system

86
00:02:57,280 --> 00:02:58,640
and expect it to fix things.

87
00:02:58,640 --> 00:03:00,080
In reality, it does the opposite.

88
00:03:00,080 --> 00:03:01,920
It exposes every single weakness you have.

89
00:03:01,920 --> 00:03:04,160
It shows you exactly where context is being lost

90
00:03:04,160 --> 00:03:06,080
and exactly where your decisions are bottlenecked.

91
00:03:06,080 --> 00:03:09,320
It reveals where people have access to information they should never see

92
00:03:09,320 --> 00:03:13,080
and where they are blocked from seeing the information they actually need to do their jobs.

93
00:03:13,080 --> 00:03:15,120
That visibility is uncomfortable.

94
00:03:15,120 --> 00:03:17,200
Once you see the problem, you cannot unsee it

95
00:03:17,200 --> 00:03:18,560
and that means you have to fix it.

96
00:03:18,560 --> 00:03:21,880
You have to redesign how decisions happen and change what you measure.

97
00:03:21,880 --> 00:03:23,800
You have to align your governance structures

98
00:03:23,800 --> 00:03:27,600
with how work actually flows instead of how the org chart says it should flow.

99
00:03:27,600 --> 00:03:29,120
That is what this shift is really about.

100
00:03:29,120 --> 00:03:30,680
It is not about faster productivity.

101
00:03:30,680 --> 00:03:33,880
It is about organizational intelligence finally becoming visible

102
00:03:33,880 --> 00:03:37,400
and organizations usually hate visibility into their own dysfunction.

103
00:03:37,400 --> 00:03:40,120
How Microsoft 365 was built?

104
00:03:40,120 --> 00:03:42,760
To understand why work IQ had to exist,

105
00:03:42,760 --> 00:03:45,880
you first need to look at how Microsoft 365 was actually built.

106
00:03:45,880 --> 00:03:47,800
It was never designed to be a reasoning system.

107
00:03:47,800 --> 00:03:50,600
Instead, the creators built a collection of specialized tools

108
00:03:50,600 --> 00:03:52,760
where each one solved a single specific problem.

109
00:03:52,760 --> 00:03:55,640
Outlook handles your email, teams manages your chats,

110
00:03:55,640 --> 00:03:57,960
and SharePoint takes care of document storage.

111
00:03:57,960 --> 00:04:01,360
Because each tool has its own data model and its own permissions logic,

112
00:04:01,360 --> 00:04:03,000
they stay siloed by design.

113
00:04:03,000 --> 00:04:04,680
They are connected at the surface,

114
00:04:04,680 --> 00:04:07,200
but they are not integrated at the intelligence level.

115
00:04:07,200 --> 00:04:09,520
Microsoft recognized this liability early on,

116
00:04:09,520 --> 00:04:12,480
so they built Microsoft Graph to unify the data access layer.

117
00:04:12,480 --> 00:04:14,720
Graph is a massive technical achievement

118
00:04:14,720 --> 00:04:17,400
because it simplified how applications talk to your data.

119
00:04:17,400 --> 00:04:21,880
Before Graph existed, if you wanted to build an app that touched your calendar and your files,

120
00:04:21,880 --> 00:04:25,000
you had to manage several different authentication flows

121
00:04:25,000 --> 00:04:27,000
and stitch together different data models.

122
00:04:27,000 --> 00:04:30,840
Graph changed that by providing one API and one permission framework for everything,

123
00:04:30,840 --> 00:04:33,000
but here is the problem that Graph didn't solve.

124
00:04:33,000 --> 00:04:34,600
It never unified the reasoning layer.

125
00:04:34,600 --> 00:04:37,480
Graph is excellent at answering one specific question.

126
00:04:37,480 --> 00:04:40,040
What data exists and who is allowed to see it?

127
00:04:40,040 --> 00:04:42,920
It returns objects like files, messages and meetings.

128
00:04:42,920 --> 00:04:46,120
You can ask Graph to show you every file shared with a specific user

129
00:04:46,120 --> 00:04:48,920
or find every email from last year about a certain project.

130
00:04:48,920 --> 00:04:51,000
Graph will find them and respect your permissions,

131
00:04:51,000 --> 00:04:53,400
but it won't tell you why those files actually matter.

132
00:04:53,400 --> 00:04:57,080
The system doesn't understand how these pieces relate to each other in the real world.

133
00:04:57,080 --> 00:05:01,640
It won't realize that three different emails are describing the same problem from three different angles.

134
00:05:01,640 --> 00:05:04,680
It won't know that the solution is actually buried in a SharePoint document

135
00:05:04,680 --> 00:05:07,880
that nobody thought to search for because it wasn't tagged correctly.

136
00:05:07,880 --> 00:05:10,280
You can retrieve information from Microsoft 365,

137
00:05:10,280 --> 00:05:13,320
but the system has no idea why that information is important to your team.

138
00:05:13,320 --> 00:05:17,480
This architecture is the reason organizations became data rich but intelligence poor.

139
00:05:17,480 --> 00:05:20,440
You can dump every meeting transcript and document into the system,

140
00:05:20,440 --> 00:05:22,200
and you can index it all for search.

141
00:05:22,200 --> 00:05:25,320
But the system itself has no memory and no understanding of patterns.

142
00:05:25,320 --> 00:05:28,680
Every time a user interacts with it, the process starts from zero.

143
00:05:28,680 --> 00:05:29,960
Every search is independent,

144
00:05:29,960 --> 00:05:32,280
and every application has to rebuild context from scratch

145
00:05:32,280 --> 00:05:35,480
because Graph simply wasn't built to do that work for you.

146
00:05:35,480 --> 00:05:37,480
The original design reflected in all the assumption

147
00:05:37,480 --> 00:05:39,880
that people always know exactly what they are looking for.

148
00:05:39,880 --> 00:05:41,720
If you need an email, you search for it,

149
00:05:41,720 --> 00:05:44,440
and if you need to schedule a meeting, you open your calendar.

150
00:05:44,440 --> 00:05:46,680
The system was just a digital filing cabinet.

151
00:05:46,680 --> 00:05:48,760
Its only job was to store and retrieve

152
00:05:48,760 --> 00:05:51,400
while your job was to know what you needed and ask for it.

153
00:05:51,400 --> 00:05:54,680
That logic made sense back in 2016, but it doesn't work anymore.

154
00:05:54,680 --> 00:05:57,640
Work has become too fragmented for that model to hold up,

155
00:05:57,640 --> 00:05:59,560
information lives in too many places,

156
00:05:59,560 --> 00:06:02,440
and making a good decision now requires connecting signals

157
00:06:02,440 --> 00:06:05,000
from email, meetings and collaboration patterns.

158
00:06:05,000 --> 00:06:07,480
Right now, the system just hands you the raw material

159
00:06:07,480 --> 00:06:09,800
and expects you to do all the reasoning yourself.

160
00:06:09,800 --> 00:06:11,240
It tells you to figure it out.

161
00:06:11,240 --> 00:06:13,480
And that is exactly where the model breaks.

162
00:06:13,480 --> 00:06:15,160
The Graph backbone.

163
00:06:15,160 --> 00:06:17,160
Microsoft Graph is the actual infrastructure

164
00:06:17,160 --> 00:06:18,760
that runs through the entire suite.

165
00:06:18,760 --> 00:06:21,160
It isn't just a marketing term, it is the data spine

166
00:06:21,160 --> 00:06:24,040
for every user, file and email in your organization.

167
00:06:24,040 --> 00:06:27,160
Graph exposes all of that data and enforces the permissions on it.

168
00:06:27,160 --> 00:06:30,280
It is the backbone that holds the entire ecosystem together.

169
00:06:30,280 --> 00:06:32,600
Graph is very good at answering who can access what?

170
00:06:32,600 --> 00:06:35,240
When you tell Graph to get a user's emails from the last month,

171
00:06:35,240 --> 00:06:37,720
it finds them quickly and checks the permissions.

172
00:06:37,720 --> 00:06:40,760
It is fast and consistent, which is why developers can build apps

173
00:06:40,760 --> 00:06:43,160
that touch multiple services without rewriting

174
00:06:43,160 --> 00:06:45,240
the security logic a dozen times.

175
00:06:45,240 --> 00:06:47,480
It provides a solid foundation for data access,

176
00:06:47,480 --> 00:06:49,400
but that is where its capabilities end.

177
00:06:49,400 --> 00:06:51,880
The big issue is that Graph is completely stateless.

178
00:06:51,880 --> 00:06:54,520
Every single query you send starts from a blank slate.

179
00:06:54,520 --> 00:06:56,680
Graph doesn't remember what you asked a minute ago

180
00:06:56,680 --> 00:06:58,840
and it has no idea what you were working on yesterday.

181
00:06:58,840 --> 00:07:00,600
It won't connect an email to a document

182
00:07:00,600 --> 00:07:03,160
and a meeting to say they are all part of the same project.

183
00:07:03,160 --> 00:07:05,000
Graph returns exactly what you ask for,

184
00:07:05,000 --> 00:07:07,960
but it never infers what you actually need to get your job done.

185
00:07:07,960 --> 00:07:10,600
If a developer wants to build an app on top of Graph,

186
00:07:10,600 --> 00:07:12,760
they have to do all the heavy lifting themselves.

187
00:07:12,760 --> 00:07:14,600
To build a tool that summarizes a project,

188
00:07:14,600 --> 00:07:16,760
you have to call the API five different times

189
00:07:16,760 --> 00:07:19,080
to get the emails, meetings, and people involved,

190
00:07:19,080 --> 00:07:22,200
then you have to write custom code to connect those dots.

191
00:07:22,200 --> 00:07:24,520
You have to teach the app which documents were discussed

192
00:07:24,520 --> 00:07:27,160
in which meetings, because Graph just hands you raw data

193
00:07:27,160 --> 00:07:29,960
and walks away, because there is no shared understanding.

194
00:07:29,960 --> 00:07:32,600
Every application ends up rebuilding context separately.

195
00:07:32,600 --> 00:07:35,160
Your calendar tool, your email client and your project manager

196
00:07:35,160 --> 00:07:37,560
are all trying to figure out the same thing in isolation.

197
00:07:37,560 --> 00:07:39,000
There is no organizational memory.

198
00:07:39,000 --> 00:07:41,320
There is no persistent model that understands

199
00:07:41,320 --> 00:07:43,160
how signals from different systems

200
00:07:43,160 --> 00:07:44,600
actually connect to each other.

201
00:07:44,600 --> 00:07:46,840
This is how you end up with perfect data access

202
00:07:46,840 --> 00:07:49,400
but zero organizational intelligence.

203
00:07:49,400 --> 00:07:51,960
Graph handles the first problem of access beautifully,

204
00:07:51,960 --> 00:07:54,200
but it doesn't solve the second problem of meaning.

205
00:07:54,200 --> 00:07:55,720
It gives you the what?

206
00:07:55,720 --> 00:07:57,000
But never the why.

207
00:07:57,000 --> 00:07:59,080
Think about how this feels for you as a user.

208
00:07:59,080 --> 00:08:01,000
When you are in a meeting and need to understand

209
00:08:01,000 --> 00:08:02,040
the history of a project,

210
00:08:02,040 --> 00:08:04,120
you have to manually search through old notes

211
00:08:04,120 --> 00:08:06,200
and scroll through endless email threads.

212
00:08:06,200 --> 00:08:07,880
You are the one piecing the story together

213
00:08:07,880 --> 00:08:09,480
because the system doesn't maintain a model

214
00:08:09,480 --> 00:08:11,240
of what was decided or what comes next.

215
00:08:11,240 --> 00:08:14,120
It doesn't remember that you care about this specific project

216
00:08:14,120 --> 00:08:15,960
and it doesn't connect today's conversation

217
00:08:15,960 --> 00:08:17,320
to a decision made last week.

218
00:08:17,320 --> 00:08:18,840
Graph is a powerful foundation,

219
00:08:18,840 --> 00:08:20,520
but having power over data is not the same

220
00:08:20,520 --> 00:08:22,040
as having intelligence about work.

221
00:08:22,040 --> 00:08:24,200
Graph gives you the access you need to find files.

222
00:08:24,200 --> 00:08:26,360
Work IQ gives you the understanding you need

223
00:08:26,360 --> 00:08:28,360
to actually finish the job.

224
00:08:28,360 --> 00:08:29,640
Enter Work IQ.

225
00:08:29,640 --> 00:08:32,360
Microsoft didn't solve the graph problem by replacing it

226
00:08:32,360 --> 00:08:34,840
but by building something entirely new on top of it.

227
00:08:34,840 --> 00:08:37,480
Work IQ sits directly on that graph backbone

228
00:08:37,480 --> 00:08:39,800
and adds the one thing graph was always missing.

229
00:08:39,800 --> 00:08:42,680
It uses three distinct layers to turn your raw data

230
00:08:42,680 --> 00:08:44,840
into actual organizational understanding.

231
00:08:44,840 --> 00:08:47,000
The first layer is all about the data itself.

232
00:08:47,000 --> 00:08:50,200
Work IQ takes permission away access to everything graph sees,

233
00:08:50,200 --> 00:08:52,440
including your emails, meetings, files

234
00:08:52,440 --> 00:08:54,840
and even the subtle collaboration signals between teams.

235
00:08:54,840 --> 00:08:56,760
It feeds all of this into a semantic engine

236
00:08:56,760 --> 00:08:57,880
that never stops learning.

237
00:08:57,880 --> 00:08:59,640
This isn't just a basic index of your files.

238
00:08:59,640 --> 00:09:02,040
It's an active analysis that builds associations

239
00:09:02,040 --> 00:09:03,480
and learns what matters by watching

240
00:09:03,480 --> 00:09:06,040
how people actually interact with information every day.

241
00:09:06,040 --> 00:09:07,480
Then we move into the context layer,

242
00:09:07,480 --> 00:09:09,160
which is where the real reasoning happens.

243
00:09:09,160 --> 00:09:11,720
Work IQ develops a deep understanding of relationships

244
00:09:11,720 --> 00:09:13,560
that goes far beyond simple connections.

245
00:09:13,560 --> 00:09:16,040
It doesn't just see that person a emailed person b.

246
00:09:16,040 --> 00:09:18,360
Instead it understands that both people are collaborating

247
00:09:18,360 --> 00:09:21,560
on project gamma and have been debating risk mitigation for weeks.

248
00:09:21,560 --> 00:09:24,920
It knows there are three open decisions waiting on the finance team.

249
00:09:24,920 --> 00:09:27,960
It understands your projects, your workflows and your priorities.

250
00:09:27,960 --> 00:09:30,920
It sees how decisions flow and where information gets stuck.

251
00:09:30,920 --> 00:09:32,520
This builds a model of your company

252
00:09:32,520 --> 00:09:35,160
that makes the official org chart look like a toy.

253
00:09:35,160 --> 00:09:37,720
An org chart tells you who reports to whom

254
00:09:37,720 --> 00:09:40,200
but it can't tell you who actually drives the decisions

255
00:09:40,200 --> 00:09:43,400
or who sits at the center of the most critical information flows.

256
00:09:43,400 --> 00:09:45,000
The third layer is the memory layer

257
00:09:45,000 --> 00:09:47,160
and this is the part graph could never do on its own.

258
00:09:47,160 --> 00:09:50,600
Work IQ maintains a persistent and evolving understanding

259
00:09:50,600 --> 00:09:53,080
of how every person and team actually gets things done.

260
00:09:53,080 --> 00:09:55,080
It remembers your typical collaborators

261
00:09:55,080 --> 00:09:56,920
and the projects you usually touch.

262
00:09:56,920 --> 00:09:58,680
It learns your communication style

263
00:09:58,680 --> 00:10:00,360
and how you prefer to make decisions.

264
00:10:00,360 --> 00:10:01,960
It knows what information you need

265
00:10:01,960 --> 00:10:03,960
and exactly when you usually need it.

266
00:10:03,960 --> 00:10:05,320
This isn't just for one session.

267
00:10:05,320 --> 00:10:09,320
It remembers across weeks and months to build a behavioral profile of your work.

268
00:10:09,320 --> 00:10:10,440
The goal isn't surveillance

269
00:10:10,440 --> 00:10:12,520
but rather understanding you well enough to help.

270
00:10:12,520 --> 00:10:14,680
When an agent tries to assist you,

271
00:10:14,680 --> 00:10:16,600
it can make a smart guess about what you need

272
00:10:16,600 --> 00:10:19,080
instead of a stupid guess based on a keyword search.

273
00:10:19,080 --> 00:10:21,080
This combination creates a work graph

274
00:10:21,080 --> 00:10:24,840
which is fundamentally different from the entity graph you get with standard tools.

275
00:10:24,840 --> 00:10:29,000
It is a living breathing model of how your organization functions in the real world.

276
00:10:29,000 --> 00:10:32,200
It tracks who works with whom, why a specific task matters

277
00:10:32,200 --> 00:10:33,720
and what needs to happen next.

278
00:10:33,720 --> 00:10:35,640
It maps the bottlenecks and the decision flows

279
00:10:35,640 --> 00:10:37,000
that actually run the business.

280
00:10:37,000 --> 00:10:39,720
This shift is crucial for anyone building applications at scale,

281
00:10:39,720 --> 00:10:40,920
especially with co-pilot.

282
00:10:40,920 --> 00:10:43,000
Because Work IQ has already done the heavy lifting,

283
00:10:43,000 --> 00:10:45,000
the system now has memory and context.

284
00:10:45,000 --> 00:10:48,600
Co-pilot doesn't have to start from zero every time you ask a question.

285
00:10:48,600 --> 00:10:51,960
It doesn't need to be retrained on your specific world every single morning.

286
00:10:51,960 --> 00:10:54,680
Work IQ has already built that persistent model of what matters.

287
00:10:54,680 --> 00:10:56,040
It knows your pending decisions

288
00:10:56,040 --> 00:10:57,560
and what you've tried in the past.

289
00:10:57,560 --> 00:11:00,360
It even knows why similar problems weren't solved years ago.

290
00:11:00,360 --> 00:11:02,920
This is the real reason co-pilot works at scale.

291
00:11:02,920 --> 00:11:05,560
It isn't just that the large language model got smarter.

292
00:11:05,560 --> 00:11:09,400
It's because the intelligence layer underneath it became persistent and contextual.

293
00:11:09,400 --> 00:11:11,800
Co-pilot isn't just looking at raw text anymore.

294
00:11:11,800 --> 00:11:14,440
It is reasoning over an organizational understanding

295
00:11:14,440 --> 00:11:17,560
that Work IQ has already synthesized into something useful.

296
00:11:17,560 --> 00:11:20,440
The power here comes from the fact that Work IQ learns from you.

297
00:11:20,440 --> 00:11:23,480
It improves over time and gets better at predicting what you'll need next.

298
00:11:23,480 --> 00:11:26,200
It connects scattered signals into a coherent picture

299
00:11:26,200 --> 00:11:29,160
and routes information to the right people before they even ask for it.

300
00:11:29,160 --> 00:11:30,440
The system isn't static.

301
00:11:30,440 --> 00:11:34,600
It is constantly refining its own model of how work actually happens in your office.

302
00:11:34,600 --> 00:11:38,360
Moving from simple data retrieval to persistent reasoning changes everything.

303
00:11:38,360 --> 00:11:42,120
It changes what is possible, what is required, and what is at risk.

304
00:11:42,120 --> 00:11:46,600
When you have an intelligence layer that understands your company better than your executives do,

305
00:11:46,600 --> 00:11:48,440
governance becomes a massive priority.

306
00:11:48,440 --> 00:11:51,800
This system sees patterns that nobody has the time to track manually.

307
00:11:51,800 --> 00:11:57,080
Your risks become visible and your organizational flaws are suddenly exposed for everyone to see.

308
00:11:57,080 --> 00:11:59,000
The aggregation problem.

309
00:11:59,000 --> 00:12:03,160
We need to talk about what happens when you concentrate all that intelligence in one single place.

310
00:12:03,160 --> 00:12:05,000
That is exactly what Work IQ does.

311
00:12:05,000 --> 00:12:06,600
It doesn't just build understanding.

312
00:12:06,600 --> 00:12:09,720
It consolidates every signal your company produces.

313
00:12:09,720 --> 00:12:13,240
It pulls from Outlook, Teams, SharePoint, and your OneDrive.

