From AI Hype to Business Value with Kayode Ajayi [MVP]

1
00:00:00,000 --> 00:00:04,960
Welcome to another edition of the N65 podcast.

2
00:00:04,960 --> 00:00:07,600
The podcast where we explore the people, technologies,

3
00:00:07,600 --> 00:00:10,200
and ideas shaping the Microsoft ecosystem.

4
00:00:10,200 --> 00:00:12,520
Today's guest is Kayyod Ajaji.

5
00:00:12,520 --> 00:00:14,800
I hope we pronounce this right.

6
00:00:14,800 --> 00:00:17,400
He is a Microsoft MVP solution architect,

7
00:00:17,400 --> 00:00:20,120
Microsoft certified trainer and a recognized expert

8
00:00:20,120 --> 00:00:22,880
in the Microsoft Power Platform, Microsoft Copilot,

9
00:00:22,880 --> 00:00:25,600
Copilot Studio and Enterprise solution architecture.

10
00:00:25,600 --> 00:00:27,080
For more than seven years,

11
00:00:27,080 --> 00:00:29,920
Kay has helped organization designing

12
00:00:29,920 --> 00:00:32,440
and implementations, scalable business solutions

13
00:00:32,440 --> 00:00:35,720
that combines local development, AI and cloud technologies.

14
00:00:35,720 --> 00:00:38,960
He's passionate about helping business innovate faster

15
00:00:38,960 --> 00:00:41,080
while maintaining the governance, security,

16
00:00:41,080 --> 00:00:43,920
and architecture, control foundations,

17
00:00:43,920 --> 00:00:45,920
requirement for enterprise success.

18
00:00:45,920 --> 00:00:49,400
Many of us know him also at the co-host

19
00:00:49,400 --> 00:00:51,360
of the Power Platform Deep Dive podcast,

20
00:00:51,360 --> 00:00:54,280
where he regularly shares practical insights, lessons learned

21
00:00:54,280 --> 00:00:56,640
from customer project and his perspective

22
00:00:56,640 --> 00:01:00,680
on the rapid evolving Microsoft ecosystem.

23
00:01:00,680 --> 00:01:02,320
In today's conversation,

24
00:01:02,320 --> 00:01:05,360
we will explore the reality of AI implementation

25
00:01:05,360 --> 00:01:09,440
where organizations succeeding, where they are struggling

26
00:01:09,440 --> 00:01:13,360
and what it really takes to move from AI experimentation

27
00:01:13,360 --> 00:01:15,640
to measurable business value.

28
00:01:15,640 --> 00:01:18,080
So welcome to the show.

29
00:01:18,080 --> 00:01:20,320
- Yes, thank you very much, I'm ocher.

30
00:01:20,320 --> 00:01:23,640
Well, that was a very wonderful introduction.

31
00:01:23,640 --> 00:01:25,480
Thank you so much for that.

32
00:01:25,480 --> 00:01:26,760
- Yeah, I'm glad to be here.

33
00:01:26,760 --> 00:01:30,320
I just see your articles, your podcasts and stuff.

34
00:01:30,320 --> 00:01:32,760
So yeah, really, really well done.

35
00:01:32,760 --> 00:01:34,720
I wish I had the energy you have.

36
00:01:34,720 --> 00:01:37,200
(laughs)

37
00:01:37,200 --> 00:01:40,760
- Yeah, can you tell a little bit about your background,

38
00:01:40,760 --> 00:01:44,000
how did you first get into technology,

39
00:01:44,000 --> 00:01:47,000
especially Microsoft ecosystem?

40
00:01:47,000 --> 00:01:48,920
- Yeah, I will say for me,

41
00:01:48,920 --> 00:01:53,800
I kind of stumbled into the whole Microsoft ecosystem.

42
00:01:53,800 --> 00:01:55,640
I think I was at the point in my life

43
00:01:55,640 --> 00:01:58,680
when I was thinking about what my career was gonna be

44
00:01:58,680 --> 00:02:01,760
'cause at that point I had done many things,

45
00:02:01,760 --> 00:02:05,920
I had done things around programming, 3D animation,

46
00:02:05,920 --> 00:02:08,480
graphic design, figure editing.

47
00:02:08,480 --> 00:02:10,760
So I tried my hands of so many things,

48
00:02:10,760 --> 00:02:13,600
but I didn't really know what my career was gonna be

49
00:02:13,600 --> 00:02:14,960
at that point then.

50
00:02:14,960 --> 00:02:18,200
One way or the other, I discovered the power platform

51
00:02:18,200 --> 00:02:21,760
and since then it's just been me going deeper

52
00:02:21,760 --> 00:02:24,200
and deeper into the Microsoft ecosystem.

53
00:02:24,200 --> 00:02:28,720
So you're falling in love with the power platform.

54
00:02:28,720 --> 00:02:30,120
- Oh yeah, 100%.

55
00:02:30,120 --> 00:02:32,400
And I think one of the things that I really loved

56
00:02:32,400 --> 00:02:35,720
was just how quickly you could do stuff with it

57
00:02:35,720 --> 00:02:39,920
and have something production and enterprise ready.

58
00:02:39,920 --> 00:02:44,240
Using local tools, before then I don't stuff

59
00:02:44,240 --> 00:02:46,960
like Python programming, Java programming,

60
00:02:46,960 --> 00:02:49,760
but there's like a lot that I would have had

61
00:02:49,760 --> 00:02:53,240
to put together install on my computer

62
00:02:53,240 --> 00:02:55,160
before I could actually get things to work

63
00:02:55,160 --> 00:02:56,680
with the power platform.

64
00:02:56,680 --> 00:03:00,520
All I just did was log into a web interface

65
00:03:00,520 --> 00:03:03,120
and I could just start creating things.

66
00:03:03,120 --> 00:03:05,840
So that just made me love the power platform.

67
00:03:05,840 --> 00:03:09,680
- And you're very active also in the community.

68
00:03:09,680 --> 00:03:13,760
You have the podcast, you're an NDP and also an MCT.

69
00:03:13,760 --> 00:03:18,760
So what motivated you were to be so active in the community?

70
00:03:19,560 --> 00:03:22,680
- Well, also for me I've always been a lover of learning

71
00:03:22,680 --> 00:03:26,960
and sharing what I know even since my university days

72
00:03:26,960 --> 00:03:31,280
have been like bringing together communities.

73
00:03:31,280 --> 00:03:35,240
There was some stuff that I did in school called Exo-Africa

74
00:03:35,240 --> 00:03:38,560
so which is just bringing together other students

75
00:03:38,560 --> 00:03:41,960
within my university teaching them about technology

76
00:03:41,960 --> 00:03:44,840
'cause by then I was learning stuff like Java, Python

77
00:03:44,840 --> 00:03:47,800
and a couple of my colleagues didn't really know so much

78
00:03:47,800 --> 00:03:50,560
even though we were all studying computer science.

79
00:03:50,560 --> 00:03:53,640
So for the most cases, even since then

80
00:03:53,640 --> 00:03:56,320
I've started bringing communities together.

81
00:03:56,320 --> 00:04:00,360
So coming to the UK, I think initially I was very quiet

82
00:04:00,360 --> 00:04:03,240
because I was doing my master's degree.

83
00:04:03,240 --> 00:04:05,320
So during that time when I didn't have time

84
00:04:05,320 --> 00:04:08,880
for any community stuff 'cause during the master's degree

85
00:04:08,880 --> 00:04:11,520
it's not easy and interestingly it actually came

86
00:04:11,520 --> 00:04:14,320
out with a distinction so it was worth it.

87
00:04:14,320 --> 00:04:18,640
But after then I started getting more active in the community

88
00:04:18,640 --> 00:04:22,520
attending user groups and I finally internet from myself

89
00:04:22,520 --> 00:04:25,280
speaking at some of this user groups and yeah,

90
00:04:25,280 --> 00:04:28,000
I just tried to get involved at the end events

91
00:04:28,000 --> 00:04:30,080
and show why new basically.

92
00:04:30,080 --> 00:04:35,680
- Yeah, I think that was really nice career I think.

93
00:04:35,680 --> 00:04:38,920
You're so deep in the power platform

94
00:04:38,920 --> 00:04:41,880
and I think most people still see the power platform

95
00:04:41,880 --> 00:04:44,000
as a citizen develop a platform.

96
00:04:44,000 --> 00:04:46,440
How will doing to describe it today?

97
00:04:46,440 --> 00:04:51,880
- So that's very interesting 'cause especially even when you see

98
00:04:51,880 --> 00:04:54,840
Microsoft messaging about the power platform

99
00:04:54,840 --> 00:04:58,720
is kind of geared towards that citizen development mindset

100
00:04:58,720 --> 00:05:02,920
and then when you now see, well, what's your career then?