314
00:12:13,240 --> 00:12:16,520
It looks at your directory and every collaboration pattern it can measure.

315
00:12:16,520 --> 00:12:21,080
It ingests and synthesizes all of it into one continuous model of your organization.

316
00:12:21,080 --> 00:12:23,480
That concentration of data is incredibly powerful,

317
00:12:23,480 --> 00:12:25,080
but it also creates a massive problem.

318
00:12:25,080 --> 00:12:28,520
Work IQ creates a high value target because it aggregates everything.

319
00:12:28,520 --> 00:12:31,720
Before this, your risks were distributed across different silos.

320
00:12:31,720 --> 00:12:34,840
Your email lived in one place while your documents lived in another.

321
00:12:34,840 --> 00:12:37,560
Your conversations and your relationships were all separated.

322
00:12:37,560 --> 00:12:40,600
If permissions were broken in one system, it was a contained problem.

323
00:12:40,600 --> 00:12:44,360
Now imagine an AI agent that understands all of those systems at the same time.

324
00:12:44,360 --> 00:12:47,640
This agent can instantly see every relationship, every document,

325
00:12:47,640 --> 00:12:50,760
and every thread of context that exists anywhere in your company.

326
00:12:50,760 --> 00:12:54,200
If your permission model is weak, work IQ won't hide it.

327
00:12:54,200 --> 00:12:56,600
It will expose that weakness to your agents immediately.

328
00:12:56,600 --> 00:13:01,800
An agent can now discover exactly which sensitive documents a user technically has access to,

329
00:13:01,800 --> 00:13:02,680
even if they shouldn't.

330
00:13:02,680 --> 00:13:06,680
It doesn't matter if it's a legacy permission from five years ago that nobody cleaned up.

331
00:13:06,680 --> 00:13:09,400
The semantic layer in Work IQ will find the connection.

332
00:13:09,400 --> 00:13:13,640
It will realize the document relates to the user's current task and surface it.

333
00:13:13,640 --> 00:13:15,080
That isn't a helpful feature.

334
00:13:15,080 --> 00:13:17,080
It's a liability waiting to explode.

335
00:13:17,080 --> 00:13:19,320
Think about the oversharing problem in SharePoint.

336
00:13:19,320 --> 00:13:22,120
In most companies, oversharing isn't seen as a governance failure,

337
00:13:22,120 --> 00:13:23,800
but as a normal way of working.

338
00:13:23,800 --> 00:13:26,840
Someone creates a site and shares it with everyone just to make things easy.

339
00:13:26,840 --> 00:13:29,880
Years go by and that site becomes part of the institutional memory.

340
00:13:29,880 --> 00:13:35,560
Nobody audits the access or revisits the policy because leaving it open is easier than managing it.

341
00:13:35,560 --> 00:13:39,800
In the old world, this was just a boring compliance issue for the auditors to flag.

342
00:13:39,800 --> 00:13:43,080
With Work IQ, that oversharing becomes an operational nightmare.

343
00:13:43,080 --> 00:13:47,320
An agent can instantly discover that sensitive financial plans are accessible to half the company

344
00:13:47,320 --> 00:13:49,880
because of one lazy decision made three years ago.

345
00:13:49,880 --> 00:13:52,200
The agent might not intentionally leak the info,

346
00:13:52,200 --> 00:13:56,360
but the fact that it can see the data means your risk is now visible and measurable.

347
00:13:56,360 --> 00:13:57,880
It becomes impossible to ignore.

348
00:13:57,880 --> 00:14:00,520
This is the point where most organizations start to panic.

349
00:14:00,520 --> 00:14:04,600
The moment Work IQ starts looking at your tenant, it reveals what is actually in there.

350
00:14:04,600 --> 00:14:06,040
It doesn't care what you think is in there.

351
00:14:06,040 --> 00:14:07,320
It finds the dark matter.

352
00:14:07,320 --> 00:14:12,120
It finds the stale SharePoint sites and the often teams groups that nobody remembered to archive.

353
00:14:12,120 --> 00:14:14,120
It sees the broken permission inheritance

354
00:14:14,120 --> 00:14:17,480
and the file sitting in personal folders that should be in a govern space.

355
00:14:17,480 --> 00:14:21,560
It even sees the confidential email threads that were forwarded to the wrong people.

356
00:14:21,560 --> 00:14:24,760
All of that dark matter exists in every real organization

357
00:14:24,760 --> 00:14:27,160
and it all becomes visible to Work IQ.

358
00:14:27,160 --> 00:14:31,240
The system doesn't see these as compliance violations but as operational reality.

359
00:14:31,240 --> 00:14:34,440
It's just another factor the system has to account for when it tries to help you.

360
00:14:34,440 --> 00:14:37,480
The AI can't ignore it and it can't pretend the data isn't there.

361
00:14:37,480 --> 00:14:39,880
You can no longer hide your data hygiene problems.

362
00:14:39,880 --> 00:14:42,920
Not from Work IQ and certainly not from the agents that use it.

363
00:14:42,920 --> 00:14:46,840
The visibility is now complete and when everything is visible you are forced to make a choice.

364
00:14:46,840 --> 00:14:49,880
You either fix the underlying problem or you choose to accept the risk.

365
00:14:49,880 --> 00:14:52,040
You can't pretend the danger doesn't exist anymore.

366
00:14:52,040 --> 00:14:55,800
This is exactly why governance has to happen before you ever hit the deploy button.

367
00:14:55,800 --> 00:14:56,840
The permission model.

368
00:14:56,840 --> 00:15:00,520
Microsoft made a very specific design choice here that changes everything.

369
00:15:00,520 --> 00:15:03,560
Work IQ does not introduce a brand new security model

370
00:15:03,560 --> 00:15:06,600
and it doesn't try to build a separate framework for access control.

371
00:15:06,600 --> 00:15:11,400
It never bypasses the existing permission structures you already have inside Microsoft 365.

372
00:15:11,400 --> 00:15:15,160
Instead the system operates entirely within those established boundaries.

373
00:15:15,160 --> 00:15:17,800
Every agent you build and every query a user makes

374
00:15:17,800 --> 00:15:22,440
runs through the same identity layers that protect your email, your files and your team's chats.

375
00:15:22,440 --> 00:15:26,440
On the surface this sounds incredibly reassuring and for the most part it is.

376
00:15:26,440 --> 00:15:28,840
There is no new data exposure to worry about

377
00:15:28,840 --> 00:15:33,240
because Work IQ cannot let an agent see something a human user couldn't already open.

378
00:15:33,240 --> 00:15:36,200
If you lack the permissions to read a specific file in SharePoint

379
00:15:36,200 --> 00:15:38,920
Work IQ simply won't show that file to the agent.

380
00:15:38,920 --> 00:15:41,240
When you aren't a member of a specific team's channel

381
00:15:41,240 --> 00:15:43,640
the agent is unable to pull conversations from it.

382
00:15:43,640 --> 00:15:46,040
If you don't have access to a specific thread in Outlook

383
00:15:46,040 --> 00:15:49,240
Work IQ won't include that data when it tries to synthesize context.

384
00:15:49,240 --> 00:15:51,000
The permission boundary is absolute

385
00:15:51,000 --> 00:15:53,480
but this is exactly where governance starts to get complicated.

386
00:15:53,480 --> 00:15:55,960
There is a massive gap between what a person can access

387
00:15:55,960 --> 00:15:57,800
and what they actually should access.

388
00:15:57,800 --> 00:16:00,840
You have to look at the difference between what the technical system allows

389
00:16:00,840 --> 00:16:04,200
and what the organization actually intends for its employees to see.

390
00:16:04,200 --> 00:16:08,440
In almost every organization access permissions tend to get broader as time goes on.

391
00:16:08,440 --> 00:16:11,960
A sensitive file gets shared with everyone in finance

392
00:16:11,960 --> 00:16:13,640
just because it was convenient at the moment.

393
00:16:13,640 --> 00:16:16,040
A SharePoint site ends up with wide open access

394
00:16:16,040 --> 00:16:19,880
because it was easier to share the whole folder than to manage individual permissions.

395
00:16:19,880 --> 00:16:22,840
Teams channels often include people from five different departments

396
00:16:22,840 --> 00:16:24,440
because they collaborated once

397
00:16:24,440 --> 00:16:27,240
and removing them later feels socially awkward.

398
00:16:27,240 --> 00:16:30,040
The permission model usually reflects historical convenience

399
00:16:30,040 --> 00:16:32,040
rather than current business intent.

400
00:16:32,040 --> 00:16:34,600
Work IQ is going to expose this gap immediately.

401
00:16:34,600 --> 00:16:37,400
Instead of a human manually hunting for sensitive documents

402
00:16:37,400 --> 00:16:39,400
which is a slow and incomplete process

403
00:16:39,400 --> 00:16:43,400
an AI agent can reason across everything a user is technically allowed to touch

404
00:16:43,400 --> 00:16:46,360
an agent can instantly tell a user that they have access

405
00:16:46,360 --> 00:16:48,840
to strategic planning documents, budget forecasts

406
00:16:48,840 --> 00:16:51,560
and executive communications about an upcoming acquisition.

407
00:16:51,560 --> 00:16:54,680
The agent isn't actually breaking any rules or violating permissions.

408
00:16:54,680 --> 00:16:56,840
It is simply making the invisible visible

409
00:16:56,840 --> 00:16:59,400
by showing what the permission model already allows.

410
00:16:59,400 --> 00:17:01,800
In most companies the permission model allows for things

411
00:17:01,800 --> 00:17:03,560
that nobody ever intended to make public.

412
00:17:03,560 --> 00:17:06,200
This reality makes sensitivity labels and DLP policies

413
00:17:06,200 --> 00:17:07,800
your most critical infrastructure.

414
00:17:07,800 --> 00:17:09,880
Work IQ respects these labels in real time

415
00:17:09,880 --> 00:17:12,040
which means a document marked highly confidential

416
00:17:12,040 --> 00:17:14,440
does more than just restrict external sharing.

417
00:17:14,440 --> 00:17:18,200
It actually restricts what the AI agent is allowed to do with that information.

418
00:17:18,200 --> 00:17:20,040
The agent cannot summarize that document

419
00:17:20,040 --> 00:17:21,800
it cannot include it in morning briefings

420
00:17:21,800 --> 00:17:24,040
and it cannot synthesize it with other files.

421
00:17:24,040 --> 00:17:26,200
The label acts as a digital gate for the AI

422
00:17:26,200 --> 00:17:27,960
just as much as it does for the human.

423
00:17:27,960 --> 00:17:31,160
This shift moves governance from a passive state to an active one.

424
00:17:31,160 --> 00:17:34,280
Traditionally, sensitivity labels were used for prevention

425
00:17:34,280 --> 00:17:36,760
like stopping a document from being shared externally

426
00:17:36,760 --> 00:17:38,040
or blocked from being copied.

427
00:17:38,040 --> 00:17:40,200
The system would wait for a human to take an action

428
00:17:40,200 --> 00:17:42,360
and then reactively block it if it broke a rule.

429
00:17:42,360 --> 00:17:45,560
Work IQ turns that model into a system of continuous enforcement.

430
00:17:45,560 --> 00:17:47,640
The label doesn't just block a final action

431
00:17:47,640 --> 00:17:50,760
but it actively controls what the AI is allowed to think about.

432
00:17:50,760 --> 00:17:52,760
It dictates what the agent see

433
00:17:52,760 --> 00:17:55,160
and what gets included in a final summary.

434
00:17:55,160 --> 00:17:57,320
Because every request runs through a policy engine

435
00:17:57,320 --> 00:17:58,920
that understands context,

436
00:17:58,920 --> 00:18:01,720
the system doesn't just ask if a user has access.

437
00:18:01,720 --> 00:18:03,400
It asks if the specific request

438
00:18:03,400 --> 00:18:05,000
from that specific identity

439
00:18:05,000 --> 00:18:07,560
makes sense within the current context and company policy.

440
00:18:07,560 --> 00:18:10,520
This is real-time evaluation happening during every single query

441
00:18:10,520 --> 00:18:11,560
and every single inference.

442
00:18:11,560 --> 00:18:13,720
This level of granularity is necessary

443
00:18:13,720 --> 00:18:15,960
but it is also brand new for most teams.

444
00:18:15,960 --> 00:18:18,200
Most organizations do not have governance frameworks

445
00:18:18,200 --> 00:18:20,600
that are sophisticated enough to work at this speed.

446
00:18:20,600 --> 00:18:22,920
They rely on broad, static policies

447
00:18:22,920 --> 00:18:24,920
where a specific role can do one thing

448
00:18:24,920 --> 00:18:26,600
and a security group can do another

449
00:18:26,600 --> 00:18:29,640
that works fine when humans are clicking through folders manually.

450
00:18:29,640 --> 00:18:31,960
It fails completely when agents are making

451
00:18:31,960 --> 00:18:34,120
thousands of micro decisions every second.

452
00:18:34,120 --> 00:18:36,120
If your policy framework isn't precise

453
00:18:36,120 --> 00:18:39,480
those micro decisions will eventually expose sensitive information.

454
00:18:39,480 --> 00:18:41,480
The permission model is secured by design

455
00:18:41,480 --> 00:18:44,600
but that security is only as good as the labels you've applied.

456
00:18:44,600 --> 00:18:47,720
Your safety depends entirely on the quality of your policy design

457
00:18:47,720 --> 00:18:50,040
and how well you maintain those rules over time.

458
00:18:50,040 --> 00:18:50,920
If you get this right,

459
00:18:50,920 --> 00:18:54,040
work IQ actually becomes more secure than your current setup

460
00:18:54,040 --> 00:18:56,760
because it enforces policy with perfect consistency.

461
00:18:56,760 --> 00:18:57,880
If you get it wrong,

462
00:18:57,880 --> 00:19:00,280
you've created the exact problem you were trying to avoid

463
00:19:00,280 --> 00:19:03,080
by giving intelligent agents far too much room to roam.

464
00:19:03,080 --> 00:19:04,600
The data hygiene crisis.

465
00:19:04,600 --> 00:19:06,680
We have to talk about an uncomfortable reality

466
00:19:06,680 --> 00:19:09,160
that most leadership teams are currently ignoring.

467
00:19:09,160 --> 00:19:11,320
If you opened up your SharePoint environment right now

468
00:19:11,320 --> 00:19:12,760
to perform a real audit,

469
00:19:12,760 --> 00:19:14,120
you would probably hate what you found.

470
00:19:14,120 --> 00:19:16,280
You likely have sites that haven't been touched in three years

471
00:19:16,280 --> 00:19:17,800
and teams groups created for projects

472
00:19:17,800 --> 00:19:19,800
that ended back in 2021.

473
00:19:19,800 --> 00:19:21,960
Files are sitting in personal one-drive folders

474
00:19:21,960 --> 00:19:23,720
that should be in official repositories

475
00:19:23,720 --> 00:19:26,840
but they stayed local because it was easier for the employee.

476
00:19:26,840 --> 00:19:29,240
Your naming conventions have probably changed several times,

477
00:19:29,240 --> 00:19:31,560
leaving your content inconsistent and confusing.

478
00:19:31,560 --> 00:19:34,440
This isn't just a compliance headache in the old fashion sense.

479
00:19:34,440 --> 00:19:36,600
It is a fundamental operational reality

480
00:19:36,600 --> 00:19:38,440
that every single organization deals with.

481
00:19:38,440 --> 00:19:40,440
Most companies just learn to live with the mess.

482
00:19:40,440 --> 00:19:42,440
The general attitude is that things are cluttered

483
00:19:42,440 --> 00:19:44,440
but we manage to get our work done anyway.

484
00:19:44,440 --> 00:19:46,600
Audits might find a few issues that get fixed

485
00:19:46,600 --> 00:19:48,840
but the rest of the data just continues to drift.

486
00:19:48,840 --> 00:19:50,120
It isn't the perfect system,

487
00:19:50,120 --> 00:19:52,600
but it works because people know how to navigate the chaos.

488
00:19:52,600 --> 00:19:55,000
They know where the important files actually live,

489
00:19:55,000 --> 00:19:57,720
even if the folder structure says they should be somewhere else.

490
00:19:57,720 --> 00:19:59,880
Work IQ changes that math entirely

491
00:19:59,880 --> 00:20:02,840
and AI doesn't navigate a mess the same way a human does.

492
00:20:02,840 --> 00:20:05,160
It doesn't use common sense or apply human judgment

493
00:20:05,160 --> 00:20:06,680
to skip over old folders.

494
00:20:06,680 --> 00:20:08,440
It won't say that a file technically exists

495
00:20:08,440 --> 00:20:09,960
but nobody would ever think to look for it.

496
00:20:09,960 --> 00:20:13,000
Instead the AI sees that the file is discoverable

497
00:20:13,000 --> 00:20:16,280
and accessible so it pulls it into the conversation.

498
00:20:16,280 --> 00:20:19,400
An agent powered by Work IQ can instantly catalog

499
00:20:19,400 --> 00:20:22,920
and surface every sensitive document a user has technical access to.

500
00:20:22,920 --> 00:20:26,360
It doesn't care what the organization intended for that user to see.

501
00:20:26,360 --> 00:20:29,080
It only cares what the broken permission model actually allows.

502
00:20:29,080 --> 00:20:30,120
This isn't a helpful feature

503
00:20:30,120 --> 00:20:32,280
but rather a massive liability disguised

504
00:20:32,280 --> 00:20:33,560
as a productivity tool.

505
00:20:33,560 --> 00:20:35,400
Think about the actual consequences of this.

506
00:20:35,400 --> 00:20:37,560
You might have financial planning documents

507
00:20:37,560 --> 00:20:40,120
that were shared with a finance group five years ago

508
00:20:40,120 --> 00:20:42,600
and nobody has checked that group's membership since.

509
00:20:42,600 --> 00:20:44,680
Work IQ will find them and surface them.

510
00:20:44,680 --> 00:20:48,280
You might have executive emails about sensitive strategic decisions

511
00:20:48,280 --> 00:20:50,680
that were forwarded to the wrong people years ago.

512
00:20:50,680 --> 00:20:52,040
Work IQ will find those too.

513
00:20:52,040 --> 00:20:54,600
Confidential HR documents might be sitting in a sharepoint site

514
00:20:54,600 --> 00:20:57,160
from a hiring decision that ended in 2019

515
00:20:57,160 --> 00:20:58,920
but because the share was never removed,

516
00:20:58,920 --> 00:21:00,600
Work IQ will connect the dots.

517
00:21:00,600 --> 00:21:02,520
This is the moment where organizations realize

518
00:21:02,520 --> 00:21:04,360
that governance cannot be an afterthought.

519
00:21:04,360 --> 00:21:06,200
You cannot deploy Work IQ first

520
00:21:06,200 --> 00:21:07,960
and then try to clean up the mess later.

521
00:21:07,960 --> 00:21:10,280
You have to fix the foundation before the intelligence layer

522
00:21:10,280 --> 00:21:13,880
turns your digital clutter into a full-blown operational emergency.

523
00:21:13,880 --> 00:21:17,080
Microsoft calls this the copilot-ready data problem

524
00:21:17,080 --> 00:21:18,520
and they are right to focus on it.

525
00:21:18,520 --> 00:21:20,760
But this isn't just about one specific tool.

526
00:21:20,760 --> 00:21:23,720
It is about the danger of deploying an intelligent agent

527
00:21:23,720 --> 00:21:25,640
that sees everything without the human context

528
00:21:25,640 --> 00:21:27,160
that makes a mess tolerable.

529
00:21:27,160 --> 00:21:29,560
Once an AI starts reasoning over your data,

530
00:21:29,560 --> 00:21:32,520
the dark matter in your files becomes an urgent priority.

531
00:21:32,520 --> 00:21:34,520
Stale sites become liabilities.

532
00:21:34,520 --> 00:21:36,440
Often groups become security risks

533
00:21:36,440 --> 00:21:38,920
and broken permissions become impossible to ignore.

534
00:21:38,920 --> 00:21:42,280
Cleaning up your data is the absolute prerequisite for this technology.

535
00:21:42,280 --> 00:21:43,720
It isn't a follow-up project

536
00:21:43,720 --> 00:21:46,200
or something you can handle during phase two of your rollout.

537
00:21:46,200 --> 00:21:47,880
It is the gate you have to pass through.