101
00:05:02,920 --> 00:05:06,440
Basically I mean, look who to develop and I think

102
00:05:06,440 --> 00:05:09,920
I love being a look who to develop because as I said

103
00:05:09,920 --> 00:05:11,440
when you're doing through goods stuff

104
00:05:11,440 --> 00:05:13,800
there's a lot of extra admin

105
00:05:13,800 --> 00:05:16,440
that you need to do things you have to install

106
00:05:16,440 --> 00:05:20,960
and stuff like that but with a local development platform

107
00:05:20,960 --> 00:05:24,360
you're kind of building based on something

108
00:05:24,360 --> 00:05:25,800
that's been established.

109
00:05:25,800 --> 00:05:29,640
But I've been able to exist in this career

110
00:05:29,640 --> 00:05:32,160
all this well because at the end of the day

111
00:05:32,160 --> 00:05:34,840
we still build enterprise solutions

112
00:05:34,840 --> 00:05:37,400
and one of the messaging about local is

113
00:05:37,400 --> 00:05:42,400
you can experiment quickly, you can go to markets very quickly

114
00:05:43,320 --> 00:05:44,960
then you would have been able to do

115
00:05:44,960 --> 00:05:47,160
with traditional development.

116
00:05:47,160 --> 00:05:51,600
So yeah, there's a lot of value in the power platform

117
00:05:51,600 --> 00:05:54,440
and as much as it's geared towards like

118
00:05:54,440 --> 00:05:56,440
that citizen develop a mindset,

119
00:05:56,440 --> 00:06:01,800
we see that we actually have organizations

120
00:06:01,800 --> 00:06:04,280
using this for their enterprise solutions.

121
00:06:04,280 --> 00:06:08,240
So part of why do we organizations is in some cases

122
00:06:08,240 --> 00:06:10,640
it's citizen develop a within the organization

123
00:06:10,640 --> 00:06:14,680
that's put together a solution as a proof of concept

124
00:06:14,680 --> 00:06:18,760
then we now step in as experts to get its productionized.

125
00:06:18,760 --> 00:06:20,920
So even we did a local space,

126
00:06:20,920 --> 00:06:23,560
there's still a citizen developer space

127
00:06:23,560 --> 00:06:28,560
which is less technical and there is the expert space

128
00:06:28,560 --> 00:06:30,280
which is where I sit in.

129
00:06:30,280 --> 00:06:32,000
What do you say?

130
00:06:32,000 --> 00:06:37,000
It's also power platforms also ready for enterprise solutions.

131
00:06:37,000 --> 00:06:39,000
Oh yes, it's always been,

132
00:06:40,000 --> 00:06:41,720
I was having a conversation with my colleague

133
00:06:41,720 --> 00:06:43,280
on our podcast very recently

134
00:06:43,280 --> 00:06:45,760
and she was talking with someone and they said,

135
00:06:45,760 --> 00:06:50,280
well, you know, local developers actually don't design

136
00:06:50,280 --> 00:06:53,560
for scalability and well,

137
00:06:53,560 --> 00:06:55,440
the person explained what it means to her

138
00:06:55,440 --> 00:06:58,560
and what they were just saying is because of the way

139
00:06:58,560 --> 00:07:00,720
a lot of people then look at,

140
00:07:00,720 --> 00:07:03,200
they don't lend it with that enterprise mindset

141
00:07:03,200 --> 00:07:05,560
and I would say have been guilty of that through

142
00:07:05,560 --> 00:07:08,040
from the onset and when starting

143
00:07:08,040 --> 00:07:09,960
we local down power platforms,

144
00:07:09,960 --> 00:07:13,640
the solutions that abuse and after a few months

145
00:07:13,640 --> 00:07:17,440
I realized oh, that was actually a bad architectural decision

146
00:07:17,440 --> 00:07:21,800
and adding some updates will now lead like major redesign

147
00:07:21,800 --> 00:07:25,320
because at that point I didn't design for scalability

148
00:07:25,320 --> 00:07:30,080
and enterprise but now I build solutions that are supported,

149
00:07:30,080 --> 00:07:32,040
they are used across the world

150
00:07:32,040 --> 00:07:34,600
that actually build platforms on the power platforms

151
00:07:34,600 --> 00:07:38,080
that all that developers be take and they use it

152
00:07:38,080 --> 00:07:40,080
to extend their solutions.

153
00:07:40,080 --> 00:07:45,080
So it's about the mindset and the training basically.

154
00:07:45,080 --> 00:07:47,840
If you do things shabily well then your solution

155
00:07:47,840 --> 00:07:49,800
is not going to be enterprise ready

156
00:07:49,800 --> 00:07:51,360
but if you do this properly,

157
00:07:51,360 --> 00:07:53,720
we good architectural principles

158
00:07:53,720 --> 00:07:56,680
then you would have an enterprise ready solution.

159
00:07:56,680 --> 00:07:58,320
So you could have the same tools

160
00:07:58,320 --> 00:08:00,680
but use them very differently.

161
00:08:02,200 --> 00:08:07,200
And what the thing you were missing about the companies

162
00:08:07,200 --> 00:08:09,680
what is the biggest misconceptions

163
00:08:09,680 --> 00:08:13,080
executives have about the local platforms?

164
00:08:13,080 --> 00:08:19,760
I think one thing because a lot of messaging

165
00:08:19,760 --> 00:08:24,160
about local gains towards speed of execution

166
00:08:24,160 --> 00:08:28,520
and then they just want everything immediately, right?

167
00:08:28,520 --> 00:08:32,600
And that causes some issues down the line because

168
00:08:32,600 --> 00:08:35,320
that means you do not do a proper discovery,

169
00:08:35,320 --> 00:08:37,800
you do not plan out your solution, right?

170
00:08:37,800 --> 00:08:40,080
You do not do a proper project plan

171
00:08:40,080 --> 00:08:45,080
and we local is very easy to just keep all of those extra things

172
00:08:45,080 --> 00:08:47,880
and those jump to the solution

173
00:08:47,880 --> 00:08:50,880
because as I said, you just need to go on the browser

174
00:08:50,880 --> 00:08:53,000
and you just start working.

175
00:08:53,000 --> 00:08:56,080
So I feel like that's the common misconception

176
00:08:56,080 --> 00:08:59,320
that you can just open up a blank app

177
00:08:59,320 --> 00:09:04,320
and just stop using without doing a proper solution design

178
00:09:04,320 --> 00:09:07,080
without doing proper discovery.

179
00:09:07,080 --> 00:09:08,920
If you find yourself doing that

180
00:09:08,920 --> 00:09:11,080
you probably have issues down the line.

181
00:09:11,080 --> 00:09:16,840
And what did you think about an in the power platform?

182
00:09:16,840 --> 00:09:20,880
How good is for governance, security and risk topics?

183
00:09:20,880 --> 00:09:25,480
- Yeah, so everything is evolving, right?

184
00:09:25,480 --> 00:09:29,560
And I think that's one of the reasons why I love being in the

185
00:09:29,560 --> 00:09:34,120
Microsoft space and in the power platform space in particular

186
00:09:34,120 --> 00:09:36,120
because everything evolves.

187
00:09:36,120 --> 00:09:38,200
I'll say a couple of years ago,

188
00:09:38,200 --> 00:09:40,520
there were not so much governance

189
00:09:40,520 --> 00:09:41,880
and administrative features

190
00:09:41,880 --> 00:09:45,120
but that always keeps on evolving.

191
00:09:45,120 --> 00:09:47,960
Now we have more tools.

192
00:09:47,960 --> 00:09:51,040
We've also have community developed tools

193
00:09:51,040 --> 00:09:56,840
that kind of extend what Microsoft offers out of the box

194
00:09:56,840 --> 00:10:00,920
and Microsoft as we've seen as even in the recent months

195
00:10:00,920 --> 00:10:04,320
and recent years, they're also trying to meet up

196
00:10:04,320 --> 00:10:06,080
with those community developed tools

197
00:10:06,080 --> 00:10:09,640
so that there's like a quality tool

198
00:10:09,640 --> 00:10:13,520
within the Microsoft and power platform space for the governance.

199
00:10:13,520 --> 00:10:17,480
So I'll say sufficient and any areas

200
00:10:17,480 --> 00:10:20,760
that Microsoft doesn't provide out of the box

201
00:10:20,760 --> 00:10:24,080
we do have opportunities to extend

202
00:10:24,080 --> 00:10:25,560
because at the end of the day,

203
00:10:25,560 --> 00:10:27,520
some of the tools are APIs

204
00:10:27,520 --> 00:10:30,600
and you can actually build tools around those APIs

205
00:10:30,600 --> 00:10:35,160
and CLIs to get that custom functionality you want.