538
00:21:47,880 --> 00:21:49,400
Before you let Work IQ lose,

539
00:21:49,400 --> 00:21:51,560
you have to get your sharepoint environment clean

540
00:21:51,560 --> 00:21:54,360
by archiving dead sites and fixing permission inheritance.

541
00:21:54,360 --> 00:21:56,120
You need to rationalize your team structure

542
00:21:56,120 --> 00:21:58,760
so you actually know why each group exists and who is in it.

543
00:21:58,760 --> 00:22:00,120
You also need naming conventions

544
00:22:00,120 --> 00:22:02,120
that actually means something to a machine.

545
00:22:02,120 --> 00:22:04,840
Your file structure should tell the system what the content is

546
00:22:04,840 --> 00:22:06,840
without requiring the AI to guess.

547
00:22:06,840 --> 00:22:08,840
You have to fix your sensitivity labels

548
00:22:08,840 --> 00:22:10,840
so they reflect actual business risk

549
00:22:10,840 --> 00:22:12,520
rather than theoretical categories.

550
00:22:12,520 --> 00:22:14,680
Finally, you need to look at your distribution lists

551
00:22:14,680 --> 00:22:16,920
to see who is actually receiving information

552
00:22:16,920 --> 00:22:18,920
and archive the ones that are no longer in use.

553
00:22:18,920 --> 00:22:20,120
None of this is exciting work.

554
00:22:20,120 --> 00:22:22,520
It isn't a new capability you can brag about in a meeting.

555
00:22:22,520 --> 00:22:24,040
It is basic digital housekeeping,

556
00:22:24,040 --> 00:22:25,880
but it is also essential.

557
00:22:25,880 --> 00:22:28,360
The second Work IQ starts looking at your data.

558
00:22:28,360 --> 00:22:30,120
Every structural weakness you've tolerated

559
00:22:30,120 --> 00:22:31,800
for years will become visible.

560
00:22:31,800 --> 00:22:34,360
Every shortcut your team took in 2018

561
00:22:34,360 --> 00:22:37,240
becomes a piece of operational debt that is now due.

562
00:22:37,240 --> 00:22:39,240
Every time you said, "We'll fix it later."

563
00:22:39,240 --> 00:22:41,640
You were creating a future governance crisis.

564
00:22:41,640 --> 00:22:43,560
You have to treat data clean up as a requirement,

565
00:22:43,560 --> 00:22:44,360
not an option.

566
00:22:44,360 --> 00:22:46,840
You don't do this because a software vendor told you to.

567
00:22:46,840 --> 00:22:49,720
You do it because Work IQ will no longer allow you to pretend

568
00:22:49,720 --> 00:22:51,000
that the problem doesn't exist.

569
00:22:51,000 --> 00:22:53,480
From storage to reasoning,

570
00:22:53,480 --> 00:22:55,160
this is the actual pivot point.

571
00:22:55,160 --> 00:22:56,920
We aren't talking about a feature update

572
00:22:56,920 --> 00:22:58,200
or some incremental improvement,

573
00:22:58,200 --> 00:22:59,800
but a fundamental restructuring

574
00:22:59,800 --> 00:23:01,560
of what the system is actually for.

575
00:23:01,560 --> 00:23:02,760
The old model is simple.

576
00:23:02,760 --> 00:23:03,800
You store data,

577
00:23:03,800 --> 00:23:05,800
and then users come to the system to get it back.

578
00:23:05,800 --> 00:23:07,080
They usually know what they need,

579
00:23:07,080 --> 00:23:08,440
or at least they think they do,

580
00:23:08,440 --> 00:23:10,200
and they go through the motions of retrieval.

581
00:23:10,200 --> 00:23:13,000
You open an email client and search for a specific message.

582
00:23:13,000 --> 00:23:15,080
You go to SharePoint, navigate through folders,

583
00:23:15,080 --> 00:23:16,360
and download a file.

584
00:23:16,360 --> 00:23:18,200
You check your calendar to find a meeting

585
00:23:18,200 --> 00:23:19,400
and see the details.

586
00:23:19,400 --> 00:23:21,880
In this world, the system handles storage and retrieval

587
00:23:21,880 --> 00:23:23,720
while the human handles the reasoning.

588
00:23:23,720 --> 00:23:25,720
You are the one who decides what to search for,

589
00:23:25,720 --> 00:23:27,000
what to do with the results,

590
00:23:27,000 --> 00:23:29,000
and what the information actually means.

591
00:23:29,000 --> 00:23:31,480
You connect the dots because the system is neutral.

592
00:23:31,480 --> 00:23:32,920
It holds everything equally,

593
00:23:32,920 --> 00:23:34,040
never judging what matters,

594
00:23:34,040 --> 00:23:36,200
or remembering what you cared about yesterday.

595
00:23:36,200 --> 00:23:39,160
And it certainly doesn't predict what you might need tomorrow.

596
00:23:39,160 --> 00:23:40,280
This model works fine

597
00:23:40,280 --> 00:23:42,040
if you are dealing with discrete artifacts

598
00:23:42,040 --> 00:23:44,280
like a single email or one document.

599
00:23:44,280 --> 00:23:47,640
But the reality is that knowledge work isn't discrete.

600
00:23:47,640 --> 00:23:48,920
It is continuous,

601
00:23:48,920 --> 00:23:49,880
relational,

602
00:23:49,880 --> 00:23:50,600
and layered.

603
00:23:50,600 --> 00:23:52,360
A project never lives in just one place.

604
00:23:52,360 --> 00:23:53,800
It lives across email threads,

605
00:23:53,800 --> 00:23:56,200
Teams conversations, SharePoint documents,

606
00:23:56,200 --> 00:23:56,920
meeting notes,

607
00:23:56,920 --> 00:23:58,120
and random chats with people

608
00:23:58,120 --> 00:23:59,960
across three different departments.

609
00:23:59,960 --> 00:24:01,240
To actually understand a project,

610
00:24:01,240 --> 00:24:02,760
you have to manually pull signals

611
00:24:02,760 --> 00:24:04,200
from every one of those locations

612
00:24:04,200 --> 00:24:05,400
and connect them yourself.

613
00:24:05,400 --> 00:24:06,600
You have to remember

614
00:24:06,600 --> 00:24:07,800
what you learned from each source

615
00:24:07,800 --> 00:24:09,720
and synthesize it into something coherent,

616
00:24:09,720 --> 00:24:11,080
because the system won't help you.

617
00:24:11,080 --> 00:24:12,520
It just gives you the raw material

618
00:24:12,520 --> 00:24:14,200
and leaves the hard work to you.

619
00:24:14,200 --> 00:24:15,960
Work IQ inverts this entire process.

620
00:24:15,960 --> 00:24:18,600
The new model is built to continuously reason over data,

621
00:24:18,600 --> 00:24:19,720
surface patterns,

622
00:24:19,720 --> 00:24:21,400
and anticipate what you need next.

623
00:24:21,400 --> 00:24:23,480
The system doesn't wait for you to ask a question

624
00:24:23,480 --> 00:24:25,400
because it is constantly analyzing,

625
00:24:25,400 --> 00:24:27,160
learning, and building associations.

626
00:24:27,160 --> 00:24:29,080
It updates its own model of what matters

627
00:24:29,080 --> 00:24:30,680
based on how you actually work

628
00:24:30,680 --> 00:24:32,440
rather than how you say you work.

629
00:24:32,440 --> 00:24:34,120
It looks at the patterns in your actions,

630
00:24:34,120 --> 00:24:35,560
like who you actually talk to

631
00:24:35,560 --> 00:24:36,840
and what you spend your time on.

632
00:24:36,840 --> 00:24:38,440
It identifies which decisions matter

633
00:24:38,440 --> 00:24:40,040
because they create follow-up activity

634
00:24:40,040 --> 00:24:42,920
and figures out what information actually affects your outcomes.

635
00:24:42,920 --> 00:24:43,480
Because of this,

636
00:24:43,480 --> 00:24:46,360
the role of information architecture shifts completely.

637
00:24:46,360 --> 00:24:48,120
It is no longer about organization,

638
00:24:48,120 --> 00:24:48,920
folder structures,

639
00:24:48,920 --> 00:24:50,040
or naming conventions.

640
00:24:50,040 --> 00:24:51,400
Those things become secondary.

641
00:24:51,400 --> 00:24:52,920
Architecture is now about intelligence

642
00:24:52,920 --> 00:24:55,080
and how the system is going to reason about your content.

643
00:24:55,080 --> 00:24:57,400
You have to ask what metadata matters for synthesis

644
00:24:57,400 --> 00:24:58,680
instead of just filing

645
00:24:58,680 --> 00:25:01,240
and what relationships the system needs to understand.

646
00:25:01,240 --> 00:25:03,720
The goal is to help the system learn patterns

647
00:25:03,720 --> 00:25:04,840
and represent decisions

648
00:25:04,840 --> 00:25:07,000
so it can anticipate the next step in the process.

649
00:25:07,000 --> 00:25:09,160
This changes everything for your operations.

650
00:25:09,160 --> 00:25:11,080
Work IQ doesn't just find documents,

651
00:25:11,080 --> 00:25:12,840
it understands why they matter

652
00:25:12,840 --> 00:25:15,000
in the context of what you are doing right now.

653
00:25:15,000 --> 00:25:16,600
It looks at the problems you are solving,

654
00:25:16,600 --> 00:25:17,560
the people involved,

655
00:25:17,560 --> 00:25:19,400
and the decisions that are still pending.

656
00:25:19,400 --> 00:25:22,600
It doesn't just say here is a document labeled "project status".

657
00:25:22,600 --> 00:25:25,240
Instead, it tells you that the document is a status update

658
00:25:25,240 --> 00:25:27,320
from three weeks ago regarding Project Gamma.

659
00:25:27,320 --> 00:25:29,480
It notes that two blocking decisions were resolved

660
00:25:29,480 --> 00:25:30,920
when finance signed the budget

661
00:25:30,920 --> 00:25:33,560
and legal cleared the contract but one is still open.

662
00:25:33,560 --> 00:25:35,240
It identifies that the head of product

663
00:25:35,240 --> 00:25:38,200
and the head of sales disagree on the go-to-market timeline.

664
00:25:38,200 --> 00:25:39,560
Even if you aren't directly involved,

665
00:25:39,560 --> 00:25:41,560
the system knows your team is implementing features

666
00:25:41,560 --> 00:25:42,520
that depend on this

667
00:25:42,520 --> 00:25:44,920
and you will be blocked in two weeks if it isn't fixed.

668
00:25:44,920 --> 00:25:47,320
That isn't retrieval, that is true understanding.

669
00:25:47,320 --> 00:25:48,760
The system tracks relationships

670
00:25:48,760 --> 00:25:51,160
at a level human simply cannot maintain manually.

671
00:25:51,160 --> 00:25:52,680
It sees who works with whom,

672
00:25:52,680 --> 00:25:53,720
how decisions flow,

673
00:25:53,720 --> 00:25:56,200
and which people are central to specific outcomes.

674
00:25:56,200 --> 00:25:58,520
It identifies which teams have friction

675
00:25:58,520 --> 00:26:00,760
and which cross-functional relationships are breaking down.

676
00:26:00,760 --> 00:26:03,720
The system doesn't build this view from surveys or org charts

677
00:26:03,720 --> 00:26:06,600
but from observable patterns in how work actually happens.

678
00:26:06,600 --> 00:26:08,040
It sees where information flows

679
00:26:08,040 --> 00:26:09,240
and where it gets stuck,

680
00:26:09,240 --> 00:26:11,080
distinguishing between real collaboration

681
00:26:11,080 --> 00:26:12,520
and ceremonial meetings.

682
00:26:12,520 --> 00:26:14,120
Since the system is reasoning continuously,

683
00:26:14,120 --> 00:26:16,440
it can answer the questions humans never bother to ask

684
00:26:16,440 --> 00:26:18,120
because the manual work would take hours.

685
00:26:18,120 --> 00:26:19,560
It can tell you what should happen next

686
00:26:19,560 --> 00:26:21,320
based on patterns, history,

687
00:26:21,320 --> 00:26:23,480
and what similar situations led to in the past.

688
00:26:23,480 --> 00:26:24,920
It knows who needs to be involved

689
00:26:24,920 --> 00:26:27,560
and what is typically required before the next phase can start.

690
00:26:27,560 --> 00:26:29,240
This is not just a small change.

691
00:26:29,240 --> 00:26:32,040
It restructures how an entire organization operates.

692
00:26:32,040 --> 00:26:33,960
The inference risk.

693
00:26:33,960 --> 00:26:36,840
Work IQ doesn't just consolidate the things you explicitly store.

694
00:26:36,840 --> 00:26:37,560
It infers.

695
00:26:37,560 --> 00:26:40,280
That distinction matters more than most organizations realize.

696
00:26:40,280 --> 00:26:42,440
Infrance is fundamentally different from retrieval.

697
00:26:42,440 --> 00:26:43,480
When you search for a file,

698
00:26:43,480 --> 00:26:46,040
that file already exists and you are just finding it.

699
00:26:46,040 --> 00:26:47,400
But inference creates knowledge

700
00:26:47,400 --> 00:26:49,000
that was never actually documented.

701
00:26:49,000 --> 00:26:51,640
Work IQ analyzes patterns to generate conclusions

702
00:26:51,640 --> 00:26:53,000
about relationships and behaviors

703
00:26:53,000 --> 00:26:54,680
that nobody ever formally encoded.

704
00:26:54,680 --> 00:26:56,760
It observes how people interact with information,

705
00:26:56,760 --> 00:26:58,040
watches communication flows,

706
00:26:58,040 --> 00:26:59,640
and tracks decision sequences.

707
00:26:59,640 --> 00:27:01,640
From those observations, it infers structure.

708
00:27:01,640 --> 00:27:03,000
It finds influence networks,

709
00:27:03,000 --> 00:27:05,240
project dependencies, and decision bottlenecks.

710
00:27:05,240 --> 00:27:07,640
It figures out which people are central to outcomes

711
00:27:07,640 --> 00:27:09,320
and which teams hold authority

712
00:27:09,320 --> 00:27:11,320
that the official org chart doesn't show.

713
00:27:11,320 --> 00:27:13,000
This capability is incredibly valuable

714
00:27:13,000 --> 00:27:15,160
but it is also dangerous if it's misused.

715
00:27:15,160 --> 00:27:17,080
Think about what an inference can become.

716
00:27:17,080 --> 00:27:19,400
Work IQ might observe that you are consistently involved

717
00:27:19,400 --> 00:27:20,440
in high stakes decisions

718
00:27:20,440 --> 00:27:21,640
and that people reach out to you

719
00:27:21,640 --> 00:27:23,160
whenever things are uncertain.

720
00:27:23,160 --> 00:27:24,840
It sees that projects don't move forward

721
00:27:24,840 --> 00:27:26,200
until you give your input.

722
00:27:26,200 --> 00:27:28,680
That is an objective observation about your influence.

723
00:27:28,680 --> 00:27:30,440
And the system isn't making a judgment.

724
00:27:30,440 --> 00:27:32,040
It's just documenting a pattern.

725
00:27:32,040 --> 00:27:33,880
But the moment that inference becomes visible

726
00:27:33,880 --> 00:27:35,480
to management things change.

727
00:27:35,480 --> 00:27:36,680
When someone points to the data

728
00:27:36,680 --> 00:27:38,200
and says you are a bottleneck,

729
00:27:38,200 --> 00:27:40,920
that inference transforms into a performance metric.

730
00:27:40,920 --> 00:27:42,520
What started as a descriptive observation

731
00:27:42,520 --> 00:27:44,120
becomes prescriptive pressure.

732
00:27:44,120 --> 00:27:46,200
Now, you are expected to move faster.

733
00:27:46,200 --> 00:27:48,760
Stay out of certain decisions and delegate more.

734
00:27:48,760 --> 00:27:50,680
The inference that described your actual role

735
00:27:50,680 --> 00:27:53,560
is suddenly weaponized as a measurement of your effectiveness.

736
00:27:53,560 --> 00:27:55,480
We see the same thing with behavioral patterns.

737
00:27:55,480 --> 00:27:57,720
Work IQ observes exactly when you work,

738
00:27:57,720 --> 00:27:59,560
noting that you spend all day in meetings

739
00:27:59,560 --> 00:28:01,480
but respond to emails at 10 pm.

740
00:28:01,480 --> 00:28:03,240
It sees which conversations you join

741
00:28:03,240 --> 00:28:05,320
and which ones you just watch from the sidelines.

742
00:28:05,320 --> 00:28:06,840
It tracks how often you switch tools

743
00:28:06,840 --> 00:28:08,680
and how fast you respond to messages.

744
00:28:08,680 --> 00:28:10,200
All of this is infeurable from data

745
00:28:10,200 --> 00:28:11,800
that already exists in your tenant

746
00:28:11,800 --> 00:28:13,800
like email timestamps and calendar activity.

747
00:28:13,800 --> 00:28:16,120
The system isn't intentionally surveilling you

748
00:28:16,120 --> 00:28:18,680
but the inferences that emerge can feel like judgment.

749
00:28:18,680 --> 00:28:20,760
It might conclude that a person over works,

750
00:28:20,760 --> 00:28:22,200
responds slowly to feedback

751
00:28:22,200 --> 00:28:24,040
or avoids difficult conversations.

752
00:28:24,040 --> 00:28:26,600
Even if the system is being entirely objective,

753
00:28:26,600 --> 00:28:28,280
those conclusions feel like surveillance.

754
00:28:28,280 --> 00:28:30,280
Here is the problem.

755
00:28:30,280 --> 00:28:32,600
Infrance creates knowledge that didn't exist before.

756
00:28:32,600 --> 00:28:35,320
You never explicitly said you were a bottleneck

757
00:28:35,320 --> 00:28:36,760
or that you were overworking.

758
00:28:36,760 --> 00:28:38,120
The system inferred those truths

759
00:28:38,120 --> 00:28:39,800
by analyzing your behavior.

760
00:28:39,800 --> 00:28:41,880
Because this knowledge is derived from data

761
00:28:41,880 --> 00:28:43,400
rather than your own declarations,

762
00:28:43,400 --> 00:28:44,920
it is often more threatening.

763
00:28:44,920 --> 00:28:46,040
You can't really argue with it

764
00:28:46,040 --> 00:28:47,960
because you didn't create it or publish it.

765
00:28:47,960 --> 00:28:49,960
It just emerged from the systems analysis

766
00:28:49,960 --> 00:28:51,960
of how you actually behave every day.

767
00:28:51,960 --> 00:28:53,880
Organizations have to define boundaries

768
00:28:53,880 --> 00:28:55,880
on what inferences are actually permissible.

769
00:28:55,880 --> 00:28:57,720
This isn't about what is technically possible

770
00:28:57,720 --> 00:29:00,760
but what is ethically and operationally appropriate for the culture.

771
00:29:00,760 --> 00:29:02,520
We have to ask if the system should be allowed

772
00:29:02,520 --> 00:29:04,120
to infer influence networks.

773
00:29:04,120 --> 00:29:05,320
The answer is probably yes

774
00:29:05,320 --> 00:29:06,520
because that helps us understand

775
00:29:06,520 --> 00:29:08,040
how work gets coordinated.

776
00:29:08,040 --> 00:29:09,640
But should those inferences be visible

777
00:29:09,640 --> 00:29:11,480
to managers for performance reviews?

778
00:29:11,480 --> 00:29:14,360
Probably not because that turns intelligence into a weapon.

779
00:29:14,360 --> 00:29:15,880
The system can technically infer

780
00:29:15,880 --> 00:29:17,400
that you work outside normal hours

781
00:29:17,400 --> 00:29:19,560
but whether it flags that as a burn out risk

782
00:29:19,560 --> 00:29:20,760
or a sign of commitment

783
00:29:20,760 --> 00:29:23,400
depends entirely on the organization's values.

784
00:29:23,400 --> 00:29:25,720
Our privacy frameworks need to evolve much faster

785
00:29:25,720 --> 00:29:27,240
than the technology itself.

786
00:29:27,240 --> 00:29:29,720
Right now most organizations don't have any governance structures

787
00:29:29,720 --> 00:29:31,000
that account for inference.

788
00:29:31,000 --> 00:29:33,160
They have plenty of policies about data collection,

789
00:29:33,160 --> 00:29:34,360
access and retention,

790
00:29:34,360 --> 00:29:36,600
but they are silent on what kinds of patterns

791
00:29:36,600 --> 00:29:37,960
the system is allowed to derive.

792
00:29:37,960 --> 00:29:39,400
They haven't decided what counts

793
00:29:39,400 --> 00:29:40,760
as organizational intelligence

794
00:29:40,760 --> 00:29:42,760
and what counts as individual surveillance.

795
00:29:42,760 --> 00:29:43,960
Without those boundaries,

796
00:29:43,960 --> 00:29:46,600
the power of inference becomes a massive liability

797
00:29:46,600 --> 00:29:47,640
instead of an advantage.

798
00:29:47,640 --> 00:29:49,560
The memory problem.

799
00:29:49,560 --> 00:29:52,280
Work IQ doesn't just look for patterns in the moment.

800
00:29:52,280 --> 00:29:53,400
It remembers them.

801
00:29:53,400 --> 00:29:55,960
That is the critical distinction we need to understand.

802
00:29:55,960 --> 00:29:57,880
While inference describes what the system learns

803
00:29:57,880 --> 00:29:59,720
by looking at what you're doing right now,

804
00:29:59,720 --> 00:30:01,720
memory is what the system actually keeps.

805
00:30:01,720 --> 00:30:03,160
It's what the system carries forward

806
00:30:03,160 --> 00:30:04,760
to make decisions about the next hour,

807
00:30:04,760 --> 00:30:05,960
the next day and the next month.

808
00:30:05,960 --> 00:30:08,280
This memory capability is incredibly valuable

809
00:30:08,280 --> 00:30:09,800
because co-pilot shouldn't have to start

810
00:30:09,800 --> 00:30:12,280
from zero every single time you ask a question.

811
00:30:12,280 --> 00:30:14,680
It should know what you were working on last Tuesday

812
00:30:14,680 --> 00:30:16,520
and it should remember the specific decisions

813
00:30:16,520 --> 00:30:18,120
your team made in that meeting.

814
00:30:18,120 --> 00:30:19,320
The system needs to understand

815
00:30:19,320 --> 00:30:21,320
that the problem you're wrestling with today

816
00:30:21,320 --> 00:30:22,840
is actually linked to a conversation

817
00:30:22,840 --> 00:30:24,040
you had three weeks ago

818
00:30:24,040 --> 00:30:26,040
with someone in a different department.

819
00:30:26,040 --> 00:30:28,440
By building a persistent model of how you work,

820
00:30:28,440 --> 00:30:31,000
the system learns who your typical collaborators are

821
00:30:31,000 --> 00:30:32,840
and which projects you actually own.

822
00:30:32,840 --> 00:30:34,840
It tracks the information you usually need

823
00:30:34,840 --> 00:30:37,800
when you need it and exactly how you prefer it to be presented.

824
00:30:37,800 --> 00:30:40,680
That behavioral memory is what makes co-pilot feel like a partner

825
00:30:40,680 --> 00:30:42,600
instead of just a faster search engine.

826
00:30:42,600 --> 00:30:45,320
But here's the problem, memory concentrates risk.

827
00:30:45,320 --> 00:30:48,120
What the system remembers is ultimately a governance question

828
00:30:48,120 --> 00:30:49,960
that most companies haven't answered yet.

829
00:30:49,960 --> 00:30:52,040
We have to ask how long this data stays,

830
00:30:52,040 --> 00:30:54,120
who can see it and whether you can delete

831
00:30:54,120 --> 00:30:56,040
or correct it if the system gets it wrong.

832
00:30:56,040 --> 00:30:57,320
You might want to opt out of memory

833
00:30:57,320 --> 00:30:59,000
for certain sensitive activities

834
00:30:59,000 --> 00:31:01,400
but right now the controls aren't always clear.

835
00:31:01,400 --> 00:31:04,280
Most organizations are currently flying blind on these questions.

836
00:31:04,280 --> 00:31:07,400
They have standard retention policies for email and documents

837
00:31:07,400 --> 00:31:09,240
which are usually pretty straightforward.

838
00:31:09,240 --> 00:31:11,400
Email lives for seven years and then it's gone

839
00:31:11,400 --> 00:31:13,560
while documents follow the standard share point rules.

840
00:31:13,560 --> 00:31:15,400
But memory about your behavioral patterns,

841
00:31:15,400 --> 00:31:17,480
your work preferences and your decision history

842
00:31:17,480 --> 00:31:18,840
is entirely new territory.

843
00:31:18,840 --> 00:31:21,560
There is no established playbook for how long that data should be kept

844
00:31:21,560 --> 00:31:24,200
or what happens to it when an employee leaves the company.

845
00:31:24,200 --> 00:31:26,200
We don't know if that memory should transfer

846
00:31:26,200 --> 00:31:27,640
when you change roles

847
00:31:27,640 --> 00:31:29,640
or if it can be used to influence

848
00:31:29,640 --> 00:31:31,320
what the system shows to other people.

849
00:31:31,320 --> 00:31:33,240
Think about how this looks in practice.

850
00:31:33,240 --> 00:31:34,920
The system remembers that you are the person

851
00:31:34,920 --> 00:31:37,880
who always reviews financial decisions for certain projects

852
00:31:37,880 --> 00:31:40,120
and it knows you prefer short executive summaries

853
00:31:40,120 --> 00:31:41,240
over long briefings.

854
00:31:41,240 --> 00:31:43,320
It sees that you typically pull in the same three people

855
00:31:43,320 --> 00:31:45,000
when a decision gets complicated

856
00:31:45,000 --> 00:31:46,600
and it recognizes that your projects

857
00:31:46,600 --> 00:31:48,600
tend to be high risk and fast moving.

858
00:31:48,600 --> 00:31:49,640
All of that is true

859
00:31:49,640 --> 00:31:51,160
and it helps co-pilot serve you better.

860
00:31:51,160 --> 00:31:53,240
But what happens when you want to break that pattern?

861
00:31:53,240 --> 00:31:55,320
If you're burned out and want to stop being the bottleneck

862
00:31:55,320 --> 00:31:56,120
for every decision,

863
00:31:56,120 --> 00:31:58,200
the system's memory might actually trap you.

864
00:31:58,200 --> 00:32:00,920
It still thinks you're the person who needs to be involved

865
00:32:00,920 --> 00:32:03,320
so it keeps pushing you into the same old loops.

866
00:32:03,320 --> 00:32:05,080
There is also the issue of what happens

867
00:32:05,080 --> 00:32:07,000
when someone leaves the organization.

868
00:32:07,000 --> 00:32:08,680
That person's communication style,

869
00:32:08,680 --> 00:32:10,760
their relationships and their role in the hierarchy

870
00:32:10,760 --> 00:32:13,000
all stay behind in the work IQ memory.

871
00:32:13,000 --> 00:32:14,760
If the company later needs to figure out

872
00:32:14,760 --> 00:32:16,760
why a specific decision was made,

873
00:32:16,760 --> 00:32:18,440
the system can provide deep insights

874
00:32:18,440 --> 00:32:20,760
into how that departed person used to operate.

875
00:32:20,760 --> 00:32:22,840
While that might be great for institutional learning,

876
00:32:22,840 --> 00:32:24,920
it also creates a permanent surveillance archive

877
00:32:24,920 --> 00:32:27,320
of someone who isn't there to defend their record.

878
00:32:27,320 --> 00:32:28,680
Users need real visibility

879
00:32:28,680 --> 00:32:31,000
into what is being remembered about them.

880
00:32:31,000 --> 00:32:33,240
It isn't enough to know that memory exists.

881
00:32:33,240 --> 00:32:35,240
People need to see the specific patterns

882
00:32:35,240 --> 00:32:36,680
the system has documented

883
00:32:36,680 --> 00:32:39,560
and the conclusions it has drawn about their work habits.

884
00:32:39,560 --> 00:32:40,760
Users also need control,

885
00:32:40,760 --> 00:32:43,160
including the ability to edit or delete memory

886
00:32:43,160 --> 00:32:44,360
if it becomes inaccurate

887
00:32:44,360 --> 00:32:45,880
or if their circumstances change.

888
00:32:45,880 --> 00:32:48,360
Admins need these same controls over retention and decay.

889
00:32:48,360 --> 00:32:50,520
If the system forgets everything too quickly,

890
00:32:50,520 --> 00:32:51,960
Copilot won't be able to connect you

891
00:32:51,960 --> 00:32:53,720
with your own historical expertise

892
00:32:53,720 --> 00:32:55,720
when a problem resurfaces a year later.

893
00:32:55,720 --> 00:32:57,400
But if memory lasts forever,

894
00:32:57,400 --> 00:32:59,880
the system will continue to misrepresent you long after

895
00:32:59,880 --> 00:33:01,720
you've moved on to new ways of working.

896
00:33:01,720 --> 00:33:03,800
Memory governance is a brand new discipline

897
00:33:03,800 --> 00:33:06,280
and until organizations build the right frameworks,

898
00:33:06,280 --> 00:33:08,280
work IQ cannot operate safely.

899
00:33:08,280 --> 00:33:11,000
Agent 365 and Gov. autonomy.

900
00:33:11,000 --> 00:33:14,120
This is the point where governance turns into organizational design.

901
00:33:14,120 --> 00:33:16,040
We have to stop thinking of agents as tools

902
00:33:16,040 --> 00:33:17,560
and start seeing them as actors.

903
00:33:17,560 --> 00:33:19,080
In the past, every tool you used

904
00:33:19,080 --> 00:33:21,240
in Microsoft 365 was passive.

905
00:33:21,240 --> 00:33:23,720
Outlook never sent an email unless you clicked send

906
00:33:23,720 --> 00:33:25,080
and SharePoint didn't just decide

907
00:33:25,080 --> 00:33:27,000
to move your files around on its own.

908
00:33:27,000 --> 00:33:29,000
You were the one who initiated the action,

909
00:33:29,000 --> 00:33:31,400
which meant you were the one responsible for the outcome.

910
00:33:31,400 --> 00:33:32,680
The tool was just an instrument

911
00:33:32,680 --> 00:33:34,120
and you held all the agency.

912
00:33:34,120 --> 00:33:36,840
Agents powered by Work IQ changed that dynamic

913
00:33:36,840 --> 00:33:39,320
because they can initiate, propose, and act

914
00:33:39,320 --> 00:33:41,480
without waiting for a human to tell them what to do.

915
00:33:41,480 --> 00:33:44,280
They operate autonomously within the boundaries you set

916
00:33:44,280 --> 00:33:46,360
and the moment an agent takes an independent step

917
00:33:46,360 --> 00:33:48,520
the question of accountability becomes urgent.

918
00:33:48,520 --> 00:33:50,280
If an agent makes a massive mistake,

919
00:33:50,280 --> 00:33:52,360
we need to know exactly who is responsible.

920
00:33:52,360 --> 00:33:54,600
Microsoft is addressing this through Agent 365.

921
00:33:54,600 --> 00:33:56,120
This isn't a new product you buy,

922
00:33:56,120 --> 00:33:57,960
but rather a governance framework

923
00:33:57,960 --> 00:34:00,600
that treats agents like real members of the organization.

924
00:34:00,600 --> 00:34:02,760
Agents get their own Entra ID accounts,

925
00:34:02,760 --> 00:34:03,960
their own licenses,

926
00:34:03,960 --> 00:34:05,720
and their own specific policies.

927
00:34:05,720 --> 00:34:08,520
You can assign them to roles, monitor their activity,

928
00:34:08,520 --> 00:34:10,280
and audit their decisions just like you would

929
00:34:10,280 --> 00:34:11,480
with a human employee.

930
00:34:11,480 --> 00:34:13,960
They aren't shadow ET or invisible scripts,

931
00:34:13,960 --> 00:34:16,920
they are managed and visible actors in your infrastructure.

932
00:34:16,920 --> 00:34:19,080
This approach is architecturally solid,

933
00:34:19,080 --> 00:34:20,840
but it creates a level of complexity

934
00:34:20,840 --> 00:34:22,760
that most companies aren't ready to handle.

935
00:34:22,760 --> 00:34:25,480
An agent working inside Work IQ can act on your behalf

936
00:34:25,480 --> 00:34:27,720
by sending emails, updating CRM records,

937
00:34:27,720 --> 00:34:28,920
or posting in Teams.

938
00:34:28,920 --> 00:34:30,600
Because the agent inherits the permissions

939
00:34:30,600 --> 00:34:31,880
of the user it represents,

940
00:34:31,880 --> 00:34:34,040
it can't do anything you aren't allowed to do.

941
00:34:34,040 --> 00:34:35,240
That keeps things secure,

942
00:34:35,240 --> 00:34:37,240
but it isn't enough for true governance.

943
00:34:37,240 --> 00:34:39,720
You don't just need to know what an agent is allowed to do,

944
00:34:39,720 --> 00:34:40,920
you need to know what it should do.

945
00:34:40,920 --> 00:34:43,800
You have to decide which actions require a human to sign off

946
00:34:43,800 --> 00:34:45,880
and which decisions the agent can make on its own.

947
00:34:45,880 --> 00:34:47,480
Let's look at a real world scenario.

948
00:34:47,480 --> 00:34:50,520
Imagine an agent is scanning your emails and documents

949
00:34:50,520 --> 00:34:52,920
and finds a contract that has been stuck in a review queue

950
00:34:52,920 --> 00:34:53,640
for six weeks.

951
00:34:53,640 --> 00:34:55,000
This is a major customer,

952
00:34:55,000 --> 00:34:57,240
and the delay is starting to cost the company money.

953
00:34:57,240 --> 00:34:58,680
The agent understands the context

954
00:34:58,680 --> 00:35:01,320
and realizes the delay was caused by a legal concern

955
00:35:01,320 --> 00:35:03,560
that was actually resolved in an old email thread.

956
00:35:03,560 --> 00:35:05,000
The agent could send a clarification

957
00:35:05,000 --> 00:35:07,640
to the legal team right now and unblock the entire deal.

958
00:35:07,640 --> 00:35:09,640
Should the agent send that email automatically

959
00:35:09,640 --> 00:35:11,320
or should it wait for you to hit a button?

960
00:35:11,320 --> 00:35:13,720
Technically, there is no reason it couldn't just do it.

961
00:35:13,720 --> 00:35:15,160
You have permission to talk to legal

962
00:35:15,160 --> 00:35:16,680
and the agent is acting for you.

963
00:35:16,680 --> 00:35:18,280
From a security perspective, it's fine,

964
00:35:18,280 --> 00:35:20,360
but from a governance perspective, it's a mess.

965
00:35:20,360 --> 00:35:23,000
We have to decide if that is helpful automation

966
00:35:23,000 --> 00:35:24,600
or if it's a dangerous precedent

967
00:35:24,600 --> 00:35:26,920
for agents to interfere in sensitive business matters.

968
00:35:26,920 --> 00:35:29,800
You need a clear organizational policy to answer these questions.

969
00:35:29,800 --> 00:35:32,360
This isn't a technical setting, it's a decision policy.

970
00:35:32,360 --> 00:35:34,920
You have to define which categories of work allow

971
00:35:34,920 --> 00:35:37,480
for full autonomy and which ones require the agent

972
00:35:37,480 --> 00:35:39,160
to stop and wait for approval.

973
00:35:39,160 --> 00:35:40,520
For high-risk decisions,

974
00:35:40,520 --> 00:35:43,320
the agent should probably stay in an advisory role

975
00:35:43,320 --> 00:35:44,840
rather than becoming an actor.

976
00:35:44,840 --> 00:35:46,520
This is an organizational design challenge

977
00:35:46,520 --> 00:35:48,520
because it forces you to clarify decision rights

978
00:35:48,520 --> 00:35:50,120
that are usually left unsaid.

979
00:35:50,120 --> 00:35:52,200
Most companies run on implicit understandings

980
00:35:52,200 --> 00:35:54,280
where Jane handles certain approvals

981
00:35:54,280 --> 00:35:56,120
and Tom manages resource allocation.

982
00:35:56,120 --> 00:35:58,280
That works fine when humans are using their judgment,

983
00:35:58,280 --> 00:36:00,840
but it breaks when you try to give an agent rules.

984
00:36:00,840 --> 00:36:02,360
Agents don't have judgment,

985
00:36:02,360 --> 00:36:04,200
they only have the rules you give them.

986
00:36:04,200 --> 00:36:07,880
And every rule has an edge case where it might be technically right

987
00:36:07,880 --> 00:36:09,640
but organizationally wrong.

988
00:36:09,640 --> 00:36:11,640
Your governance structure has to be crystal clear

989
00:36:11,640 --> 00:36:13,240
before these agents start running.

990
00:36:13,240 --> 00:36:14,680
You need to know who approves what,

991
00:36:14,680 --> 00:36:15,880
how mistakes get fixed,

992
00:36:15,880 --> 00:36:17,240
and how to resolve a conflict

993
00:36:17,240 --> 00:36:19,800
when an agent's decision clashes with what are human ones.

994
00:36:19,800 --> 00:36:23,000
Without that clarity, agent365 becomes a liability.

995
00:36:23,000 --> 00:36:25,160
It's an autonomous system moving into areas

996
00:36:25,160 --> 00:36:27,480
where authority was never formally defined,

997
00:36:27,480 --> 00:36:29,880
making decisions that might seem logical to a machine

998
00:36:29,880 --> 00:36:32,520
but end up surprising the entire organization.

999
00:36:32,520 --> 00:36:35,240
And this is exactly where we find the connection to ROI.

1000
00:36:35,240 --> 00:36:36,840
Why AI projects fail?

1001
00:36:36,840 --> 00:36:38,280
This governance problem,

1002
00:36:38,280 --> 00:36:40,280
the lack of clarity around agent autonomy

1003
00:36:40,280 --> 00:36:42,920
and decision rights isn't unique to WorkIQ.

1004
00:36:42,920 --> 00:36:45,080
It's symptomatic of a much larger failure pattern

1005
00:36:45,080 --> 00:36:47,880
that has been playing out across enterprise AI for years.

1006
00:36:47,880 --> 00:36:49,720
WorkIQ doesn't solve that pattern.

1007
00:36:49,720 --> 00:36:50,680
It exposes it.

1008
00:36:50,680 --> 00:36:52,440
Here's the reality.

1009
00:36:52,440 --> 00:36:56,600
70 to 85% of enterprise AI projects fail

1010
00:36:56,600 --> 00:36:58,040
to deliver their intended value.

1011
00:36:58,040 --> 00:36:59,800
That isn't a failure of the technology

1012
00:36:59,800 --> 00:37:02,040
but rather a failure of organizational readiness.

1013
00:37:02,040 --> 00:37:04,280
The AI works, the models are sophisticated,

1014
00:37:04,280 --> 00:37:05,960
and the data is accessible.

1015
00:37:05,960 --> 00:37:08,520
Yet somewhere between deployment and the final outcome,

1016
00:37:08,520 --> 00:37:10,600
the initiative stalls or gets abandoned.

1017
00:37:10,600 --> 00:37:13,640
It produces results that simply don't match the promise.

1018
00:37:13,640 --> 00:37:16,440
These failures cluster around a few recurring causes

1019
00:37:16,440 --> 00:37:17,640
and none of them are technical.

1020
00:37:17,640 --> 00:37:20,280
The first is misaligned objectives.

1021
00:37:20,280 --> 00:37:22,280
An organization decides to deploy AI

1022
00:37:22,280 --> 00:37:24,120
and picks a use case that sounds promising

1023
00:37:24,120 --> 00:37:26,120
but the business stakeholders, the data team

1024
00:37:26,120 --> 00:37:28,440
and the technical teams never actually align

1025
00:37:28,440 --> 00:37:29,960
on what success looks like.

1026
00:37:29,960 --> 00:37:31,400
What problem are you solving?

1027
00:37:31,400 --> 00:37:32,120
For whom?

1028
00:37:32,120 --> 00:37:33,240
How will you measure it?

1029
00:37:33,240 --> 00:37:35,240
Organizations often skip this conversation

1030
00:37:35,240 --> 00:37:37,960
because they assume alignment exists when it doesn't.

1031
00:37:37,960 --> 00:37:40,280
By the time they are six months in,

1032
00:37:40,280 --> 00:37:43,080
it becomes clear the project is optimizing for the wrong outcome

1033
00:37:43,080 --> 00:37:45,080
and by then it's too late to redirect.

1034
00:37:45,080 --> 00:37:46,520
The second cause is weak data.