206
00:10:35,160 --> 00:10:37,160
So out of the box is good

207
00:10:37,160 --> 00:10:39,520
but you can always still build upon it.

208
00:10:39,520 --> 00:10:41,920
And while being an MVP also,

209
00:10:41,920 --> 00:10:45,600
we do have ways of giving feedback to the product groups

210
00:10:45,600 --> 00:10:50,280
and to on some of the actual features that we find

211
00:10:50,280 --> 00:10:54,680
are being required in the real world, basically.

212
00:10:54,680 --> 00:10:59,560
And I think when we look at LinkedIn

213
00:10:59,560 --> 00:11:04,560
and we read about especially the topic governance

214
00:11:04,560 --> 00:11:07,880
and power platform that one buzzword,

215
00:11:07,880 --> 00:11:11,280
I would say is a COE as a set of excellence.

216
00:11:11,280 --> 00:11:16,360
What role do they play today?

217
00:11:17,840 --> 00:11:20,840
So the COE is very important

218
00:11:20,840 --> 00:11:25,840
and as I said, if the series really about ensuring

219
00:11:25,840 --> 00:11:29,680
that your solutions are very well supported

220
00:11:29,680 --> 00:11:33,160
and you know, you have your proper A&M

221
00:11:33,160 --> 00:11:35,880
and talking about COE as a concept

222
00:11:35,880 --> 00:11:38,200
is not just about your tools.

223
00:11:38,200 --> 00:11:40,080
So your tools, the tools you have,

224
00:11:40,080 --> 00:11:41,920
I involved in your CERIBLE,

225
00:11:41,920 --> 00:11:44,680
more important than the tools is actually the people.

226
00:11:44,680 --> 00:11:47,320
So how are you actually supporting your cities

227
00:11:47,320 --> 00:11:49,080
and developers?

228
00:11:49,080 --> 00:11:51,480
How are you supporting your actual makers,

229
00:11:51,480 --> 00:11:53,720
the people built into solutions?

230
00:11:53,720 --> 00:11:58,040
How are you actually admitting and governing the platform?

231
00:11:58,040 --> 00:12:02,600
So all of those things come together to become the CERI.

232
00:12:02,600 --> 00:12:05,160
So yeah, a lot of times we just think of CERI

233
00:12:05,160 --> 00:12:07,480
and we're thinking of tools and dashboards

234
00:12:07,480 --> 00:12:10,600
and maybe automation's boys beyond that

235
00:12:10,600 --> 00:12:13,600
is really how you support your makers,

236
00:12:13,600 --> 00:12:18,600
how you support the platform itself and your solutions.

237
00:12:18,600 --> 00:12:20,920
- Yeah, and what's your experience,

238
00:12:20,920 --> 00:12:26,080
how to bring in the makers into the platform

239
00:12:26,080 --> 00:12:29,280
and yeah, or make them successful?

240
00:12:29,280 --> 00:12:31,120
What's your tips here?

241
00:12:31,120 --> 00:12:35,720
- Yeah, yeah, and that's, on an organization

242
00:12:35,720 --> 00:12:37,840
we do have engagements around CERI,

243
00:12:37,840 --> 00:12:40,720
specifically CERI and governance

244
00:12:40,720 --> 00:12:45,200
and some of the things that we actually do is,

245
00:12:45,200 --> 00:12:49,640
in terms of, we don't want an IT system

246
00:12:49,640 --> 00:12:52,000
that just shuts everything down, right?

247
00:12:52,000 --> 00:12:56,200
And I do call that the old IT mindset

248
00:12:56,200 --> 00:12:59,480
because with the old IT mindset,

249
00:12:59,480 --> 00:13:01,440
they just shut everything down and believe,

250
00:13:01,440 --> 00:13:03,400
well, yes, now we're secure.

251
00:13:03,400 --> 00:13:05,280
But when you shut everything down,

252
00:13:05,280 --> 00:13:08,080
you realize that people are actually creative

253
00:13:08,080 --> 00:13:10,960
and in some cases they will find other ways

254
00:13:10,960 --> 00:13:15,160
to do some things that would actually end up

255
00:13:15,160 --> 00:13:17,600
landing in the shadow IT world.

256
00:13:17,600 --> 00:13:19,240
So rather than having that,

257
00:13:19,240 --> 00:13:21,880
you want to actually support your makers

258
00:13:21,880 --> 00:13:25,520
so that you can help them to do the right things.

259
00:13:25,520 --> 00:13:27,600
So for example, in the power plants,

260
00:13:27,600 --> 00:13:32,600
from recently new concepts called managed environments,

261
00:13:32,600 --> 00:13:34,320
Alaskan managed environments,

262
00:13:34,320 --> 00:13:37,520
when somebody creates a resource for the first time,

263
00:13:37,520 --> 00:13:40,080
maybe creates an app for the first time,

264
00:13:40,080 --> 00:13:43,640
there's like an automation that can send them

265
00:13:43,640 --> 00:13:45,880
and onboarding email and say,

266
00:13:45,880 --> 00:13:50,120
oh, welcome to the power plants from space.

267
00:13:50,120 --> 00:13:52,640
This is how we want you to do things.

268
00:13:52,640 --> 00:13:56,040
These are the connectors that are available for you to use.

269
00:13:56,040 --> 00:13:59,560
So in that case, you're kind of onboarding those users

270
00:13:59,560 --> 00:14:02,240
and you are directing them in the right ways.

271
00:14:02,240 --> 00:14:07,040
So that's an example of how you can kind of help

272
00:14:07,040 --> 00:14:09,080
those makers.

273
00:14:09,080 --> 00:14:11,800
We also have things like deal-cute policies.

274
00:14:11,800 --> 00:14:17,720
Like SIF, for example, you do not want a user to create

275
00:14:17,720 --> 00:14:21,000
an automation that takes all your data

276
00:14:21,000 --> 00:14:23,040
and exports it to Twitter.

277
00:14:23,040 --> 00:14:27,160
And that's an example I used to do because it's very easy

278
00:14:27,160 --> 00:14:28,720
if you have the Twitter connector,

279
00:14:28,720 --> 00:14:31,040
you could just give that automation expert data

280
00:14:31,040 --> 00:14:34,240
from data that's sending to Twitter every morning.

281
00:14:34,240 --> 00:14:37,760
But you don't want your users to be able to do that.

282
00:14:37,760 --> 00:14:40,480
So you can actually call deal-cute policies

283
00:14:40,480 --> 00:14:45,040
that define for every environment what should be possible,

284
00:14:45,040 --> 00:14:46,880
what users should be able to do.

285
00:14:46,880 --> 00:14:48,200
So as part of that onboarding,

286
00:14:48,200 --> 00:14:50,760
you can always give them that information.

287
00:14:50,760 --> 00:14:55,760
But you can get the platform itself to do some governance.

288
00:14:55,760 --> 00:14:58,480
That's an example of deal-cute policies

289
00:14:58,480 --> 00:15:01,000
that get other tools within the power platform

290
00:15:01,000 --> 00:15:04,480
that can allow the platform to kind of govern itself

291
00:15:04,480 --> 00:15:06,640
and restrict what people are able to do.

292
00:15:06,640 --> 00:15:10,120
So you can still have them work safely, right?

293
00:15:10,120 --> 00:15:12,160
But yes, don't lock it down.

294
00:15:12,160 --> 00:15:14,560
But what you want to try to do is you want to give them

295
00:15:14,560 --> 00:15:17,720
a safe space that they can explore and try things out.

296
00:15:17,720 --> 00:15:21,200
Because I say this on almost all my engagements.

297
00:15:21,200 --> 00:15:23,520
And because I've actually found it to be true

298
00:15:23,520 --> 00:15:27,160
that a lot of great ideas come from the business users

299
00:15:27,160 --> 00:15:31,440
because they are the one facing those challenges, right?

300
00:15:31,440 --> 00:15:34,720
And if you give them the tools to actually build some things,

301
00:15:34,720 --> 00:15:38,440
that can be a good level zero, right?

302
00:15:38,440 --> 00:15:41,600
It could be a good spring zero in the sense that

303
00:15:41,600 --> 00:15:46,320
they give you a foundation that you can outtake on an IT

304
00:15:46,320 --> 00:15:47,920
and make it production-ready.

305
00:15:47,920 --> 00:15:53,360
I think in the power platform, the sets range,

306
00:15:53,360 --> 00:16:01,120
yeah, in the last years, especially since co-pilot, I think,

307
00:16:01,120 --> 00:16:04,560
and co-pilot, we were in the market, I could still.

308
00:16:04,560 --> 00:16:10,280
What will you say, what is the biggest change

309
00:16:10,280 --> 00:16:12,680
what AI brings to the power platform?

310
00:16:12,680 --> 00:16:17,000
So, yeah, this is very began.

311
00:16:17,000 --> 00:16:21,080
A lot of my team are developments in we do have

312
00:16:21,080 --> 00:16:23,760
lots of conversations around this a lot.

313
00:16:23,760 --> 00:16:29,920
Because one thing that I myself have seen to change is

314
00:16:29,920 --> 00:16:33,400
there are loads of conversations about AI now

315
00:16:33,400 --> 00:16:36,320
much more than it used to be before.

316
00:16:36,320 --> 00:16:38,680
Before in the power platform, space is really about

317
00:16:38,680 --> 00:16:41,720
apps and automation where we had peripheral agents,

318
00:16:41,720 --> 00:16:44,360
but not many clients were engaging with those

319
00:16:44,360 --> 00:16:46,320
weak peripheral agents.

320
00:16:46,320 --> 00:16:48,320
But now, if there is like a curve,

321
00:16:48,320 --> 00:16:52,120
I think it's kind of turned over because a lot of the

322
00:16:52,120 --> 00:16:55,760
conversations we are having now is AI first,

323
00:16:55,760 --> 00:16:59,880
then as part of the AI, then we build apps and automations

324
00:16:59,880 --> 00:17:01,120
to support it.

325
00:17:01,120 --> 00:17:05,320
So, that language is kind of changing where a lot is

326
00:17:05,320 --> 00:17:07,120
having to do with AI.

327
00:17:07,120 --> 00:17:09,560
And even from Microsoft, we do see that,

328
00:17:09,560 --> 00:17:11,440
well, a lot of the new features on the power

329
00:17:11,440 --> 00:17:14,000
platform are actually AI focused, right?

330
00:17:14,000 --> 00:17:20,320
So, the conversation is kind of shifted a lot to AI first.

331
00:17:20,320 --> 00:17:23,080
And Microsoft even comes and says, well,

332
00:17:23,080 --> 00:17:25,360
Copa is the UI for AI.

333
00:17:25,360 --> 00:17:30,480
And they are actually building tools to enable users

334
00:17:30,480 --> 00:17:36,320
have AI as their first way of getting work done.

335
00:17:36,320 --> 00:17:39,840
So, you start with AI, then it branches into other tools

336
00:17:39,840 --> 00:17:41,120
when necessary.

337
00:17:41,120 --> 00:17:44,520
So, yeah, I feel like that's the major shift on the power

338
00:17:44,520 --> 00:17:46,120
platform now.

339
00:17:46,120 --> 00:17:48,360
You are also an expert in Copa.

340
00:17:48,360 --> 00:17:50,560
It's studio.

341
00:17:50,560 --> 00:17:54,200
And I think a lot of people think the Copa.

342
00:17:54,200 --> 00:17:56,680
It's studio is the new, I don't know, power,

343
00:17:56,680 --> 00:17:59,280
pages, power, automate.

344
00:17:59,280 --> 00:18:02,800
What think you different power or the Copa.

345
00:18:02,800 --> 00:18:06,040
It's studio from the power platform?

346
00:18:06,040 --> 00:18:06,880
Well, the Copa.

347
00:18:06,880 --> 00:18:09,520
The studio is part of the power platform,

348
00:18:09,520 --> 00:18:10,880
anyways.

349
00:18:10,880 --> 00:18:15,800
So, it's not like different because the power platform

350
00:18:15,800 --> 00:18:18,160
is like a sort of application.

351
00:18:18,160 --> 00:18:21,680
So, we have power, we have power to make what we use to have

352
00:18:21,680 --> 00:18:24,600
Power BI, Power BI technically.

353
00:18:24,600 --> 00:18:27,200
It's still talked about in the context of Power

354
00:18:27,200 --> 00:18:28,280
Plats with Power BI.

355
00:18:28,280 --> 00:18:30,480
I'll see right now is a different project,

356
00:18:30,480 --> 00:18:34,400
product on the zone, especially with Fabric in the picture.

357
00:18:34,400 --> 00:18:37,040
Because Power BI is more like a Fabric thing now.

358
00:18:37,040 --> 00:18:40,160
But if you look at documentation, you may probably still

359
00:18:40,160 --> 00:18:42,840
see part of the area as being part of the power platform.

360
00:18:42,840 --> 00:18:46,320
Then we also have Copa pilot, Copa pilot studio.

361
00:18:46,320 --> 00:18:50,960
So, it's one of the tools and a lot of solutions, especially

362
00:18:50,960 --> 00:18:53,480
when you're talking of enterprise solutions,

363
00:18:53,480 --> 00:18:56,840
they do not just sit on one of those two.

364
00:18:56,840 --> 00:18:59,840
So, you will hardly build an app that does

365
00:18:59,840 --> 00:19:01,720
it require an automation.

366
00:19:01,720 --> 00:19:06,040
You would actually build a Copa pilot studio, agents

367
00:19:06,040 --> 00:19:10,360
that doesn't have maybe like an app interface somewhere,

368
00:19:10,360 --> 00:19:13,600
maybe for data entry or for viewing bulk data.

369
00:19:13,600 --> 00:19:16,200
So, all of these things, they come together

370
00:19:16,200 --> 00:19:20,840
and that's a whole idea of solution architecture anyway.

371
00:19:20,840 --> 00:19:23,160
Where you're looking at all the tools are available

372
00:19:23,160 --> 00:19:26,920
and you're deciding what tools are actually needed

373
00:19:26,920 --> 00:19:29,720
to solve this business problem.

374
00:19:29,720 --> 00:19:33,080
So, the Copa pilot studio is not outside the power platform.

375
00:19:33,080 --> 00:19:38,440
It's just one of the tools are available within the power platform.

376
00:19:38,440 --> 00:19:40,640
OK.

377
00:19:40,640 --> 00:19:46,400
So, we will a little bit talk about the AI to business value.

378
00:19:46,400 --> 00:19:53,760
So, can you say what is AI hype actually

379
00:19:53,760 --> 00:19:58,480
and what delivered real value with Copa pilot AI?

380
00:19:58,480 --> 00:19:59,480
Yeah.

381
00:19:59,480 --> 00:20:02,400
So, yeah, the whole idea of hype, right?

382
00:20:02,400 --> 00:20:07,480
And I think even this morning, I was still listening to

383
00:20:07,480 --> 00:20:11,800
a recording by someone and they were talking about how

384
00:20:11,800 --> 00:20:15,440
well is the AI bubble actually going to burst anytime soon?

385
00:20:15,440 --> 00:20:20,040
And there is always this notion about hype around AI.

386
00:20:20,040 --> 00:20:25,120
And when I hear AI hype, it kind of reminds me of the dotcom bubble

387
00:20:25,120 --> 00:20:29,240
and in a lot of ways people kind of relate this old AI

388
00:20:29,240 --> 00:20:31,200
to the dotcom bubble.

389
00:20:31,200 --> 00:20:33,880
But I was actually thinking of it before this call.

390
00:20:33,880 --> 00:20:40,240
And as much as the dotcom bubble ended up bursting right,

391
00:20:40,240 --> 00:20:41,600
the fact still remains.

392
00:20:41,600 --> 00:20:44,240
It was a technology revolution.

393
00:20:44,240 --> 00:20:48,680
And before the dotcom era, well, you could probably only browse

394
00:20:48,680 --> 00:20:51,080
the internet through your university.

395
00:20:51,080 --> 00:20:54,960
And even that was very novel and not everybody had access to it.

396
00:20:54,960 --> 00:20:59,800
But the dotcom bubble brought access to the internet

397
00:20:59,800 --> 00:21:04,680
to everyone and everybody could start getting on the kind

398
00:21:04,680 --> 00:21:07,080
of bandwagon of the dotcom.

399
00:21:07,080 --> 00:21:12,720
So, if you look at AI hype, I'm sure, yes, very possibly,

400
00:21:12,720 --> 00:21:15,160
there is a hype somewhere in the sense that,

401
00:21:15,160 --> 00:21:17,640
but the hype really comes a lot from the business

402
00:21:17,640 --> 00:21:19,560
and investment side of things.

403
00:21:19,560 --> 00:21:22,120
Same thing we saw with the dotcom bubble.

404
00:21:22,120 --> 00:21:25,200
Technically, it was amazing.

405
00:21:25,200 --> 00:21:26,440
It was a breakthrough.

406
00:21:26,440 --> 00:21:30,840
But, yes, some people kind of wrote on that breakthrough

407
00:21:30,840 --> 00:21:34,680
and built on sustainable business models on it

408
00:21:34,680 --> 00:21:37,360
and those ended up crushing.