1035
00:37:46,520 --> 00:37:48,280
This isn't about having too little data

1036
00:37:48,280 --> 00:37:50,440
but about data quality and consistency.

1037
00:37:50,440 --> 00:37:52,040
Many organizations have tons of data

1038
00:37:52,040 --> 00:37:53,560
that no one really understands,

1039
00:37:53,560 --> 00:37:55,480
which means they don't know where it came from,

1040
00:37:55,480 --> 00:37:57,800
how it was collected or if it's biased.

1041
00:37:57,800 --> 00:38:00,520
You can build sophisticated AI on top of garbage data

1042
00:38:00,520 --> 00:38:02,680
but the result will just be sophisticated garbage.

1043
00:38:02,680 --> 00:38:05,160
It will feel authoritative, but it will be wrong.

1044
00:38:05,160 --> 00:38:07,320
The third issue is unclear ownership.

1045
00:38:07,320 --> 00:38:10,120
When there is no named person responsible for the outcome,

1046
00:38:10,120 --> 00:38:12,920
a project gets built by a team that eventually moves on.

1047
00:38:12,920 --> 00:38:14,920
The system operates but nobody is accountable

1048
00:38:14,920 --> 00:38:15,960
for whether it's working

1049
00:38:15,960 --> 00:38:18,440
and nobody is incentivized to fix it when it breaks.

1050
00:38:18,440 --> 00:38:20,040
Because nobody owns the relationship

1051
00:38:20,040 --> 00:38:22,760
between the AI output and the business decision it supports,

1052
00:38:22,760 --> 00:38:24,040
the tool becomes orphaned.

1053
00:38:24,040 --> 00:38:25,640
The organization assumes it's working

1054
00:38:25,640 --> 00:38:27,080
because nobody is complaining,

1055
00:38:27,080 --> 00:38:29,160
but in reality it's just being ignored.

1056
00:38:29,160 --> 00:38:30,680
The fourth cause is structural.

1057
00:38:30,680 --> 00:38:32,280
Organizations deploy AI

1058
00:38:32,280 --> 00:38:34,120
but don't redesign their workflows

1059
00:38:34,120 --> 00:38:35,880
or change how decisions are made.

1060
00:38:35,880 --> 00:38:38,840
They treat AI as a tool to accelerate the existing process

1061
00:38:38,840 --> 00:38:41,560
instead of an opportunity to reimagine that process entirely.

1062
00:38:41,560 --> 00:38:43,080
Copilot might write faster emails

1063
00:38:43,080 --> 00:38:45,080
but you're still writing the same number of emails.

1064
00:38:45,080 --> 00:38:46,520
AI generates quicker summaries

1065
00:38:46,520 --> 00:38:49,080
but you're still sitting in the same number of meetings reading them.

1066
00:38:49,080 --> 00:38:50,120
The tool got faster

1067
00:38:50,120 --> 00:38:52,040
but the organization didn't transform.

1068
00:38:52,040 --> 00:38:54,600
Along with that they measure adoption instead of outcomes.

1069
00:38:54,600 --> 00:38:56,200
They track how many people use copilot,

1070
00:38:56,200 --> 00:38:57,240
how many prompts they send

1071
00:38:57,240 --> 00:38:59,240
or how many documents are processed.

1072
00:38:59,240 --> 00:39:00,920
Adoption metrics are easy to measure

1073
00:39:00,920 --> 00:39:02,280
but outcome metrics are harder

1074
00:39:02,280 --> 00:39:04,920
because they require baseline data and patience.

1075
00:39:04,920 --> 00:39:06,600
It's easier to count users than to prove

1076
00:39:06,600 --> 00:39:09,080
that decision speed increased or rework decreased.

1077
00:39:09,080 --> 00:39:11,480
Consequently, organizations report adoption success

1078
00:39:11,480 --> 00:39:13,640
while their actual outcomes stagnate.

1079
00:39:13,640 --> 00:39:15,480
WorkIQ amplifies this problem

1080
00:39:15,480 --> 00:39:17,560
because it's powerful enough to expose it immediately.

1081
00:39:17,560 --> 00:39:20,200
If you deploy WorkIQ with unclear objectives,

1082
00:39:20,200 --> 00:39:22,440
the system will generate outputs that contradict

1083
00:39:22,440 --> 00:39:24,440
organizational assumptions within weeks.

1084
00:39:24,440 --> 00:39:26,440
The objectives become obviously misaligned.

1085
00:39:26,440 --> 00:39:28,200
If you deploy it with weak data,

1086
00:39:28,200 --> 00:39:31,160
the system will confidently reason over bad information

1087
00:39:31,160 --> 00:39:34,120
and produce plausible sounding answers that are completely wrong.

1088
00:39:34,120 --> 00:39:36,200
If you deploy it without clear ownership,

1089
00:39:36,200 --> 00:39:37,400
there's no one responsible

1090
00:39:37,400 --> 00:39:39,640
when the agent takes an action nobody expected.

1091
00:39:39,640 --> 00:39:41,480
If you deploy it without workflow redesign,

1092
00:39:41,480 --> 00:39:42,680
the system works fine

1093
00:39:42,680 --> 00:39:44,360
but adoption stalls because people don't know

1094
00:39:44,360 --> 00:39:46,200
why they should change what they're doing.

1095
00:39:46,200 --> 00:39:49,000
WorkIQ doesn't hide these failures, it surfaces them.

1096
00:39:49,000 --> 00:39:51,160
It accelerates the timeline to visibility,

1097
00:39:51,160 --> 00:39:52,200
which is valuable,

1098
00:39:52,200 --> 00:39:55,160
but only if the organization is ready to act on that visibility.

1099
00:39:55,160 --> 00:39:58,200
Governance clarity must exist before deployment.

1100
00:39:58,200 --> 00:39:59,960
This is where organizational intelligence

1101
00:39:59,960 --> 00:40:01,160
becomes the real constraint.

1102
00:40:01,160 --> 00:40:03,240
The management model problem.

1103
00:40:03,240 --> 00:40:05,080
The management model in most organizations

1104
00:40:05,080 --> 00:40:07,800
is built on an assumption that became obsolete decades ago.

1105
00:40:07,800 --> 00:40:10,600
It's the idea that activity correlates with productivity

1106
00:40:10,600 --> 00:40:12,680
that hours worked predict output

1107
00:40:12,680 --> 00:40:15,160
and that meetings attended reflect engagement.

1108
00:40:15,160 --> 00:40:18,280
Manages assume that if people look busy, they're being effective.

1109
00:40:18,280 --> 00:40:21,240
Management by activity is easier than management by outcomes

1110
00:40:21,240 --> 00:40:23,480
because activity is visible and simple to count.

1111
00:40:23,480 --> 00:40:25,080
You count hours, you count meetings,

1112
00:40:25,080 --> 00:40:26,040
and you count deliverables

1113
00:40:26,040 --> 00:40:27,080
so you don't have to think about

1114
00:40:27,080 --> 00:40:28,920
whether any of it actually matters.

1115
00:40:28,920 --> 00:40:30,840
WorkIQ doesn't care about activity.

1116
00:40:30,840 --> 00:40:32,680
It cares about outcomes, it observes

1117
00:40:32,680 --> 00:40:34,120
what actually drives decisions

1118
00:40:34,120 --> 00:40:36,920
and what actually moves projects forward to create value.

1119
00:40:36,920 --> 00:40:38,920
What it observes in most organizations

1120
00:40:38,920 --> 00:40:41,000
would be deeply uncomfortable to see your leadership

1121
00:40:41,000 --> 00:40:42,920
if they had to articulate it clearly.

1122
00:40:42,920 --> 00:40:44,120
The system can show you exactly

1123
00:40:44,120 --> 00:40:45,560
who sits in meetings all day

1124
00:40:45,560 --> 00:40:47,160
and never decides anything.

1125
00:40:47,160 --> 00:40:50,440
You can see the calendar filled with nine hours of back-to-back meetings

1126
00:40:50,440 --> 00:40:52,680
which looks productive and sounds exhausting.

1127
00:40:52,680 --> 00:40:55,000
However, WorkIQ can correlate that meeting load

1128
00:40:55,000 --> 00:40:57,400
with actual decision output and closed items.

1129
00:40:57,400 --> 00:40:59,000
Often the correlation is zero.

1130
00:40:59,000 --> 00:41:01,880
The person in nine hours of meetings generated no decisions

1131
00:41:01,880 --> 00:41:04,200
started no new initiatives and closed nothing.

1132
00:41:04,200 --> 00:41:06,680
The meetings created the illusion of productivity

1133
00:41:06,680 --> 00:41:09,240
and the activity created the appearance of contribution

1134
00:41:09,240 --> 00:41:10,680
but the outcome was nothing.

1135
00:41:10,680 --> 00:41:13,880
WorkIQ can surface who collaborates across teams extensively

1136
00:41:13,880 --> 00:41:15,480
but never drives closure.

1137
00:41:15,480 --> 00:41:17,880
They are in the conversations, they are involved in the discussions

1138
00:41:17,880 --> 00:41:19,320
and they are included in the decisions

1139
00:41:19,320 --> 00:41:22,200
but the decisions don't actually happen until they leave the room.

1140
00:41:22,200 --> 00:41:25,080
Outcomes don't move until someone else takes ownership.

1141
00:41:25,080 --> 00:41:28,120
The collaboration is real but the impact is a phantom.

1142
00:41:28,120 --> 00:41:31,240
They are networked extensively but central to nothing.

1143
00:41:31,240 --> 00:41:34,440
WorkIQ can also identify who has access to critical information

1144
00:41:34,440 --> 00:41:35,560
but never uses it.

1145
00:41:35,560 --> 00:41:37,320
These people are on distribution lists,

1146
00:41:37,320 --> 00:41:38,920
they are invited to meetings

1147
00:41:38,920 --> 00:41:40,280
and they are added to the channels

1148
00:41:40,280 --> 00:41:41,560
where decisions are documented

1149
00:41:41,560 --> 00:41:44,120
but they don't read the information, they don't act on it

1150
00:41:44,120 --> 00:41:45,800
and they don't integrate it into their work.

1151
00:41:45,800 --> 00:41:48,440
The access is granted but the consumption is absent.

1152
00:41:48,440 --> 00:41:50,600
This pattern is visible in how they operate

1153
00:41:50,600 --> 00:41:52,280
as projects advance without their input

1154
00:41:52,280 --> 00:41:54,280
because they were never actually paying attention.

1155
00:41:54,280 --> 00:41:56,200
This visibility is operationally valuable

1156
00:41:56,200 --> 00:41:58,760
and could drive real organizational transformation.

1157
00:41:58,760 --> 00:42:01,080
Managers could see clearly where energy is wasted

1158
00:42:01,080 --> 00:42:03,560
and where activity is mistaken for contribution.

1159
00:42:03,560 --> 00:42:05,960
They could see where people are kept busy but not productive

1160
00:42:05,960 --> 00:42:08,200
and where structures create the appearance of collaboration

1161
00:42:08,200 --> 00:42:09,800
without generating outcomes

1162
00:42:09,800 --> 00:42:12,040
but this visibility is also threatening to managers

1163
00:42:12,040 --> 00:42:13,480
who depend on opacity.

1164
00:42:13,480 --> 00:42:16,680
These are managers whose authority rests on controlling the flow of information

1165
00:42:16,680 --> 00:42:18,920
and who benefit from ambiguous decision rights.

1166
00:42:18,920 --> 00:42:20,520
They measure subordinates by activity

1167
00:42:20,520 --> 00:42:24,120
because measuring by outcomes would expose that the work doesn't matter.

1168
00:42:24,120 --> 00:42:27,080
They keep people in meetings because meetings look like management.

1169
00:42:27,080 --> 00:42:30,200
Their own value proposition depends on decisions being hard to trace

1170
00:42:30,200 --> 00:42:31,960
and outcomes being hard to measure.

1171
00:42:31,960 --> 00:42:34,360
That's the reality work IQ threatens to expose.

1172
00:42:34,360 --> 00:42:37,240
In many organizations, management layers exist to manage opacity

1173
00:42:37,240 --> 00:42:40,120
and translate activity into the appearance of progress.

1174
00:42:40,120 --> 00:42:42,600
They protect people from having to answer the hardest question.

1175
00:42:42,600 --> 00:42:44,440
What did you actually accomplish?

1176
00:42:44,440 --> 00:42:47,080
Work IQ makes that question impossible to avoid.

1177
00:42:47,080 --> 00:42:51,000
When the system shows clearly that your team generated two decisions in a month

1178
00:42:51,000 --> 00:42:52,920
while the adjacent team generated 12,

1179
00:42:52,920 --> 00:42:55,000
you can't argue that they're equally productive.

1180
00:42:55,000 --> 00:42:57,800
When the system demonstrates that half your meeting time is ceremonial

1181
00:42:57,800 --> 00:42:59,160
and half is consequential,

1182
00:42:59,160 --> 00:43:01,480
you can't argue for the current meeting culture.

1183
00:43:01,480 --> 00:43:03,080
When the system maps decision flows

1184
00:43:03,080 --> 00:43:05,320
and shows that critical decisions are made by three people

1185
00:43:05,320 --> 00:43:07,720
while 30 others are kept in consultative loops,

1186
00:43:07,720 --> 00:43:09,320
you can't argue that those loops matter.

1187
00:43:10,120 --> 00:43:14,440
Organizations that build management systems around activity will resist work IQ fiercely.

1188
00:43:14,440 --> 00:43:15,880
They won't do it consciously.

1189
00:43:15,880 --> 00:43:18,120
Publicly, they will say they welcome transparency,

1190
00:43:18,120 --> 00:43:20,440
but operationally, the resistance will be real.

1191
00:43:20,440 --> 00:43:24,200
Adoption will stall and the system will be tolerated but not genuinely used.

1192
00:43:24,200 --> 00:43:27,640
Genuine use requires changing how manages themselves are evaluated.

1193
00:43:27,640 --> 00:43:31,240
It requires holding them accountable for outcomes instead of activity

1194
00:43:31,240 --> 00:43:34,280
and demanding clarity about what each person actually does.

1195
00:43:34,280 --> 00:43:37,160
This is why adoption will be slower than Microsoft expects.

1196
00:43:37,160 --> 00:43:38,360
The culture collision.

1197
00:43:38,360 --> 00:43:41,640
Work IQ does more than just point out where individuals are struggling.

1198
00:43:41,640 --> 00:43:44,680
It makes the invisible friction inside your company visible.

1199
00:43:44,680 --> 00:43:47,480
It builds a map of exactly where things are breaking down.

1200
00:43:47,480 --> 00:43:48,840
You see the communication gaps.

1201
00:43:48,840 --> 00:43:50,520
You see the bottlenecks that kill your speed.

1202
00:43:50,520 --> 00:43:52,840
You see the silos where information goes to die.

1203
00:43:52,840 --> 00:43:55,560
It shows you teams that are working against each other,

1204
00:43:55,560 --> 00:43:59,000
simply because nobody actually knows what anyone else is doing.

1205
00:43:59,000 --> 00:44:02,040
When leaders look at this map, they can finally point to the root cause.

1206
00:44:02,040 --> 00:44:05,880
They can use real data to show that one division is operating like an island

1207
00:44:05,880 --> 00:44:08,760
and that isolation is costing the company months of progress.

1208
00:44:08,760 --> 00:44:11,880
They can see that a single decision is waiting on five different people

1209
00:44:11,880 --> 00:44:14,600
who don't even realize they're all involved in the same process.

1210
00:44:14,600 --> 00:44:18,200
They can see a project failing because it depends on data from another department

1211
00:44:18,200 --> 00:44:19,960
that nobody is actually updating.

1212
00:44:19,960 --> 00:44:21,560
The dysfunction isn't a feeling anymore.

1213
00:44:21,560 --> 00:44:22,440
It's quantifiable.

1214
00:44:22,440 --> 00:44:24,680
It's real and it's undeniable.

1215
00:44:24,680 --> 00:44:25,720
But here's the problem.

1216
00:44:25,720 --> 00:44:28,200
Seeing the mess and actually being willing to clean it up

1217
00:44:28,200 --> 00:44:29,800
are two very different things.

1218
00:44:29,800 --> 00:44:33,320
Most organizations will fight against fixing the problems that work IQ reveals.

1219
00:44:33,320 --> 00:44:35,720
It isn't because they don't see the value in a better system

1220
00:44:35,720 --> 00:44:38,760
it's because fixing the system means changing how power works.

1221
00:44:38,760 --> 00:44:39,960
It means moving authority.

1222
00:44:39,960 --> 00:44:43,240
It means tearing down the very silos that certain leaders have spent years building

1223
00:44:43,240 --> 00:44:45,640
silos create domains and domains create fiefdoms.

1224
00:44:45,640 --> 00:44:48,600
If you lead a silo, you have total authority over that space.

1225
00:44:48,600 --> 00:44:51,000
If you break the silo, your authority disappears.

1226
00:44:51,000 --> 00:44:53,240
Nobody wants to volunteer for that kind of demotion.

1227
00:44:53,240 --> 00:44:56,040
The path of least resistance is to let work IQ find the floors

1228
00:44:56,040 --> 00:44:57,640
and then do absolutely nothing about them.

1229
00:44:57,640 --> 00:45:00,360
It's easy to acknowledge the insight and just keep moving.

1230
00:45:00,360 --> 00:45:02,920
Leaders will say they've known about the problem for years

1231
00:45:02,920 --> 00:45:06,440
but they'll claim a reorganization is too difficult to handle right now.

1232
00:45:06,440 --> 00:45:07,880
Later never actually comes.

1233
00:45:07,880 --> 00:45:09,400
The dysfunction stays visible

1234
00:45:09,400 --> 00:45:11,960
but the structures that created it stay exactly where they are.

1235
00:45:11,960 --> 00:45:13,880
Some companies will take an even darker path.

1236
00:45:13,880 --> 00:45:16,040
They won't use this visibility to improve the system.

1237
00:45:16,040 --> 00:45:17,400
They'll use it to watch the people.

1238
00:45:17,400 --> 00:45:19,720
They'll use work IQ to track who isn't hitting their numbers

1239
00:45:19,720 --> 00:45:21,400
or who's spending too much time in meetings.

1240
00:45:21,400 --> 00:45:23,320
They'll monitor who isn't collaborating enough

1241
00:45:23,320 --> 00:45:25,400
or who's talking to the wrong people.

1242
00:45:25,400 --> 00:45:27,640
The same data that could have redesigned the company

1243
00:45:27,640 --> 00:45:29,720
gets turned into a weapon for surveillance.

1244
00:45:29,720 --> 00:45:31,880
The system stops being a way to understand work

1245
00:45:31,880 --> 00:45:33,880
and starts being a way to measure humans.

1246
00:45:33,880 --> 00:45:36,120
This distinction is everything because it's what determines

1247
00:45:36,120 --> 00:45:38,360
if your team trusts the system or hates it.

1248
00:45:38,360 --> 00:45:41,880
If you use the data to fix broken workflows, people feel supported.

1249
00:45:41,880 --> 00:45:44,840
They see the system as a tool that helps them understand

1250
00:45:44,840 --> 00:45:46,600
their own process and work better.

1251
00:45:46,600 --> 00:45:48,120
Trust goes up.

1252
00:45:48,120 --> 00:45:51,000
But if you use it to judge individuals against hidden standards,

1253
00:45:51,000 --> 00:45:52,440
people feel exposed.

1254
00:45:52,440 --> 00:45:55,160
They realize the system is just a digital supervisor taking notes

1255
00:45:55,160 --> 00:45:56,520
and building a case against them.

1256
00:45:56,520 --> 00:45:59,880
The moment that feeling of surveillance kicks in, adoption is dead.

1257
00:45:59,880 --> 00:46:02,680
People will find workarounds. They'll hide what they're doing.

1258
00:46:02,680 --> 00:46:05,160
They'll start optimizing for the metrics they think you're watching

1259
00:46:05,160 --> 00:46:07,240
instead of focusing on the work that actually matters.

1260
00:46:07,240 --> 00:46:08,680
The system becomes the enemy.

1261
00:46:08,680 --> 00:46:10,200
The technology itself doesn't care.

1262
00:46:10,200 --> 00:46:12,280
The data is the same and the patterns are identical

1263
00:46:12,280 --> 00:46:13,640
regardless of how you use them.

1264
00:46:13,640 --> 00:46:15,160
The difference is entirely in your culture.