409
00:21:37,360 --> 00:21:41,240
So, what I'm really trying to say is, yes, there's hype

410
00:21:41,240 --> 00:21:44,400
from the business side of things.

411
00:21:44,400 --> 00:21:48,120
Because, well, everybody now wants AI.

412
00:21:48,120 --> 00:21:50,320
Everybody wants an AI company now.

413
00:21:50,320 --> 00:21:54,240
And at some point, I think in the 2020s, we had an era

414
00:21:54,240 --> 00:21:57,480
where if your AI is in your company,

415
00:21:57,480 --> 00:21:59,880
you just have your investor's dream money at you.

416
00:21:59,880 --> 00:22:04,120
But, I don't want that to take away the facts

417
00:22:04,120 --> 00:22:07,640
that were actually in a revolution.

418
00:22:07,640 --> 00:22:11,000
And, interestingly, between 2020 and today,

419
00:22:11,000 --> 00:22:14,760
there's been so much innovation around AI.

420
00:22:14,760 --> 00:22:17,960
So, let's not take away that fact

421
00:22:17,960 --> 00:22:20,560
and just call it a bubble and end it there.

422
00:22:20,560 --> 00:22:22,840
Well, there might be some bits of bubbles

423
00:22:22,840 --> 00:22:26,720
around investments going around money exchange in hands.

424
00:22:26,720 --> 00:22:29,520
But, the fact that it's still a technology revolution.

425
00:22:29,520 --> 00:22:33,840
And, we are, it now comes into play me being an architect

426
00:22:33,840 --> 00:22:36,600
in a consultancy.

427
00:22:36,600 --> 00:22:39,360
I just want to add value to organizations, right?

428
00:22:39,360 --> 00:22:41,880
I want to take AI as a teach today

429
00:22:41,880 --> 00:22:46,680
and give them business value and deliver real projects with it.

430
00:22:46,680 --> 00:22:50,160
And, that's what we've been doing even since, you know,

431
00:22:50,160 --> 00:22:53,280
co-pilot, that a COVID-19 co-pilot studio,

432
00:22:53,280 --> 00:22:57,400
where, pervertual events got rebranded as co-pilot studio.

433
00:22:57,400 --> 00:23:00,040
We've started engaging with clients

434
00:23:00,040 --> 00:23:02,600
and we've started delivering value around that.

435
00:23:02,600 --> 00:23:05,600
It's just about understanding what the current limitations

436
00:23:05,600 --> 00:23:09,400
of the technology are and copy things properly.

437
00:23:09,400 --> 00:23:16,160
Sorry, and, scoping things properly to fit within those limitations

438
00:23:16,160 --> 00:23:21,160
if you do that you can actually have a successful implementation.

439
00:23:21,160 --> 00:23:27,160
And, and what are, I say, metrics or KPIs,

440
00:23:27,160 --> 00:23:33,160
you showed or company showed, track if this is a successful

441
00:23:33,160 --> 00:23:37,160
or a good AI investment.

442
00:23:37,160 --> 00:23:44,160
So, one metric that is easy and to track is,

443
00:23:44,160 --> 00:23:48,480
when you're looking at things like your ROI or return of investment,

444
00:23:48,480 --> 00:23:54,480
a lot of times one of the very obvious ones is time-saving.

445
00:23:54,480 --> 00:23:58,720
And, you know, if you implement an AI solution

446
00:23:58,720 --> 00:24:02,720
and people are actually spending more time,

447
00:24:02,720 --> 00:24:06,480
doing the same things they were doing without AI

448
00:24:06,480 --> 00:24:08,960
and now they take it longer to do it there,

449
00:24:08,960 --> 00:24:11,480
that's automatically a failure.

450
00:24:11,480 --> 00:24:15,000
And, I think in the early days of all these AI rollouts,

451
00:24:15,000 --> 00:24:17,920
there was a research that I heard about

452
00:24:17,920 --> 00:24:20,320
and I kind of read about too,

453
00:24:20,320 --> 00:24:25,400
where the research that said, well, employees have nine to five.

454
00:24:25,400 --> 00:24:27,360
Now we've given them AI.

455
00:24:27,360 --> 00:24:30,080
Do they actually save,

456
00:24:30,080 --> 00:24:34,120
do they actually, can they stop working by four o'clock?

457
00:24:34,120 --> 00:24:36,720
And, that was a very interesting research

458
00:24:36,720 --> 00:24:39,560
because what they actually found out was,

459
00:24:39,560 --> 00:24:42,680
well, employees still works nine to five,

460
00:24:42,680 --> 00:24:48,160
but the key difference was they could actually do more

461
00:24:48,160 --> 00:24:50,280
within that nine to five hours.

462
00:24:50,280 --> 00:24:52,120
So, I feel like as a human race,

463
00:24:52,120 --> 00:24:54,040
we will always have problems to solve,

464
00:24:54,040 --> 00:24:55,920
we will always have things to do.

465
00:24:55,920 --> 00:25:00,480
If we have tools that solve our current problems,

466
00:25:00,480 --> 00:25:02,600
we don't necessarily get that time back

467
00:25:02,600 --> 00:25:04,320
and just go and hold it in.

468
00:25:04,320 --> 00:25:07,400
We will really hope we all have problems.

469
00:25:07,400 --> 00:25:10,360
So, I feel like that was very interesting.

470
00:25:10,360 --> 00:25:12,280
So, both time saving,

471
00:25:12,280 --> 00:25:15,120
the fact that your employees are able to do more

472
00:25:15,120 --> 00:25:19,720
within the time frame they used to originally do things.

473
00:25:19,720 --> 00:25:22,640
I feel like that's one of the easiest measures.

474
00:25:22,640 --> 00:25:25,480
Some other measures to do things like adoption,

475
00:25:25,480 --> 00:25:28,080
which might not be so easy to track,

476
00:25:28,080 --> 00:25:30,640
but if you have an efficient way of tracking it,

477
00:25:30,640 --> 00:25:35,880
you can also tell stories about how people actually using AI

478
00:25:35,880 --> 00:25:39,120
on our podcast, there was a time we talked about AI

479
00:25:39,120 --> 00:25:43,040
and in my colleague talked about a friend

480
00:25:43,040 --> 00:25:45,800
who, well, she had access to co-pilot,

481
00:25:45,800 --> 00:25:47,280
but she hadn't been doing it.

482
00:25:47,280 --> 00:25:50,600
So, she said after she listened to that stuff on our podcast,

483
00:25:50,600 --> 00:25:53,080
then she started doing more recuperations,

484
00:25:53,080 --> 00:25:55,720
like, oh, wow, I can actually do all this.

485
00:25:55,720 --> 00:25:59,080
So, there's that adoption piece where we could have given

486
00:25:59,080 --> 00:26:02,320
people access to AI and AI agents and all that,

487
00:26:02,320 --> 00:26:04,560
but how are they actually using it?

488
00:26:04,560 --> 00:26:06,840
How are they using it effectively?

489
00:26:06,840 --> 00:26:10,680
So, I could build a tool that saves your time,

490
00:26:10,680 --> 00:26:12,320
but if you don't use it, well,

491
00:26:12,320 --> 00:26:14,320
you're not gonna get that time saving.

492
00:26:14,320 --> 00:26:17,680
And, yeah, it's powerful, we do anyways,

493
00:26:17,680 --> 00:26:19,960
because in a lot of ways,

494
00:26:19,960 --> 00:26:23,280
imagine going to an organization that has been doing something

495
00:26:23,280 --> 00:26:26,520
a particular way for hundreds of years

496
00:26:26,520 --> 00:26:28,560
or for tens of years, right?

497
00:26:28,560 --> 00:26:31,400
And now, if you could, let's do this way with AI,

498
00:26:31,400 --> 00:26:36,280
there's that adoption piece, which a lot of projects

499
00:26:36,280 --> 00:26:38,240
don't really accommodate for that.

500
00:26:38,240 --> 00:26:41,080
They don't really plan that adoption piece into it.

501
00:26:41,080 --> 00:26:42,560
And, you know, there are some people

502
00:26:42,560 --> 00:26:46,440
that are actually adoption and change management specialists.

503
00:26:46,440 --> 00:26:50,200
So, I feel like every AI implementation should have

504
00:26:50,200 --> 00:26:53,640
some form of ACM, and hold it in,

505
00:26:53,640 --> 00:26:57,120
and just to make sure that you are tracking the rights,

506
00:26:57,120 --> 00:27:01,440
the right KPIs, and you are able to measure

507
00:27:01,440 --> 00:27:04,600
how much the new tool is being used,

508
00:27:04,600 --> 00:27:07,240
'cause if it's not being used, as I said,

509
00:27:07,240 --> 00:27:09,120
you're not gonna get value for it.

510
00:27:09,120 --> 00:27:12,600
- And how can companies say, okay,

511
00:27:12,600 --> 00:27:14,720
those are good business cases,

512
00:27:14,720 --> 00:27:17,880
and that's, I don't know, not so good business case,

513
00:27:17,880 --> 00:27:20,800
what your tips here, when they will start,

514
00:27:20,800 --> 00:27:22,760
was co-part, it's to you.

515
00:27:22,760 --> 00:27:24,000
- Yeah.

516
00:27:24,000 --> 00:27:28,680
- So, I'll say something that my boss usually says,

517
00:27:28,680 --> 00:27:32,320
and she's like, when she's speaking to players,

518
00:27:32,320 --> 00:27:37,320
she's like, okay, what part of your day don't you like?

519
00:27:37,320 --> 00:27:40,680
So, what is the, just look at,

520
00:27:40,680 --> 00:27:42,520
just picture your day and see,

521
00:27:42,520 --> 00:27:45,080
what's the activity that you do,

522
00:27:45,080 --> 00:27:48,120
that you do not like, or you feel like

523
00:27:48,120 --> 00:27:51,080
you should not be the one doing this, right?

524
00:27:51,080 --> 00:27:54,680
Those kind of activities can start showing you

525
00:27:54,680 --> 00:27:56,800
some use cases for AI.

526
00:27:56,800 --> 00:27:59,560
Because when you look at the whole idea of automation

527
00:27:59,560 --> 00:28:02,680
in the first place, and especially sometimes,

528
00:28:02,680 --> 00:28:05,520
when I'm delivering training about automation,

529
00:28:05,520 --> 00:28:08,280
we say, yeah, we wanna automate the repetitive,

530
00:28:08,280 --> 00:28:09,920
the boring tax, right?

531
00:28:09,920 --> 00:28:14,920
So, now, with AI, we can even take it a step further,

532
00:28:14,920 --> 00:28:19,360
because with AI, there's that reasoning capacity.

533
00:28:19,360 --> 00:28:23,080
So, previously, automation could only handle things

534
00:28:23,080 --> 00:28:25,960
that, well, it's just, it's more repetitive,

535
00:28:25,960 --> 00:28:28,000
and there's more no value in it.

536
00:28:28,000 --> 00:28:32,280
But now, with AI, we can actually automate some scenarios

537
00:28:32,280 --> 00:28:35,320
that actually require intelligence and reasoning.

538
00:28:35,320 --> 00:28:38,280
So, I feel like that big game changer

539
00:28:38,280 --> 00:28:40,160
that AI brings to the table.

540
00:28:40,160 --> 00:28:42,640
So, when you're trying to discover use cases,

541
00:28:42,640 --> 00:28:46,200
and yet we do have engagements around working with clients

542
00:28:46,200 --> 00:28:48,400
to come up with those use cases,

543
00:28:48,400 --> 00:28:52,800
or one of those ways is just look at your day and see,

544
00:28:52,800 --> 00:28:55,920
okay, what are the things that I do not like to do?

545
00:28:55,920 --> 00:28:58,320
Another thing you could see is, okay,

546
00:28:58,320 --> 00:29:01,480
what are the things that I would like an assistant for?

547
00:29:01,480 --> 00:29:03,880
So, imagine you have an assistant.

548
00:29:03,880 --> 00:29:08,720
What kind of tasks can I comfortably give my assistant?

549
00:29:08,720 --> 00:29:12,600
And maybe I just go through what the assistant does, right?

550
00:29:12,600 --> 00:29:14,760
You know, there will definitely be some tasks that you say,

551
00:29:14,760 --> 00:29:17,840
well, this one, I absolutely must treat myself.

552
00:29:17,840 --> 00:29:20,920
But you will definitely have some cases where

553
00:29:20,920 --> 00:29:23,320
you can delegate this one assistant.

554
00:29:23,320 --> 00:29:28,160
So, that can also be a way of finding out some of those use cases

555
00:29:28,160 --> 00:29:29,080
for AI.

556
00:29:29,080 --> 00:29:34,240
Have you a use case without exposing the company's name

557
00:29:34,240 --> 00:29:36,800
where you say, well, this was more than an amazing one?

558
00:29:36,800 --> 00:29:39,880
Okay, so, yeah, this is one,

559
00:29:39,880 --> 00:29:41,960
and I like to talk about this a lot,

560
00:29:41,960 --> 00:29:46,200
because we did it in the early days of co-pilot studio,

561
00:29:46,200 --> 00:29:51,200
and, yeah, it was more tedious than it would have been now,

562
00:29:51,200 --> 00:29:55,040
but it was a good one.

563
00:29:55,040 --> 00:30:00,040
So, basically, this organization, they do generate,

564
00:30:00,040 --> 00:30:02,520
well, they didn't generate that before,

565
00:30:02,520 --> 00:30:07,520
but they do have, they do pull together some documentation

566
00:30:07,520 --> 00:30:11,320
like sales guide documentation.

567
00:30:11,320 --> 00:30:14,520
And to do the sales guide documentation,

568
00:30:14,520 --> 00:30:18,920
they would have to do research around their competitors,

569
00:30:18,920 --> 00:30:21,720
compare the loan bars, and just try to come up with,

570
00:30:21,720 --> 00:30:26,240
come up with, kind of, differentiators

571
00:30:26,240 --> 00:30:30,440
of how their product is different from the other,

572
00:30:30,440 --> 00:30:31,800
the competitor products.

573
00:30:31,800 --> 00:30:34,040
And so, there's like that research piece,

574
00:30:34,040 --> 00:30:36,640
and there's the documentation piece,

575
00:30:36,640 --> 00:30:40,960
the messaging, writing it into an organization,

576
00:30:41,920 --> 00:30:44,280
they organize a short language the way they speak,

577
00:30:44,280 --> 00:30:45,600
and things like that.

578
00:30:45,600 --> 00:30:47,920
There was also a debate about translating

579
00:30:47,920 --> 00:30:50,320
into multiple languages, right?

580
00:30:50,320 --> 00:30:53,480
So, after they have the whole document,

581
00:30:53,480 --> 00:30:57,960
they will need to make it translated to other languages.

582
00:30:57,960 --> 00:31:01,000
So, you can see, this has like so many different components,

583
00:31:01,000 --> 00:31:01,840
right?

584
00:31:01,840 --> 00:31:04,400
But, we were able to build like an AI agent

585
00:31:04,400 --> 00:31:08,560
to do all of this, where they just upload a few such documents,

586
00:31:08,560 --> 00:31:12,880
and the AI agent goes, does all that research, brings it together,

587
00:31:12,880 --> 00:31:16,960
well, includes the sources, where it's gotten all those information

588
00:31:16,960 --> 00:31:20,440
from, and just help them through that end-to-end process.

589
00:31:20,440 --> 00:31:25,920
So, something that I would have taken, maybe a week or five days,

590
00:31:25,920 --> 00:31:28,400
well, now it's taking minutes,

591
00:31:28,400 --> 00:31:30,840
and they can still take the output,

592
00:31:30,840 --> 00:31:33,840
and, you know, make sure it's valid,

593
00:31:33,840 --> 00:31:35,280
and just do some checking.

594
00:31:35,280 --> 00:31:37,120
And one thing about this thing is,

595
00:31:37,120 --> 00:31:41,800
over time, then, the confidence can increase in the solution,

596
00:31:41,800 --> 00:31:43,280
you know, when you see something,

597
00:31:43,280 --> 00:31:47,760
and you're saying that, well, it's 90% of the time is getting it right,

598
00:31:47,760 --> 00:31:50,120
then you begin to trust it more.

599
00:31:50,120 --> 00:31:52,000
So, by the early stages, you might be like,

600
00:31:52,000 --> 00:31:53,320
"Oh, I need to check everything,

601
00:31:53,320 --> 00:31:57,280
but as time goes on, when you see it behaving predictably,

602
00:31:57,280 --> 00:32:02,440
consistent, then, you'll kind of start getting more of those times saved."

603
00:32:02,440 --> 00:32:05,040
So, I'll say, that's an example.

604
00:32:05,040 --> 00:32:11,440
Another example is an organization that is smaller than an admission board,

605
00:32:11,440 --> 00:32:15,360
where they have some criteria for admission.

606
00:32:15,360 --> 00:32:21,680
And so, this one is kind of where AI gives recommendations,

607
00:32:21,680 --> 00:32:23,920
or AI don't make a decision.