1265
00:46:15,160 --> 00:46:17,160
Is this a tool for understanding how we work?

1266
00:46:17,160 --> 00:46:19,160
Or is it a tool for judging how well we work?

1267
00:46:19,160 --> 00:46:21,480
Are we making the system visible so we can fix it?

1268
00:46:21,480 --> 00:46:23,720
Or making the people visible so we can rank them?

1269
00:46:23,720 --> 00:46:25,240
Companies that have built real trust

1270
00:46:25,240 --> 00:46:27,080
will use work IQ as an accelerator.

1271
00:46:27,080 --> 00:46:29,400
They have the maturity to face what the data shows them

1272
00:46:29,400 --> 00:46:31,720
and the flexibility to actually change their habits.

1273
00:46:31,720 --> 00:46:34,360
But organizations that run on secrets and use busyness

1274
00:46:34,360 --> 00:46:38,440
as a proxy for value will either resist this technology or corrupt it.

1275
00:46:38,440 --> 00:46:41,160
They'll block it from being used or they'll turn it into a spy tool.

1276
00:46:41,160 --> 00:46:44,840
Your culture is what decides if this intelligence layer

1277
00:46:44,840 --> 00:46:46,760
becomes a way to transform your business

1278
00:46:46,760 --> 00:46:49,000
or just another way to control your employees.

1279
00:46:49,000 --> 00:46:51,880
And that is exactly where governance and ethics meet.

1280
00:46:51,880 --> 00:46:53,160
The governance framework.

1281
00:46:53,160 --> 00:46:55,880
This is the point where you move from just talking about the problem

1282
00:46:55,880 --> 00:46:58,120
to building the actual infrastructure to manage it.

1283
00:46:58,120 --> 00:47:00,920
None of the things we've talked about agents working on their own,

1284
00:47:00,920 --> 00:47:03,080
memory that lasts or data being pulled together

1285
00:47:03,080 --> 00:47:05,640
can happen safely without a solid governance framework.

1286
00:47:05,640 --> 00:47:07,800
That framework has to cover four specific areas.

1287
00:47:07,800 --> 00:47:09,720
The first area is data access.

1288
00:47:09,720 --> 00:47:11,720
You have to decide which parts of your company

1289
00:47:11,720 --> 00:47:13,480
work IQ is allowed to look at.

1290
00:47:13,480 --> 00:47:17,000
Not every database or file system needs to be part of the intelligence layer.

1291
00:47:17,000 --> 00:47:19,640
You need to make a deliberate choice about what's in and what's out.

1292
00:47:19,640 --> 00:47:22,200
Your main Microsoft 365 environment is a yes.

1293
00:47:22,200 --> 00:47:23,640
Your CRM is likely a yes.

1294
00:47:23,640 --> 00:47:26,120
But your old archives or a developer's private sandbox

1295
00:47:26,120 --> 00:47:27,720
those should probably stay off limits.

1296
00:47:27,720 --> 00:47:29,640
You aren't saying the system can never see them.

1297
00:47:29,640 --> 00:47:31,720
You're deciding which systems you're comfortable with

1298
00:47:31,720 --> 00:47:34,360
and AI reasoning about every single day.

1299
00:47:34,360 --> 00:47:36,280
The second area is agent behavior.

1300
00:47:36,280 --> 00:47:38,680
Once an agent is inside those data boundaries,

1301
00:47:38,680 --> 00:47:40,360
what is it actually allowed to do?

1302
00:47:40,360 --> 00:47:42,680
Which actions need a human to click approve

1303
00:47:42,680 --> 00:47:44,520
and which ones can happen automatically?

1304
00:47:44,520 --> 00:47:46,600
This is where your policy engine becomes vital.

1305
00:47:46,600 --> 00:47:48,280
You aren't just writing static rules.

1306
00:47:48,280 --> 00:47:50,360
You're writing policies that understand context.

1307
00:47:50,360 --> 00:47:51,800
If an agent is managing your email,

1308
00:47:51,800 --> 00:47:54,760
it might be fine for it to move messages into folders on its own.

1309
00:47:54,760 --> 00:47:57,080
But that same agent shouldn't be allowed to delete an email

1310
00:47:57,080 --> 00:47:58,040
without you saying so.

1311
00:47:58,040 --> 00:47:59,640
If an agent is building a project plan,

1312
00:47:59,640 --> 00:48:01,160
it can draft the proposal.

1313
00:48:01,160 --> 00:48:03,560
But it shouldn't be able to assign tasks to your team

1314
00:48:03,560 --> 00:48:05,160
until they've seen the plan first.

1315
00:48:05,160 --> 00:48:08,120
Every policy is a line you draw between a suggestion

1316
00:48:08,120 --> 00:48:09,240
and an action.

1317
00:48:09,240 --> 00:48:10,520
The third area is memory.

1318
00:48:10,520 --> 00:48:12,520
What is the system allowed to remember and for how long?

1319
00:48:12,520 --> 00:48:13,720
Who gets to see that memory?

1320
00:48:13,720 --> 00:48:15,320
If the system gets something wrong,

1321
00:48:15,320 --> 00:48:16,680
how do you go in and fix it?

1322
00:48:16,680 --> 00:48:18,280
These aren't just technical details.

1323
00:48:18,280 --> 00:48:20,120
Memory is where the system gets its value,

1324
00:48:20,120 --> 00:48:22,600
but it's also where the risk of surveillance is the highest.

1325
00:48:22,600 --> 00:48:24,040
If you don't govern memory,

1326
00:48:24,040 --> 00:48:27,160
the system becomes a permanent record of every mistake

1327
00:48:27,160 --> 00:48:29,000
that can be used against someone later.

1328
00:48:29,000 --> 00:48:31,240
You need clear rules for how long data is kept

1329
00:48:31,240 --> 00:48:33,720
and you need ways for users to see what's being remembered.

1330
00:48:33,720 --> 00:48:36,360
You want a system that learns your preferences

1331
00:48:36,360 --> 00:48:39,000
without keeping a permanent file on your behavior.

1332
00:48:39,000 --> 00:48:41,720
The fourth area is accountability for outcomes.

1333
00:48:41,720 --> 00:48:43,080
When an agent makes a move,

1334
00:48:43,080 --> 00:48:44,680
who is responsible for the result?

1335
00:48:44,680 --> 00:48:46,840
If an agent sends a customer to the wrong department,

1336
00:48:46,840 --> 00:48:48,040
who owns that mistake?

1337
00:48:48,040 --> 00:48:49,480
Is it the person who built the agent

1338
00:48:49,480 --> 00:48:51,320
or the person the agent was working for?

1339
00:48:51,320 --> 00:48:52,600
You need total clarity here

1340
00:48:52,600 --> 00:48:54,680
because responsibility without a clear owner

1341
00:48:54,680 --> 00:48:56,040
just leads to finger pointing.

1342
00:48:56,040 --> 00:48:58,360
Every action an agent takes has a consequence

1343
00:48:58,360 --> 00:49:00,840
and that means every action needs a human owner.

1344
00:49:00,840 --> 00:49:01,640
If something breaks,

1345
00:49:01,640 --> 00:49:03,640
someone needs to be able to explain why it happened

1346
00:49:03,640 --> 00:49:04,840
and how to fix it.

1347
00:49:04,840 --> 00:49:07,800
These four areas work together as one single system.

1348
00:49:07,800 --> 00:49:09,880
Access rules decide what the agency,

1349
00:49:09,880 --> 00:49:11,720
behavior rules decide what they do,

1350
00:49:11,720 --> 00:49:14,120
memory rules decide what stays in the system

1351
00:49:14,120 --> 00:49:16,280
and accountability ensures that when things go wrong,

1352
00:49:16,280 --> 00:49:17,960
there's a path to a solution.

1353
00:49:17,960 --> 00:49:20,360
This isn't a "set it and forget it" configuration.

1354
00:49:20,360 --> 00:49:22,040
Governance is a constant discipline.

1355
00:49:22,040 --> 00:49:23,720
It's a function of how you run the company.

1356
00:49:23,720 --> 00:49:26,360
You need people in specific roles to manage this.

1357
00:49:26,360 --> 00:49:29,320
You need AI stewards who can help teams design agents

1358
00:49:29,320 --> 00:49:30,520
that follow the rules.

1359
00:49:30,520 --> 00:49:32,280
You need work designers who can figure out

1360
00:49:32,280 --> 00:49:34,200
where an agent should be autonomous

1361
00:49:34,200 --> 00:49:35,800
and where a human needs to step in.

1362
00:49:35,800 --> 00:49:37,560
You need leaders who can adjust the policies

1363
00:49:37,560 --> 00:49:39,640
as the company learns what works and what doesn't.

1364
00:49:39,640 --> 00:49:41,720
The framework isn't there to slow you down.

1365
00:49:41,720 --> 00:49:43,880
It exists to make the technology safe enough

1366
00:49:43,880 --> 00:49:45,480
so that you can actually move fast.

1367
00:49:45,480 --> 00:49:46,520
Without these rules,

1368
00:49:46,520 --> 00:49:47,640
organizations get nervous

1369
00:49:47,640 --> 00:49:49,240
and the whole project stores out.

1370
00:49:49,240 --> 00:49:51,080
The real power of work IQ stays locked away

1371
00:49:51,080 --> 00:49:52,840
because the company is too afraid to use it.

1372
00:49:52,840 --> 00:49:54,520
But with the right structure in place,

1373
00:49:54,520 --> 00:49:55,800
you can empower your teams

1374
00:49:55,800 --> 00:49:58,440
to build and deploy agents with total confidence.

1375
00:49:58,440 --> 00:50:00,280
But even the best governance won't matter

1376
00:50:00,280 --> 00:50:02,040
if you don't know what you're trying to achieve.

1377
00:50:02,040 --> 00:50:03,400
You have to know what you're measuring.

1378
00:50:03,400 --> 00:50:05,640
Measuring intelligence?

1379
00:50:05,640 --> 00:50:06,760
Not activity.

1380
00:50:06,760 --> 00:50:08,680
The moment you stop tracking emails,

1381
00:50:08,680 --> 00:50:10,680
send and meetings attended,

1382
00:50:10,680 --> 00:50:12,440
you have to answer a much harder question.

1383
00:50:12,440 --> 00:50:14,120
What are you actually measuring instead?

1384
00:50:14,120 --> 00:50:16,440
The old metrics at least had the virtue of being simple

1385
00:50:16,440 --> 00:50:17,800
because you just counted things

1386
00:50:17,800 --> 00:50:19,240
and watched the numbers go up.

1387
00:50:19,240 --> 00:50:20,440
That looked like progress.

1388
00:50:20,440 --> 00:50:22,360
But new metrics require work IQ

1389
00:50:22,360 --> 00:50:23,480
and they ask about patterns

1390
00:50:23,480 --> 00:50:24,920
that don't exist in activity logs.

1391
00:50:24,920 --> 00:50:26,200
They exist in outcomes.

1392
00:50:26,200 --> 00:50:28,600
Decision speed is a genuine operational metric

1393
00:50:28,600 --> 00:50:31,000
but it isn't about how fast people respond to messages.

1394
00:50:31,000 --> 00:50:32,360
That's just more activity.

1395
00:50:32,360 --> 00:50:35,000
Real decision speed means information enters the organization

1396
00:50:35,000 --> 00:50:36,200
and a decision gets made.

1397
00:50:36,200 --> 00:50:38,200
So you have to look at how long that cycle takes

1398
00:50:38,200 --> 00:50:39,960
and where the information gets stuck.

1399
00:50:39,960 --> 00:50:41,240
Work IQ can map this

1400
00:50:41,240 --> 00:50:43,160
because it understands when information appears,

1401
00:50:43,160 --> 00:50:43,960
when it's reviewed,

1402
00:50:43,960 --> 00:50:45,880
and when the right people finally get involved

1403
00:50:45,880 --> 00:50:46,920
to communicate a result.

1404
00:50:46,920 --> 00:50:48,680
The system sees the entire flow

1405
00:50:48,680 --> 00:50:50,200
rather than just individual actions.

1406
00:50:50,440 --> 00:50:51,880
In most organizations,

1407
00:50:51,880 --> 00:50:54,360
you simply can't measure this without work IQ

1408
00:50:54,360 --> 00:50:56,680
because you'd have to manually trace every detail

1409
00:50:56,680 --> 00:50:58,840
through email meetings and documents.

1410
00:50:58,840 --> 00:51:00,760
The system does it for you continuously.

1411
00:51:00,760 --> 00:51:03,240
Cross-team collaboration quality is also measurable

1412
00:51:03,240 --> 00:51:05,080
but not in the way most leaders think.

1413
00:51:05,080 --> 00:51:06,760
It's not about the volume of emails

1414
00:51:06,760 --> 00:51:07,880
moving between departments.

1415
00:51:07,880 --> 00:51:10,360
It's about whether that collaboration actually produces

1416
00:51:10,360 --> 00:51:11,160
better results.

1417
00:51:11,160 --> 00:51:13,560
You have to ask if a decision involving multiple teams

1418
00:51:13,560 --> 00:51:14,840
had a lower rework rate

1419
00:51:14,840 --> 00:51:16,040
or if it moved faster

1420
00:51:16,040 --> 00:51:18,680
because the right expertise was involved early on.

1421
00:51:18,680 --> 00:51:21,480
Work IQ sees the actual structure of these decisions.

1422
00:51:21,480 --> 00:51:22,920
It knows which people were consulted

1423
00:51:22,920 --> 00:51:25,160
and which ones actually influenced the final outcome.

1424
00:51:25,160 --> 00:51:27,320
It can tell which collaborations were just for show

1425
00:51:27,320 --> 00:51:28,920
and which ones changed the result,

1426
00:51:28,920 --> 00:51:31,320
allowing you to measure the impact of teamwork directly.

1427
00:51:31,320 --> 00:51:34,360
Context reuse is where work IQ becomes essential

1428
00:51:34,360 --> 00:51:36,920
because most organizations reconstruct context constantly.

1429
00:51:36,920 --> 00:51:38,360
A person learns something in a meeting

1430
00:51:38,360 --> 00:51:39,720
and writes it in an email,

1431
00:51:39,720 --> 00:51:41,000
then someone else reads that email

1432
00:51:41,000 --> 00:51:43,080
and has the same conversation with a third person.

1433
00:51:43,080 --> 00:51:45,480
The same context gets explained over and over again.

1434
00:51:45,480 --> 00:51:47,960
Work IQ measures whether information is actually flowing

1435
00:51:47,960 --> 00:51:49,080
and being reused

1436
00:51:49,080 --> 00:51:52,200
or if the organization is locked in a cycle of repetitive explanation.

1437
00:51:52,200 --> 00:51:54,120
If the same project context is being discussed

1438
00:51:54,120 --> 00:51:56,920
in eight separate meetings instead of being shared once,

1439
00:51:56,920 --> 00:51:58,200
the system sees that waste

1440
00:51:58,200 --> 00:52:00,680
and quantifies what better collaboration would look like.

1441
00:52:00,680 --> 00:52:02,120
Outcome quality measures,

1442
00:52:02,120 --> 00:52:05,720
whether your decisions actually produce the results you intended.

1443
00:52:05,720 --> 00:52:08,120
This sounds simple but it's rarely measured at all.

1444
00:52:08,120 --> 00:52:09,160
You make a big decision

1445
00:52:09,160 --> 00:52:11,720
but you don't track whether it had the effect you expected.

1446
00:52:11,720 --> 00:52:14,360
Work IQ allows you to correlate those choices

1447
00:52:14,360 --> 00:52:16,040
with everything that happens next.

1448
00:52:16,040 --> 00:52:18,600
If a decision was made to focus on customer acquisition,

1449
00:52:18,600 --> 00:52:20,120
did the team actually focus there

1450
00:52:20,120 --> 00:52:22,120
or did they drift back to maintenance work?

1451
00:52:22,120 --> 00:52:23,720
And when you shut down a product line,

1452
00:52:23,720 --> 00:52:25,400
did those resources redeploy

1453
00:52:25,400 --> 00:52:27,720
or did they just evaporate into other projects?

1454
00:52:27,720 --> 00:52:29,000
These are measurable outcomes

1455
00:52:29,000 --> 00:52:30,760
but they require connecting a decision

1456
00:52:30,760 --> 00:52:32,360
to what actually happens afterward.

1457
00:52:32,360 --> 00:52:34,600
The critical point is that these metrics require

1458
00:52:34,600 --> 00:52:35,720
continuous intelligence.

1459
00:52:35,720 --> 00:52:38,200
You can't measure decision speed with a spreadsheet

1460
00:52:38,200 --> 00:52:40,440
and you certainly can't track context reuse

1461
00:52:40,440 --> 00:52:42,040
by reading through old email threads.

1462
00:52:42,040 --> 00:52:44,600
You can't assess the impact of collaboration from a survey.

1463
00:52:44,600 --> 00:52:46,600
Work IQ makes these measurements automatic

1464
00:52:46,600 --> 00:52:49,640
because the system is reasoning over work patterns every single day.

1465
00:52:49,640 --> 00:52:50,840
It isn't creating new data.

1466
00:52:50,840 --> 00:52:52,760
It's analyzing the data you already have

1467
00:52:52,760 --> 00:52:56,040
in ways that expose metrics you never had visibility into before.

1468
00:52:56,040 --> 00:52:57,880
The moment you commit to measuring intelligence

1469
00:52:57,880 --> 00:52:59,080
instead of activity,

1470
00:52:59,080 --> 00:53:01,320
you've made a choice about what the organization values.

1471
00:53:01,320 --> 00:53:03,720
You're saying that you care whether decisions are fast,

1472
00:53:03,720 --> 00:53:05,240
not whether people look busy,

1473
00:53:05,240 --> 00:53:06,920
you care whether expertise is involved,

1474
00:53:06,920 --> 00:53:08,920
not how many people sat in the meeting,

1475
00:53:08,920 --> 00:53:10,840
you care whether context is being reused,

1476
00:53:10,840 --> 00:53:12,600
not how many emails are being sent,

1477
00:53:12,600 --> 00:53:14,840
you care whether decisions produce results,

1478
00:53:14,840 --> 00:53:16,680
not how decisively you announce them.

1479
00:53:16,680 --> 00:53:18,520
That commitment to measurement cascades

1480
00:53:18,520 --> 00:53:20,200
through the entire organization.

1481
00:53:20,200 --> 00:53:22,120
Leadership as orchestration.

1482
00:53:22,120 --> 00:53:24,200
The traditional manager role emerged in a world

1483
00:53:24,200 --> 00:53:26,200
where information was scarce and authority

1484
00:53:26,200 --> 00:53:27,720
was strictly hierarchical.

1485
00:53:27,720 --> 00:53:29,320
The manager was the gatekeeper.

1486
00:53:29,320 --> 00:53:31,160
They were the person with access to information

1487
00:53:31,160 --> 00:53:32,040
others didn't have

1488
00:53:32,040 --> 00:53:34,280
and the decision authority others didn't possess.

1489
00:53:34,280 --> 00:53:36,120
Managers controlled the flow of information

1490
00:53:36,120 --> 00:53:38,760
because that was how you controlled the direction of the company.

1491
00:53:38,760 --> 00:53:40,360
You released details strategically

1492
00:53:40,360 --> 00:53:41,560
and made decisions carefully

1493
00:53:41,560 --> 00:53:43,480
because you were the bottleneck by design.

1494
00:53:43,480 --> 00:53:44,920
That was what management meant.

1495
00:53:44,920 --> 00:53:46,120
The scarcity of information

1496
00:53:46,120 --> 00:53:47,560
and the clarity of authority

1497
00:53:47,560 --> 00:53:49,320
created a very specific role.

1498
00:53:49,320 --> 00:53:51,240
Work IQ eliminates that scarcity,

1499
00:53:51,240 --> 00:53:53,240
information flows continuously now

1500
00:53:53,240 --> 00:53:56,120
and patterns are visible to anyone with access to the system.

1501
00:53:56,120 --> 00:53:57,960
Decisions are mapped and traceable.

1502
00:53:57,960 --> 00:53:59,160
Authority becomes visible too

1503
00:53:59,160 --> 00:54:00,520
and I don't mean what the org chart says

1504
00:54:00,520 --> 00:54:01,640
but what the system observes

1505
00:54:01,640 --> 00:54:04,040
about who actually influences outcomes.