608
00:32:23,920 --> 00:32:26,640
And that's another distinction that is screwed,

609
00:32:26,640 --> 00:32:30,520
because as we go through and start implementing AI,

610
00:32:30,520 --> 00:32:34,040
now, we do not want AI to be the final decision

611
00:32:34,040 --> 00:32:36,040
that may kind of some cases write,

612
00:32:36,040 --> 00:32:39,880
but we just want AI to be able to give us recommendations as well,

613
00:32:39,880 --> 00:32:42,120
based on your documentation,

614
00:32:42,120 --> 00:32:45,960
based on your documents, your criteria,

615
00:32:45,960 --> 00:32:49,160
this candidates feature criteria,

616
00:32:49,160 --> 00:32:54,040
but we are not saying, well, this is the only candidate you should look at,

617
00:32:54,040 --> 00:32:57,000
so both these candidates do feature criteria,

618
00:32:57,000 --> 00:33:00,200
but we are still leaving the ultimate decision to the human

619
00:33:00,200 --> 00:33:02,600
to actually make the final decision.

620
00:33:02,600 --> 00:33:04,440
So, we don't want to meet everything,

621
00:33:04,440 --> 00:33:06,040
but AI can actually help,

622
00:33:06,040 --> 00:33:11,640
because rather than going through a 200 page document,

623
00:33:11,640 --> 00:33:16,440
AI can help you highlight some key points of key areas

624
00:33:16,440 --> 00:33:20,200
that you might have even missed if you had to sit down

625
00:33:20,200 --> 00:33:23,000
and go through the 200 page document.

626
00:33:23,000 --> 00:33:24,440
So, yeah, those are some use cases,

627
00:33:24,440 --> 00:33:28,440
but yeah, we've had so many AI use cases, as I've said,

628
00:33:28,440 --> 00:33:31,240
the conversations are really changing a lot,

629
00:33:31,240 --> 00:33:37,800
so AI first, so those are some two use cases that I find very interesting.

630
00:33:37,800 --> 00:33:39,960
That's a cool use case.

631
00:33:39,960 --> 00:33:42,920
I think a little bit about a lot of companies

632
00:33:42,920 --> 00:33:47,160
or a bus world, AI readiness or co-pilot readiness,

633
00:33:47,160 --> 00:33:50,280
what will you think is the minimum requirement

634
00:33:50,280 --> 00:33:56,840
companies have to be successful or to implement co-pilot studio?

635
00:33:56,840 --> 00:34:00,920
Okay, yeah, it's good that you actually said co-pilot studio,

636
00:34:00,920 --> 00:34:04,200
or co-pilot, because if it's co-pilot,

637
00:34:04,200 --> 00:34:09,640
it becomes a bigger conversation, because co-pilot,

638
00:34:09,640 --> 00:34:14,360
it's really hard to put co-pilot in a box,

639
00:34:14,360 --> 00:34:17,960
because co-pilot is very dependent,

640
00:34:17,960 --> 00:34:21,800
especially where you go to the more premium test

641
00:34:21,800 --> 00:34:25,080
that you have things like work IQ.

642
00:34:25,080 --> 00:34:29,480
So, with those ones, it really uses your organization data

643
00:34:29,480 --> 00:34:33,160
in the sense that it has access to your organization data.

644
00:34:33,160 --> 00:34:34,840
I don't know how else to say that.

645
00:34:34,840 --> 00:34:37,800
So, that's where it starts exposing things like,

646
00:34:37,800 --> 00:34:43,080
oh, you, where you can actually find some information

647
00:34:43,080 --> 00:34:46,840
that will HR probably didn't want you to find,

648
00:34:46,840 --> 00:34:51,400
but because you've been added to a SharePoint site,

649
00:34:51,400 --> 00:34:54,840
maybe for something, now you have access to all of that data.

650
00:34:54,840 --> 00:34:59,400
So, AI will expose that, especially when you're talking about co-pilot.

651
00:34:59,400 --> 00:35:02,280
But, I really love the world of co-pilot studio,

652
00:35:02,280 --> 00:35:05,960
because with co-pilot studio, you can actually put AI in the box.

653
00:35:05,960 --> 00:35:09,960
Now, if your organization data is no in your right place,

654
00:35:09,960 --> 00:35:15,160
as, and that's actually part of the whole AI readiness conversation,

655
00:35:15,160 --> 00:35:20,200
even at that stage, I can still build co-pilot studio agents for you,

656
00:35:20,200 --> 00:35:24,600
and you can still use it safely without having issues

657
00:35:24,600 --> 00:35:27,160
with the rest of your organization data,

658
00:35:27,160 --> 00:35:30,280
because with co-pilot studio agents, you can actually point

659
00:35:30,280 --> 00:35:33,880
to the actual knowledge sources you want it to use,

660
00:35:33,880 --> 00:35:36,040
and you can turn off things like work IQ,

661
00:35:36,040 --> 00:35:38,520
even though your organization data is in a mess,

662
00:35:38,520 --> 00:35:41,560
then you're definitely not going to be telling on work IQ,

663
00:35:41,560 --> 00:35:43,800
because that's just going to expose everything.

664
00:35:43,800 --> 00:35:48,600
So, with co-pilot studio, and that's why a lot of solutions,

665
00:35:48,600 --> 00:35:52,600
especially only on with clients, maybe they are not so far

666
00:35:52,600 --> 00:35:54,680
on that AI readiness journey,

667
00:35:54,680 --> 00:35:59,880
they could actually start with simple use cases developed in co-pilot studio,

668
00:35:59,880 --> 00:36:03,800
because with co-pilot studio, we can actually restrict

669
00:36:03,800 --> 00:36:08,120
what data sources, what access the agent has,

670
00:36:08,120 --> 00:36:12,680
and that just doesn't expose the whole mess

671
00:36:12,680 --> 00:36:15,400
that the organization data might be in.

672
00:36:15,400 --> 00:36:19,480
And I can do this without touching, I don't know, peer view,

673
00:36:19,480 --> 00:36:22,200
and so on. I can do all this at co-pilot studio.

674
00:36:23,240 --> 00:36:26,600
So, it's all false, it's all goes hand in hand.

675
00:36:26,600 --> 00:36:33,960
In, like, say, leveraging some SharePoint documents, right?

676
00:36:33,960 --> 00:36:41,160
Now, within peer view, you can specify what the label

677
00:36:41,160 --> 00:36:44,440
should be, whether it's confidential, and the likes,

678
00:36:44,440 --> 00:36:46,200
you can do that within peer view.

679
00:36:46,200 --> 00:36:51,400
Now, when that becomes ingested by co-pilot studio,

680
00:36:51,400 --> 00:36:55,960
co-pilot studio is going to leverage all that you've done in peer view.

681
00:36:55,960 --> 00:36:58,760
Now, if you've not done anything in peer view,

682
00:36:58,760 --> 00:37:01,080
co-pilot studio is to work with that data,

683
00:37:01,080 --> 00:37:04,200
or it's logger to assume any sensitivity labels.

684
00:37:04,200 --> 00:37:07,400
It's just going to treat everything as not sensitive.

685
00:37:07,400 --> 00:37:11,320
So, they work hand in hand, and if it's truly non-sensitive,

686
00:37:11,320 --> 00:37:13,640
maybe you don't have to do anything with peer view,

687
00:37:13,640 --> 00:37:17,560
but if they think that you need to categorize properly,

688
00:37:17,560 --> 00:37:19,480
then peer view comes into the picture.

689
00:37:19,480 --> 00:37:21,800
So, it's really about what you're trying to do,

690
00:37:21,800 --> 00:37:24,120
but yeah, they're definitely co-pilot use cases

691
00:37:24,120 --> 00:37:29,240
that do not need peer view because you don't need all of those sensitivity labels.

692
00:37:29,240 --> 00:37:36,520
So, it's really, you know, it depends on the extent of conversations you're having,

693
00:37:36,520 --> 00:37:39,160
and what you're really trying to achieve, really.

694
00:37:39,160 --> 00:37:44,600
And, when we look, we look a lot of technology,

695
00:37:46,040 --> 00:37:53,240
but what kind of, yeah, for success plays the company culture, change management,

696
00:37:53,240 --> 00:37:55,240
so I think all around the people.

697
00:37:55,240 --> 00:37:56,280
Yeah.

698
00:37:56,280 --> 00:38:01,640
So, if I get the question right, you're asking,

699
00:38:01,640 --> 00:38:08,440
how does the people component impacts the AI implementations?

700
00:38:08,440 --> 00:38:15,560
Yeah. So, there's the beat around governance.

701
00:38:15,560 --> 00:38:18,200
I think I'll start from governance because, well, you're like,

702
00:38:18,200 --> 00:38:21,160
oh, you asked about people, but I'm talking about governance.