1506
00:54:04,040 --> 00:54:05,800
The gatekeeping function doesn't disappear

1507
00:54:05,800 --> 00:54:08,040
but it transforms into something entirely different.

1508
00:54:08,040 --> 00:54:09,560
The new role is orchestration.

1509
00:54:09,560 --> 00:54:12,440
It isn't about making every decision or guarding every gate.

1510
00:54:12,440 --> 00:54:14,840
A manager in a work IQ environment needs to clarify

1511
00:54:14,840 --> 00:54:16,600
which decisions humans should make

1512
00:54:16,600 --> 00:54:18,200
and which ones agents should handle.

1513
00:54:18,200 --> 00:54:19,880
This isn't a philosophical exercise.

1514
00:54:19,880 --> 00:54:21,320
It's practical and operational.

1515
00:54:21,320 --> 00:54:23,560
You have to decide if a choice requires human judgment

1516
00:54:23,560 --> 00:54:24,840
because the context is new

1517
00:54:24,840 --> 00:54:28,040
or if an agent can propose options for a human to choose from.

1518
00:54:28,040 --> 00:54:31,320
You might decide an agent can act autonomously within set boundaries

1519
00:54:31,320 --> 00:54:33,480
or that a specific issue must be escalated

1520
00:54:33,480 --> 00:54:34,760
to leadership every time.

1521
00:54:34,760 --> 00:54:35,880
This level of clarity

1522
00:54:35,880 --> 00:54:37,880
doesn't exist in most organizations today.

1523
00:54:37,880 --> 00:54:39,880
Most managers inherit a lot of ambiguity

1524
00:54:39,880 --> 00:54:42,200
and operate according to informal understandings.

1525
00:54:42,200 --> 00:54:43,560
Jane handles certain approvals,

1526
00:54:43,560 --> 00:54:44,760
Tom makes staffing decisions

1527
00:54:44,760 --> 00:54:46,600
and Sarah decides what to say to the public.

1528
00:54:46,600 --> 00:54:48,360
Nobody wrote this down or formalized it.

1529
00:54:48,360 --> 00:54:49,640
It just evolved over time.

1530
00:54:49,640 --> 00:54:52,040
Everyone who's been around long enough knows the pattern

1531
00:54:52,040 --> 00:54:53,480
while new people learn it slowly

1532
00:54:53,480 --> 00:54:55,480
by doing things wrong and getting corrected.

1533
00:54:55,480 --> 00:54:58,120
With agents you can't operate on informal understanding.

1534
00:54:58,120 --> 00:54:59,240
You have to make it explicit.

1535
00:54:59,240 --> 00:55:00,600
If you don't agents will operate

1536
00:55:00,600 --> 00:55:02,760
on their best guess of what you want them to do.

1537
00:55:02,760 --> 00:55:04,200
Their guess will sometimes be right

1538
00:55:04,200 --> 00:55:06,680
but other times it will be spectacularly wrong.

1539
00:55:06,680 --> 00:55:07,560
Either way,

1540
00:55:07,560 --> 00:55:10,120
you've trained people to expect decisions to happen

1541
00:55:10,120 --> 00:55:11,400
without human involvement

1542
00:55:11,400 --> 00:55:13,800
and that changes your culture in ways you didn't anticipate.

1543
00:55:13,800 --> 00:55:16,600
So managers have to design workflows explicitly.

1544
00:55:16,600 --> 00:55:19,560
They have to decide which stages have agent proposals

1545
00:55:19,560 --> 00:55:22,200
and which ones require human reviews or specific approvals.

1546
00:55:22,200 --> 00:55:23,480
They have to figure out

1547
00:55:23,480 --> 00:55:25,160
where agents can run in parallel

1548
00:55:25,160 --> 00:55:26,920
and where they need sequential handoffs.

1549
00:55:26,920 --> 00:55:29,480
This is workflow design, not traditional decision making.

1550
00:55:29,480 --> 00:55:31,720
It's a type of design most managers have never done

1551
00:55:31,720 --> 00:55:33,480
because their workflows were always implicit

1552
00:55:33,480 --> 00:55:35,080
and their decisions were ad hoc.

1553
00:55:35,080 --> 00:55:37,160
Work IQ forces you to be deliberate.

1554
00:55:37,160 --> 00:55:38,840
Managers also have to protect something

1555
00:55:38,840 --> 00:55:40,200
that's increasingly at risk,

1556
00:55:40,200 --> 00:55:42,760
which is focus time and psychological safety.

1557
00:55:42,760 --> 00:55:44,760
As agents take on operational work,

1558
00:55:44,760 --> 00:55:46,360
people naturally have more time.

1559
00:55:46,360 --> 00:55:48,280
But organizations often fill that time

1560
00:55:48,280 --> 00:55:50,920
with more work instead of allowing people to actually think.

1561
00:55:50,920 --> 00:55:52,360
You get faster email triage

1562
00:55:52,360 --> 00:55:54,120
just so you can attend more meetings.

1563
00:55:54,120 --> 00:55:55,480
You get automatic task routing

1564
00:55:55,480 --> 00:55:57,480
just so you can focus on more projects.

1565
00:55:57,480 --> 00:55:59,160
Work IQ creates capacity,

1566
00:55:59,160 --> 00:56:02,280
but cultural defaults tend to fill that capacity immediately.

1567
00:56:02,280 --> 00:56:04,520
Managers have to actively protect that time

1568
00:56:04,520 --> 00:56:06,760
and say a team isn't taking on another initiative

1569
00:56:06,760 --> 00:56:08,600
because they need time to think.

1570
00:56:08,600 --> 00:56:10,920
They have to create safety where people feel comfortable

1571
00:56:10,920 --> 00:56:12,360
saying they can't add more,

1572
00:56:12,360 --> 00:56:14,200
even when more becomes possible.

1573
00:56:14,200 --> 00:56:16,440
They also have to think differently about productivity.

1574
00:56:16,440 --> 00:56:18,120
It's not about how much people produce,

1575
00:56:18,120 --> 00:56:20,360
but whether the work they are producing actually matters.

1576
00:56:20,360 --> 00:56:23,160
You have to ask if decisions are moving outcomes forward

1577
00:56:23,160 --> 00:56:24,920
or just creating more busy work.

1578
00:56:24,920 --> 00:56:28,120
This requires a skill set most managers simply don't have yet.

1579
00:56:28,120 --> 00:56:30,200
Most managers are optimized for getting people

1580
00:56:30,200 --> 00:56:31,880
to complete tasks efficiently,

1581
00:56:31,880 --> 00:56:34,920
not for asking whether the tasks are worth doing in the first place.

1582
00:56:35,640 --> 00:56:38,040
Work IQ makes that question impossible to avoid

1583
00:56:38,040 --> 00:56:40,680
because the system shows you which work produces outcomes

1584
00:56:40,680 --> 00:56:41,720
and which is just noise.

1585
00:56:41,720 --> 00:56:43,880
And it requires accountability for that judgment.

1586
00:56:43,880 --> 00:56:46,280
You're not accountable for making sure people stay busy anymore.

1587
00:56:46,280 --> 00:56:48,200
You're accountable for ensuring that effort

1588
00:56:48,200 --> 00:56:49,320
translates into value.

1589
00:56:49,320 --> 00:56:50,520
That's a different kind of pressure

1590
00:56:50,520 --> 00:56:51,960
and a different kind of responsibility.

1591
00:56:51,960 --> 00:56:54,440
It requires courage because it means admitting

1592
00:56:54,440 --> 00:56:56,040
when work isn't producing results

1593
00:56:56,040 --> 00:56:58,120
that requires different relationships with your team

1594
00:56:58,120 --> 00:57:00,120
because you're not commanding compliance.

1595
00:57:00,120 --> 00:57:02,760
You're collaborating on which problems actually matter.

1596
00:57:02,760 --> 00:57:06,440
This is a completely different skill set than traditional management.

1597
00:57:06,440 --> 00:57:08,200
The organizational design shift,

1598
00:57:08,200 --> 00:57:10,120
the way we structure organizations today

1599
00:57:10,120 --> 00:57:12,360
reveals exactly how we think work should happen.

1600
00:57:12,360 --> 00:57:14,840
Those assumptions are baked into every org chart,

1601
00:57:14,840 --> 00:57:15,960
every reporting line,

1602
00:57:15,960 --> 00:57:17,800
and every career path we create.

1603
00:57:17,800 --> 00:57:20,120
Most companies organize themselves around roles,

1604
00:57:20,120 --> 00:57:21,960
grouping people into sales teams,

1605
00:57:21,960 --> 00:57:24,520
operations teams, or engineering departments.

1606
00:57:24,520 --> 00:57:28,440
We group people by their function rather than what they are actually trying to accomplish.

1607
00:57:28,440 --> 00:57:30,280
The structure tells us who owns the budget

1608
00:57:30,280 --> 00:57:31,560
and who has the authority

1609
00:57:31,560 --> 00:57:34,120
and everything else in the business flows from that rigid setup.

1610
00:57:34,120 --> 00:57:37,560
Work IQ collapses the idea that structure must follow function

1611
00:57:37,560 --> 00:57:39,560
by showing us what actually drives outcomes.

1612
00:57:39,560 --> 00:57:41,960
In reality, outcomes are driven by workflows

1613
00:57:41,960 --> 00:57:44,200
like closing deals, onboarding customers,

1614
00:57:44,200 --> 00:57:45,960
or solving service requests.

1615
00:57:45,960 --> 00:57:46,760
These are not functions,

1616
00:57:46,760 --> 00:57:48,360
but rather sequences of work

1617
00:57:48,360 --> 00:57:50,600
that cut across every department in the building.

1618
00:57:50,600 --> 00:57:53,080
Closing a deal requires sales expertise,

1619
00:57:53,080 --> 00:57:55,560
but it also needs finance to approve terms,

1620
00:57:55,560 --> 00:57:57,000
legal to review contracts,

1621
00:57:57,000 --> 00:57:59,000
and operations to confirm delivery.

1622
00:57:59,000 --> 00:57:59,800
Right now,

1623
00:57:59,800 --> 00:58:01,800
that workflows across four different managers

1624
00:58:01,800 --> 00:58:03,160
and four different budgets,

1625
00:58:03,160 --> 00:58:06,760
and the coordination overhead required to keep them aligned is enormous.

1626
00:58:06,760 --> 00:58:09,800
Work IQ makes it possible to organize around these workflows

1627
00:58:09,800 --> 00:58:10,840
instead of departments.

1628
00:58:10,840 --> 00:58:13,160
You could create a dynamic deal closure team

1629
00:58:13,160 --> 00:58:16,120
where sales owns the relationship and legal owns the contract,

1630
00:58:16,120 --> 00:58:18,120
but everyone operates as a single unit.

1631
00:58:18,120 --> 00:58:19,720
They share the same visibility,

1632
00:58:19,720 --> 00:58:20,760
the same metrics,

1633
00:58:20,760 --> 00:58:23,560
and the same ownership over how fast the deal moves.

1634
00:58:23,560 --> 00:58:24,840
Once the deal is done,

1635
00:58:24,840 --> 00:58:26,520
people return to their functional homes

1636
00:58:26,520 --> 00:58:28,600
without ever leaving their official roles.

1637
00:58:28,600 --> 00:58:30,600
They simply add a temporary membership

1638
00:58:30,600 --> 00:58:33,640
to a workflow team that actually owns the final result.

1639
00:58:33,640 --> 00:58:36,360
This approach is far more agile than a traditional structure

1640
00:58:36,360 --> 00:58:38,040
because you can staff workflows

1641
00:58:38,040 --> 00:58:40,360
based on what is actually happening in the market.

1642
00:58:40,360 --> 00:58:43,160
If you are suddenly onboarding a massive wave of customers,

1643
00:58:43,160 --> 00:58:45,720
you move resources into that specific workflow.

1644
00:58:45,720 --> 00:58:48,600
Immediately, traditional structures do not flex that easily

1645
00:58:48,600 --> 00:58:51,000
because headcount is locked into specific teams

1646
00:58:51,000 --> 00:58:53,800
and changing those lines takes months of politics.

1647
00:58:53,800 --> 00:58:56,520
Work flow-centric organizing can adapt from week to week,

1648
00:58:56,520 --> 00:58:59,320
which allows the business to breathe with the demands of its customers.

1649
00:58:59,320 --> 00:59:02,840
But we have to acknowledge that this model is also more complex to govern.

1650
00:59:02,840 --> 00:59:05,560
Authority becomes fluid when you have a functional manager in sales

1651
00:59:05,560 --> 00:59:07,960
and a separate lead for the deal closure workflow.

1652
00:59:07,960 --> 00:59:10,360
If those two people disagree on your priorities,

1653
00:59:10,360 --> 00:59:11,720
you effectively have two bosses,

1654
00:59:11,720 --> 00:59:14,520
which only works if the matrix is designed with extreme care.

1655
00:59:14,520 --> 00:59:16,360
Most organizations fail at this,

1656
00:59:16,360 --> 00:59:18,200
creating ambiguity and conflict

1657
00:59:18,200 --> 00:59:20,840
instead of the clarity they were hoping to achieve.

1658
00:59:20,840 --> 00:59:23,720
Span of control also changes in some very interesting ways.

1659
00:59:23,720 --> 00:59:26,920
A typical sales manager might oversee eight people directly,

1660
00:59:26,920 --> 00:59:30,120
but if agents are handling the coordination and tracking approvals,

1661
00:59:30,120 --> 00:59:32,120
that manager could oversee 30 people.

1662
00:59:32,120 --> 00:59:35,480
The manager is no longer stuck manually routing deals or managing handoffs

1663
00:59:35,480 --> 00:59:37,640
because the agents are handling the logistics.

1664
00:59:37,640 --> 00:59:41,320
This allows the manager to focus entirely on coaching and strategic direction,

1665
00:59:41,320 --> 00:59:43,480
which could dramatically flatten organizations

1666
00:59:43,480 --> 00:59:45,720
and remove unnecessary management layers.

1667
00:59:45,720 --> 00:59:48,600
However, this only works if decision rights are crystal clear

1668
00:59:48,600 --> 00:59:52,840
because if they stay ambiguous, adding agents just adds another layer of confusion.

1669
00:59:52,840 --> 00:59:55,240
Most organizations will resist this shift fiercely

1670
00:59:55,240 --> 00:59:59,000
because our current role-centric model creates very specific career paths.

1671
00:59:59,000 --> 01:00:01,640
You progress up a ladder from salesman to director

1672
01:00:01,640 --> 01:00:04,840
and your authority and compensation are tied directly to that title.

1673
01:00:04,840 --> 01:00:07,480
Workflow-centric organization breaks that ladder

1674
01:00:07,480 --> 01:00:09,720
and replaces it with a portfolio of skills.

1675
01:00:09,720 --> 01:00:12,360
You become known for your expertise in deal closure

1676
01:00:12,360 --> 01:00:13,960
or technical implementation,

1677
01:00:13,960 --> 01:00:16,040
moving between workflows as needed.

1678
01:00:16,040 --> 01:00:18,280
Your career becomes broader and less linear

1679
01:00:18,280 --> 01:00:21,000
and your status is defined by the problems you can solve

1680
01:00:21,000 --> 01:00:23,000
rather than how many people report to you.

1681
01:00:23,000 --> 01:00:25,640
That redistribution of status threatens power structures

1682
01:00:25,640 --> 01:00:27,480
that have been stable for decades.

1683
01:00:27,480 --> 01:00:31,000
Organizations will say they are excited about this new way of working

1684
01:00:31,000 --> 01:00:33,720
but then they will spend three years doing absolutely nothing.

1685
01:00:33,720 --> 01:00:36,600
They will deploy work IQ and see exactly which workflows matter

1686
01:00:36,600 --> 01:00:38,120
but then they will ignore the data

1687
01:00:38,120 --> 01:00:40,440
because changing the structure feels too disruptive.

1688
01:00:40,440 --> 01:00:42,200
The power stays exactly where it is

1689
01:00:42,200 --> 01:00:45,000
and this is why the next 18 months are so critical.

1690
01:00:45,000 --> 01:00:49,000
The 2026 inflection point, June 16th, 2026

1691
01:00:49,000 --> 01:00:52,440
is a date that carries a lot more weight than it might seem at first.

1692
01:00:52,440 --> 01:00:54,920
That is the day Microsoft makes the work IQ APIs

1693
01:00:54,920 --> 01:00:56,680
generally available to everyone.

1694
01:00:56,680 --> 01:00:58,920
This is not a preview or a limited test

1695
01:00:58,920 --> 01:01:02,360
but a production-ready release that is backed by service level agreements.

1696
01:01:02,360 --> 01:01:04,520
It will be available to enterprises at scale

1697
01:01:04,520 --> 01:01:08,200
and that marks the moment when work IQ stops being an optional feature.

1698
01:01:08,200 --> 01:01:09,640
It becomes the core infrastructure

1699
01:01:09,640 --> 01:01:12,920
that every serious Microsoft 365 deployment will have to address.

1700
01:01:12,920 --> 01:01:15,240
The consumption-based pricing through copilot credits

1701
01:01:15,240 --> 01:01:18,200
removes the friction that usually kills enterprise adoption.

1702
01:01:18,200 --> 01:01:19,960
You do not need to negotiate a new contract

1703
01:01:19,960 --> 01:01:22,760
or wait for a new budget cycle to start experimenting.

1704
01:01:22,760 --> 01:01:25,000
The cost simply shows up on your existing bill

1705
01:01:25,000 --> 01:01:27,400
which means organizations can increase their usage

1706
01:01:27,400 --> 01:01:29,000
without needing special approval.

1707
01:01:29,000 --> 01:01:31,240
The entire business model is engineered to make

1708
01:01:31,240 --> 01:01:33,480
"let's try it" much easier than saying "no".

1709
01:01:33,480 --> 01:01:35,800
This is the true inflection point for the industry.

1710
01:01:35,800 --> 01:01:38,200
Before June of 2026,

1711
01:01:38,200 --> 01:01:40,280
work IQ is just internal infrastructure

1712
01:01:40,280 --> 01:01:42,200
that makes copilot work a little better.

1713
01:01:42,200 --> 01:01:44,680
After that date, it becomes a decision platform

1714
01:01:44,680 --> 01:01:47,400
that developers can integrate with and customers can demand.

1715
01:01:47,400 --> 01:01:50,120
Competitors who have already built their systems around this

1716
01:01:50,120 --> 01:01:53,960
will operate much faster than those still trying to figure out a governance framework.

1717
01:01:53,960 --> 01:01:57,320
But there is a question here that matters much more than the specific date.

1718
01:01:57,320 --> 01:02:00,280
Before June 16th, companies could avoid the hard questions

1719
01:02:00,280 --> 01:02:02,760
by saying they were taking a measured approach to AI.

1720
01:02:02,760 --> 01:02:04,280
That language is no longer viable

1721
01:02:04,280 --> 01:02:07,960
because the capability now exists, whether you are ready for it or not.

1722
01:02:07,960 --> 01:02:09,160
Your competitors are building

1723
01:02:09,160 --> 01:02:11,000
and the people inside your own organization

1724
01:02:11,000 --> 01:02:12,600
are already experimenting with agents

1725
01:02:12,600 --> 01:02:14,520
that generate work that needs to be managed.

1726
01:02:14,520 --> 01:02:17,240
The question is no longer whether you should use intelligent agents

1727
01:02:17,240 --> 01:02:19,480
but how you plan to govern them at scale.

1728
01:02:19,480 --> 01:02:21,400
This has stopped being a technology problem

1729
01:02:21,400 --> 01:02:23,880
and has become a question of organizational design.

1730
01:02:23,880 --> 01:02:25,880
Those answers take a long time to get right

1731
01:02:25,880 --> 01:02:29,000
and early adopters are gaining a massive structural advantage.

1732
01:02:29,000 --> 01:02:30,920
It is not because they have better tools

1733
01:02:30,920 --> 01:02:34,040
but because they have done the hard work of cleaning their data

1734
01:02:34,040 --> 01:02:35,800
and defining their frameworks.

1735
01:02:35,800 --> 01:02:39,160
These early movers will have 18 months of operational experience

1736
01:02:39,160 --> 01:02:41,160
by the time everyone else is just getting started.

1737
01:02:41,160 --> 01:02:44,280
They will have redesigned their workflows to use agents meaningfully

1738
01:02:44,280 --> 01:02:47,640
and they will have built new roles around how these agents actually operate.

1739
01:02:47,640 --> 01:02:50,840
Their leaders will be thinking about orchestration instead of gatekeeping

1740
01:02:50,840 --> 01:02:53,560
and their measurement systems will focus on outcomes

1741
01:02:53,560 --> 01:02:55,080
rather than just activity.

1742
01:02:55,080 --> 01:02:57,880
They will have learned what fails in their specific context

1743
01:02:57,880 --> 01:03:00,440
which gives them a head start that is almost impossible to close.

1744
01:03:00,440 --> 01:03:02,360
That advantage is structural and permanent.

1745
01:03:02,360 --> 01:03:04,680
It is not about being first to market with a cool app

1746
01:03:04,680 --> 01:03:07,400
but about having redesigned your entire operating model.

1747
01:03:07,400 --> 01:03:09,400
You gain clarity on decision authority

1748
01:03:09,400 --> 01:03:11,800
and learn how agents actually change the nature of work

1749
01:03:11,800 --> 01:03:13,480
instead of just making it faster.

1750
01:03:13,480 --> 01:03:16,600
That learning app will be measured in months of execution speed

1751
01:03:16,600 --> 01:03:19,560
and years of superiority over those who started late.

1752
01:03:19,560 --> 01:03:22,280
Later adopters face a very real and dangerous choice.

1753
01:03:22,280 --> 01:03:25,720
You can choose to redesign now while you still have the luxury of time to learn

1754
01:03:25,720 --> 01:03:28,200
or you can wait until the pressure becomes unbearable.

1755
01:03:28,200 --> 01:03:32,360
History suggests most companies will wait until they see what their competitors have done

1756
01:03:32,360 --> 01:03:33,400
and then they will panic.

1757
01:03:33,400 --> 01:03:36,840
They will try to force adoption during a crisis without proper governance

1758
01:03:36,840 --> 01:03:40,600
which creates the kind of chaos that confirms everyone's worst fears about AI.

1759
01:03:40,600 --> 01:03:43,480
The technology is not the bottleneck anymore because the APIs are sound

1760
01:03:43,480 --> 01:03:44,920
and the architecture is solid.

1761
01:03:44,920 --> 01:03:47,320
The only real constraint left is organizational readiness

1762
01:03:47,320 --> 01:03:49,560
and whether you can actually change how you measure success.

1763
01:03:49,560 --> 01:03:51,560
Can you grant decision authority to an agent

1764
01:03:51,560 --> 01:03:54,280
and reorganize your teams fast enough to keep pace?

1765
01:03:54,280 --> 01:03:57,480
Most organizations cannot do this, not because they lack talent

1766
01:03:57,480 --> 01:04:00,760
but because true transformation threatens the existing power structure.

1767
01:04:00,760 --> 01:04:03,800
They will continue to deploy agents in small tactical ways

1768
01:04:03,800 --> 01:04:06,920
and eventually wonder why they aren't seeing a return on their investment

1769
01:04:06,920 --> 01:04:09,400
but some organizations will be ready for this shift

1770
01:04:09,400 --> 01:04:12,040
and those are the ones that will pull away from the rest of the pack.

1771
01:04:12,040 --> 01:04:15,000
The readiness assessment

1772
01:04:15,000 --> 01:04:18,200
If the next 18 months are going to matter, you have to answer one question first.

1773
01:04:18,200 --> 01:04:20,600
Are you actually ready to use that time productively?

1774
01:04:20,600 --> 01:04:23,800
I'm not asking if you're ready to buy a license or run a small pilot.

1775
01:04:23,800 --> 01:04:25,400
I'm asking if you're ready to transform.

1776
01:04:25,400 --> 01:04:27,000
Work IQ deployment is a mirror.

1777
01:04:27,000 --> 01:04:29,800
It reveals exactly how prepared your organization is

1778
01:04:29,800 --> 01:04:34,200
and in reality most companies will discover they aren't nearly as ready as they assumed.

1779
01:04:34,200 --> 01:04:35,800
The assessment itself is straightforward.

1780
01:04:35,800 --> 01:04:37,400
It covers four specific domains.

1781
01:04:37,400 --> 01:04:39,400
You either have clarity in these areas or you don't.

1782
01:04:39,400 --> 01:04:41,800
If you have it across all four, you can move fast.

1783
01:04:41,800 --> 01:04:43,800
If you have it in some but not others,

1784
01:04:43,800 --> 01:04:45,800
you know exactly where to focus your energy first.

1785
01:04:45,800 --> 01:04:48,600
And if you have clarity in none of them, then congratulations.

1786
01:04:48,600 --> 01:04:50,600
You just identified your actual starting point.

1787
01:04:50,600 --> 01:04:52,200
The first domain is Data Hygiene.

1788
01:04:52,200 --> 01:04:54,600
I'm not talking about your high level data strategy

1789
01:04:54,600 --> 01:04:56,200
or your complex architecture.

1790
01:04:56,200 --> 01:04:59,400
I'm talking about the actual condition of your SharePoint permissions right now.

1791
01:04:59,400 --> 01:05:01,400
Can you honestly say who has access to what?

1792
01:05:01,400 --> 01:05:03,400
You might assume people follow best practices

1793
01:05:03,400 --> 01:05:06,200
but when you actually audit the system, you find out they don't.

1794
01:05:06,200 --> 01:05:09,000
You need to know which sites are active and which are dormant.

1795
01:05:09,000 --> 01:05:11,400
You need to identify which teams are business critical

1796
01:05:11,400 --> 01:05:14,200
and which are just abandoned projects that nobody bothered to archive.

1797
01:05:14,200 --> 01:05:16,200
If your file naming conventions are a mess,

1798
01:05:16,200 --> 01:05:18,200
nobody can understand what they're looking at.

1799
01:05:18,200 --> 01:05:19,800
If your metadata tags are inconsistent,

1800
01:05:19,800 --> 01:05:21,400
the whole purpose of tagging is defeated.

1801
01:05:21,400 --> 01:05:23,400
This is the messy part of the job.

1802
01:05:23,400 --> 01:05:26,200
You'll likely find thousands of archive teams taking up space

1803
01:05:26,200 --> 01:05:29,400
and shared drives where half the company has access to sensitive contracts.

1804
01:05:29,400 --> 01:05:31,800
The second domain is governance maturity.

1805
01:05:31,800 --> 01:05:35,000
This isn't about your aspirations or the frameworks you plan to build next year.

1806
01:05:35,000 --> 01:05:38,200
It's about whether you have clear and enforceable policies right now.

1807
01:05:38,200 --> 01:05:41,400
You need to be able to articulate what an agent is allowed to do

1808
01:05:41,400 --> 01:05:43,800
and more importantly, what it isn't.

1809
01:05:43,800 --> 01:05:48,200
When an agent produces an outcome that's different from what you expected,

1810
01:05:48,200 --> 01:05:49,800
do you have a defined process for that?

1811
01:05:49,800 --> 01:05:52,800
Most organizations discover they only have fragments of a plan.

1812
01:05:52,800 --> 01:05:55,800
You might have a DLP policy over here and a retention schedule over there

1813
01:05:55,800 --> 01:05:57,400
but nothing is integrated.

1814
01:05:57,400 --> 01:06:00,200
You have governance theatre instead of a governance structure.

1815
01:06:00,200 --> 01:06:03,000
If your rules don't cover agent autonomy or memory, you aren't ready.

1816
01:06:03,000 --> 01:06:05,400
The third domain is leadership alignment.

1817
01:06:05,400 --> 01:06:07,200
This question matters more than any other.

1818
01:06:07,200 --> 01:06:10,200
Are your executives committed to redesigning how work gets done

1819
01:06:10,200 --> 01:06:13,800
or are they just hoping for a tool that makes the existing process faster?

1820
01:06:13,800 --> 01:06:16,600
If leadership just wants faster email, that's exactly what you'll get.

1821
01:06:16,600 --> 01:06:18,200
But you won't get transformation.

1822
01:06:18,200 --> 01:06:19,800
You won't get the organizational redesign

1823
01:06:19,800 --> 01:06:22,400
that actually delivers a real return on investment.

1824
01:06:22,400 --> 01:06:25,400
You'll just end up with a faster version of the same broken system.

1825
01:06:25,400 --> 01:06:28,800
Real alignment means leaders are willing to change how they measure success.

1826
01:06:28,800 --> 01:06:31,200
They have to be willing to restructure around workflows

1827
01:06:31,200 --> 01:06:33,600
and grant actual decision authority to agents.

1828
01:06:33,600 --> 01:06:36,800
If they aren't ready to measure outcomes instead of just activity,

1829
01:06:36,800 --> 01:06:39,400
work IQ won't deliver any meaningful value.

1830
01:06:39,400 --> 01:06:41,400
The fourth domain is outcome clarity.

1831
01:06:41,400 --> 01:06:44,800
You have to define what winning looks like in terms of business impact.

1832
01:06:44,800 --> 01:06:48,000
Don't talk to me about features deployed or pilots completed.

1833
01:06:48,000 --> 01:06:51,200
Tell me if decision speed is going to drop from 30 days down to 5.

1834
01:06:51,200 --> 01:06:54,800
Tell me if cross-team collaboration is going to produce better results in less time.

1835
01:06:54,800 --> 01:06:57,400
When context is reused instead of reconstructed,

1836
01:06:57,400 --> 01:06:59,800
you might save each person 2 hours every week.

1837
01:06:59,800 --> 01:07:02,600
Projects should close 2 weeks faster and rework should drop

1838
01:07:02,600 --> 01:07:04,400
because decisions are clearer from the start.

1839
01:07:04,400 --> 01:07:05,600
These are measurable outcomes.

1840
01:07:05,600 --> 01:07:09,000
If you can't articulate success in terms that matter to the business,

1841
01:07:09,000 --> 01:07:10,000
you don't have a strategy.

1842
01:07:10,000 --> 01:07:11,000
You just have hopes.

1843
01:07:11,000 --> 01:07:13,400
Most organizations are going to fail this assessment.

1844
01:07:13,400 --> 01:07:15,000
To be honest, you probably will too.

1845
01:07:15,000 --> 01:07:16,600
You'll find gaps in all 4 domains.

1846
01:07:16,600 --> 01:07:18,600
Your assumptions won't hold up under pressure.

1847
01:07:18,600 --> 01:07:21,400
You'll realize your data is messier than you thought

1848
01:07:21,400 --> 01:07:23,400
and your leadership isn't actually aligned.

1849
01:07:23,400 --> 01:07:26,200
Success will mean different things to different stakeholders

1850
01:07:26,200 --> 01:07:29,200
and your governance will look like a collection of scattered policies.

1851
01:07:29,200 --> 01:07:31,400
But that isn't a failure. It's a diagnostic.

1852
01:07:31,400 --> 01:07:33,800
It's the moment you discover exactly where you need to focus.

1853
01:07:33,800 --> 01:07:36,400
Use this assessment to prioritize your next moves.

1854
01:07:36,400 --> 01:07:38,400
Clean up the domain that is the most broken.

1855
01:07:38,400 --> 01:07:41,000
Build governance in the space that matters most.

1856
01:07:41,000 --> 01:07:44,400
Align your leadership on the outcomes that will actually unlock value.

1857
01:07:44,400 --> 01:07:48,000
Every domain you strengthen makes the next one easier to tackle.

1858
01:07:48,000 --> 01:07:50,000
The first 90 days.

1859
01:07:50,000 --> 01:07:53,000
Readiness is a diagnostic tool, but it shouldn't be paralyzing.

1860
01:07:53,000 --> 01:07:56,600
You don't have to solve every problem in all 4 domains before you can start.

1861
01:07:56,600 --> 01:08:01,400
What you need to do is pick one single workflow and commit to learning everything you can from it.

1862
01:08:01,400 --> 01:08:05,400
The first 90 days aren't about scaling work IQ across the entire enterprise.

1863
01:08:05,400 --> 01:08:09,000
They are about building your capacity to learn before you try to go wide.

1864
01:08:09,000 --> 01:08:10,800
Start with one high value workflow.

1865
01:08:10,800 --> 01:08:12,600
It doesn't have to be the biggest one you have.

1866
01:08:12,600 --> 01:08:14,600
Pick the one where improvement actually matters

1867
01:08:14,600 --> 01:08:16,400
and where the friction is easy to see.

1868
01:08:16,400 --> 01:08:19,600
This could be project execution, sales deal closure or customer onboarding.

1869
01:08:19,600 --> 01:08:22,400
It might be HR hiring or finance approval chains.

1870
01:08:22,400 --> 01:08:25,200
The specific department matters less than the characteristics of the work.

1871
01:08:25,200 --> 01:08:29,200
You want a workflow that exists across multiple teams, where speed is a factor

1872
01:08:29,200 --> 01:08:32,200
and where you can measure exactly where things are getting stuck right now.

1873
01:08:32,200 --> 01:08:34,600
Once you have the workflow, map the current state.

1874
01:08:34,600 --> 01:08:37,600
Don't map the ideal state or what the official org chart says.

1875
01:08:37,600 --> 01:08:38,600
Map the reality.

1876
01:08:38,600 --> 01:08:40,800
You need to find where decisions are actually made

1877
01:08:40,800 --> 01:08:45,200
and where information is moving through hidden email chains instead of being centralised.

1878
01:08:45,200 --> 01:08:48,600
Look for the spots where handoffs are manual when they should be automated.

1879
01:08:48,600 --> 01:08:51,400
Find the people who are waiting for approvals from other departments.

1880
01:08:51,400 --> 01:08:52,600
This is detective work.

1881
01:08:52,600 --> 01:08:55,600
You are following information through real conversations and documents

1882
01:08:55,600 --> 01:08:57,600
to see who actually influences the outcome.

1883
01:08:57,600 --> 01:08:59,600
Most organizations operate on theory,

1884
01:08:59,600 --> 01:09:02,000
but this mapping breaks that theory and shows you the truth.

1885
01:09:02,000 --> 01:09:05,600
From that map, you can see where work IQ will actually reduce friction.

1886
01:09:05,600 --> 01:09:07,000
Look at information retrieval.

1887
01:09:07,000 --> 01:09:10,800
Can agents find the right context faster than a person hunting through all documents?

1888
01:09:10,800 --> 01:09:12,200
Look at context synthesis.

1889
01:09:12,200 --> 01:09:15,400
Can agents assemble everything needed for a decision in one place?

1890
01:09:15,400 --> 01:09:16,800
Then look at decision routing.

1891
01:09:16,800 --> 01:09:20,400
Can agents get that information to the right person without a human coordinator?

1892
01:09:20,400 --> 01:09:21,800
These are your leverage points.

1893
01:09:21,800 --> 01:09:25,600
These are the places where continuous intelligence actually changes the flow of work.

1894
01:09:25,600 --> 01:09:28,000
You need to pilot this with a small engaged team.

1895
01:09:28,000 --> 01:09:29,800
Don't pick your most skeptical people,

1896
01:09:29,800 --> 01:09:31,400
but don't pick the most naive ones either.

1897
01:09:31,400 --> 01:09:35,600
Find a group that is tired of how slow things are and is actually willing to work differently.

1898
01:09:35,600 --> 01:09:38,400
A group of 20 people is much better than 200.

1899
01:09:38,400 --> 01:09:40,400
You aren't trying to measure adoption yet.

1900
01:09:40,400 --> 01:09:44,200
You are trying to learn how this technology affects this specific way of working.

1901
01:09:44,200 --> 01:09:47,800
Before you deploy a single thing, you have to establish your baseline metrics.

1902
01:09:47,800 --> 01:09:49,800
I'm not talking about how many people log in.

1903
01:09:49,800 --> 01:09:51,600
I'm talking about actual workflow numbers.

1904
01:09:51,600 --> 01:09:55,800
Measure how long it takes from the moment information enters the workflow until a decision is made.

1905
01:09:55,800 --> 01:09:59,200
Track the rework rate to see how many decisions have to be redone.

1906
01:09:59,200 --> 01:10:03,600
Count how many times people have to rehash the same information in separate conversations.

1907
01:10:03,600 --> 01:10:06,400
These numbers won't be perfect, but they are real.

1908
01:10:06,400 --> 01:10:09,400
They are the baseline you will measure against six months from now.

1909
01:10:09,400 --> 01:10:11,600
Now, deploy work IQ to support that workflow.

1910
01:10:11,600 --> 01:10:13,600
Don't try to automate the whole thing at once.

1911
01:10:13,600 --> 01:10:16,200
Pick one critical stage or a single decision point.

1912
01:10:16,200 --> 01:10:19,800
Maybe the agents handle the context assembly or the follow-up automation.

1913
01:10:19,800 --> 01:10:23,000
It needs to be something tangible that people interact with every day.

1914
01:10:23,000 --> 01:10:25,200
It should be something that either works or it doesn't.

1915
01:10:25,200 --> 01:10:30,000
This generates the signals you need to understand what happens when intelligent systems touch your processes.

1916
01:10:30,000 --> 01:10:33,200
You need to measure the results at 30, 60 and 90 days.

1917
01:10:33,200 --> 01:10:35,600
If you wait until the end, you've waited too long.

1918
01:10:35,600 --> 01:10:39,600
The 30 day mark shows you if the system is being used or if people are just working around it.

1919
01:10:39,600 --> 01:10:41,600
That tells you if you have a user education problem.

1920
01:10:41,600 --> 01:10:44,800
At 60 days, you can see if behavior is actually changing.

1921
01:10:44,800 --> 01:10:47,600
People should be using agents instead of manual steps.

1922
01:10:47,600 --> 01:10:49,800
By 90 days, you should see if the outcomes are changing.

1923
01:10:49,800 --> 01:10:52,000
Is the cycle faster? Is the rework dropping?

1924
01:10:52,000 --> 01:10:53,400
Is the team moving differently?

1925
01:10:53,400 --> 01:10:55,000
Whatever you do, don't measure adoption.

1926
01:10:55,000 --> 01:10:58,400
Adoption is a lagging indicator that only tells you if someone clicked a button.

1927
01:10:58,400 --> 01:11:00,400
Outcome change is the only thing that matters.

1928
01:11:00,400 --> 01:11:03,200
The decision cycle metric will tell you if speed increased

1929
01:11:03,200 --> 01:11:06,000
and the rework metric will tell you if quality improved.

1930
01:11:06,000 --> 01:11:08,600
These are the only numbers that justify moving forward.

1931
01:11:08,600 --> 01:11:10,600
Use these results to shape your next phase.

1932
01:11:10,600 --> 01:11:13,800
If agents helped in one stage but confused people in another,

1933
01:11:13,800 --> 01:11:16,600
you need to double down on what worked and redesign what didn't.

1934
01:11:16,600 --> 01:11:20,000
If certain types of decisions improved while others stayed the same,

1935
01:11:20,000 --> 01:11:21,200
that is your signal.

1936
01:11:21,200 --> 01:11:24,800
It tells you where the intelligence layer actually changes behavior.

1937
01:11:24,800 --> 01:11:27,400
If some people use the system while others ignored it,

1938
01:11:27,400 --> 01:11:30,200
you'll know if you have an adoption problem or an alignment problem.

1939
01:11:30,200 --> 01:11:32,600
Those first 90 days are your learning cycle.

1940
01:11:32,600 --> 01:11:35,200
You aren't trying to transform the whole company in three months.

1941
01:11:35,200 --> 01:11:39,400
You are building the organizational capacity to understand how this changes the nature of working.

1942
01:11:39,400 --> 01:11:42,400
You are building the evidence that intelligent agents can improve outcomes.

1943
01:11:42,400 --> 01:11:45,400
This is how you create a foundation for everything that comes next.

1944
01:11:45,400 --> 01:11:48,600
Work IQ isn't just another productivity tool because in reality,

1945
01:11:48,600 --> 01:11:51,400
it's a complete structural overhaul of how your organization thinks.

1946
01:11:51,400 --> 01:11:56,400
Moving from simple storage to active reasoning will fundamentally change your governance and your culture.

1947
01:11:56,400 --> 01:12:00,800
Most companies will fight this shift because it exposes every hidden dysfunction in the system.

1948
01:12:00,800 --> 01:12:04,400
But the advantage goes to the leaders who redesign their models around it right now.

1949
01:12:04,400 --> 01:12:09,800
You need to start your data cleanup today because that 2026 inflection point is coming much faster than you think.