703
00:38:21,160 --> 00:38:26,440
Because what I initially wanted to say before governance was

704
00:38:26,440 --> 00:38:29,560
people doing the right things, right?

705
00:38:29,560 --> 00:38:33,400
But what governance allows you to do is,

706
00:38:33,400 --> 00:38:40,440
it allows you to kind of provide a structure for people to easily do the right things.

707
00:38:40,440 --> 00:38:44,520
So, say, talking about documents, share it.

708
00:38:44,840 --> 00:38:50,680
You can actually implement governance that restricts the way

709
00:38:50,680 --> 00:38:53,880
documents are shared in your organization, right?

710
00:38:53,880 --> 00:38:57,880
And that will just automatically help them to do the right things.

711
00:38:57,880 --> 00:39:01,800
And prevent issues where somebody has actually been shared,

712
00:39:01,800 --> 00:39:05,480
the documents that they should have been shared,

713
00:39:05,480 --> 00:39:07,880
and now they have access to the old SharePoint right.

714
00:39:07,880 --> 00:39:11,880
So, the people conversation starts from governance,

715
00:39:11,880 --> 00:39:13,320
having the right governance.

716
00:39:14,280 --> 00:39:17,320
Framework having the right governance tools in place,

717
00:39:17,320 --> 00:39:18,600
clicking the right buttons,

718
00:39:18,600 --> 00:39:19,960
disabling the right things.

719
00:39:19,960 --> 00:39:23,080
That just ensures people do the right thing.

720
00:39:23,080 --> 00:39:27,720
So, every AI conversation starts from your governance.

721
00:39:27,720 --> 00:39:30,040
And there's some clients that have come to us and said,

722
00:39:30,040 --> 00:39:32,200
well, we want to do this and this and this and this and this.

723
00:39:32,200 --> 00:39:35,160
I will like wait before we start that conversation.

724
00:39:35,160 --> 00:39:37,560
Let's actually talk about your governance,

725
00:39:37,560 --> 00:39:39,800
and let's sort this out first because,

726
00:39:39,800 --> 00:39:42,200
well, you could build a file situ,

727
00:39:42,200 --> 00:39:46,280
and it's adding value in the way it's supposed to,

728
00:39:46,280 --> 00:39:50,360
but it could start exposing some other issues

729
00:39:50,360 --> 00:39:53,240
because you've learned how to do that governance piece first.

730
00:39:53,240 --> 00:39:57,800
Yeah, that looked a little bit deep dive in the governance topic,

731
00:39:57,800 --> 00:39:58,840
especially with AI.

732
00:39:58,840 --> 00:40:04,680
How should companies start with governance,

733
00:40:04,680 --> 00:40:10,760
especially when they like to start with the world-core pilot,

734
00:40:11,320 --> 00:40:13,320
worlds, what's your tip?

735
00:40:13,320 --> 00:40:17,640
So, how should they start with governance?

736
00:40:17,640 --> 00:40:21,880
Well, the real answer to that is,

737
00:40:21,880 --> 00:40:27,000
first of all, what tools do you have available?

738
00:40:27,000 --> 00:40:32,120
And there is why I said that's a real answer is,

739
00:40:32,120 --> 00:40:34,120
within the Microsoft stack, you know,

740
00:40:34,120 --> 00:40:37,800
the different licenses stack, different licenses stack,

741
00:40:37,800 --> 00:40:39,960
sectionally-giving access to different tools.

742
00:40:40,840 --> 00:40:46,600
So, the more you pay, the more you get, basically,

743
00:40:46,600 --> 00:40:51,000
if you pay more, you get some advanced tools

744
00:40:51,000 --> 00:40:53,080
that can help you do some more things.

745
00:40:53,080 --> 00:40:58,680
So, the real, the first real, real bit of that is what tools do you have access to?

746
00:40:58,680 --> 00:41:01,800
Now, the next bit around that is,

747
00:41:01,800 --> 00:41:03,880
how do you use the tools?

748
00:41:03,880 --> 00:41:05,160
Do you use them properly?

749
00:41:05,160 --> 00:41:09,880
Now, I've spoken to many IT admins that do not actually

750
00:41:10,680 --> 00:41:13,960
know the right places to go to do the admin,

751
00:41:13,960 --> 00:41:17,000
to do the admin and governance pieces.

752
00:41:17,000 --> 00:41:19,240
And, you know, it's not necessarily the default.

753
00:41:19,240 --> 00:41:24,040
In many cases, the same IT admin is doing so many other things, right?

754
00:41:24,040 --> 00:41:27,480
And the priority is very different.

755
00:41:27,480 --> 00:41:30,600
They have, like, so many different hacks they're wearing.

756
00:41:30,600 --> 00:41:35,640
So, it's hard to be, like, very focused on just one particular bit.

757
00:41:35,640 --> 00:41:38,840
So, as I said, the first thing is, what tools do you have available?

758
00:41:39,320 --> 00:41:42,520
Following on from that, I use those tools properly,

759
00:41:42,520 --> 00:41:46,920
because if you're using the tools properly, then you have, like, better governance.

760
00:41:46,920 --> 00:41:53,240
Then, you know, never forgetting the people aspect, the bit around training,

761
00:41:53,240 --> 00:41:57,160
around empowering people to do the right things.

762
00:41:57,160 --> 00:41:59,320
So, by giving them the right frameworks,

763
00:41:59,320 --> 00:42:01,160
we're also giving them the right knowledge,

764
00:42:01,160 --> 00:42:07,320
giving them the right training on how to maximize the tools they have, and you're disposed of.

765
00:42:07,880 --> 00:42:12,440
So, yeah, governance conversations, they're always big conversations,

766
00:42:12,440 --> 00:42:16,520
and we have, like, always have big name people under conversations, like,

767
00:42:16,520 --> 00:42:20,760
infosec, compliance, office, as and things like that.

768
00:42:20,760 --> 00:42:28,600
But, yeah, as we both know, right, that's usually the first bit to start,

769
00:42:28,600 --> 00:42:34,120
because once you have good governance, in place, then you have that confidence,

770
00:42:35,080 --> 00:42:38,920
because that's why we do all of that effort, right, to have that confidence,

771
00:42:38,920 --> 00:42:43,880
and we can actually slip well at night, and know that nothing is going to break overnight.

772
00:42:43,880 --> 00:42:52,440
And another topic, I work more with, I foundry, but another topic is what I often heard,

773
00:42:52,440 --> 00:42:58,920
it's their security concerns. So, how to make, yeah,

774
00:42:58,920 --> 00:43:03,160
this co-pilot studio secure, what's your tips here?

775
00:43:04,520 --> 00:43:07,800
So, yeah, I noticed you mentioned AI Foundry there,

776
00:43:07,800 --> 00:43:14,680
and I'll say that was one of the distinctions between AI Foundry and co-pilot studio,

777
00:43:14,680 --> 00:43:20,600
in the sense that with things like AI Foundry, you get to manage a lot yourself,

778
00:43:20,600 --> 00:43:26,280
but with co-pilot studio, you manage less because you're using a managed platform.

779
00:43:26,280 --> 00:43:33,000
So, that's the whole idea of local. So, AI Foundry, you can't really call it, as much as AI Foundry

780
00:43:33,000 --> 00:43:39,640
abstracts some of the code required to do all of these AI things, but you can't necessarily call

781
00:43:39,640 --> 00:43:45,000
AI Foundry and local platform, but co-pilot studio is a local platform. So, there are things that

782
00:43:45,000 --> 00:43:53,640
Microsoft actually does to kind of manage and protect the platform itself. So, the bit you will

783
00:43:53,640 --> 00:44:00,680
need to focus on is really in terms of your agents design. So, have you designed the agents in a way

784
00:44:00,680 --> 00:44:06,040
that, and I was saying on that podcast recently, that one thing I like to do anytime I design an

785
00:44:06,040 --> 00:44:11,960
agent is ask it a very irrelevant question, and see what it says. If I ask it, an irrelevant question,

786
00:44:11,960 --> 00:44:17,000
and answering me, then I know that, you know, I've actually missed something in the design.

787
00:44:17,000 --> 00:44:23,800
So, within the co-pilot studio world, you get to just focus on more like your design, your agents

788
00:44:23,800 --> 00:44:29,640
design, and just make sure that that is tight. Microsoft handles a lot of the things around things like

789
00:44:29,640 --> 00:44:35,320
the dust protection, prompt injection, and stuff like that. So, if you're building something

790
00:44:35,320 --> 00:44:42,760
custom yourself, you will be needing to engineer those bits in, or within co-pilot studio, those

791
00:44:42,760 --> 00:44:47,960
are kind of covered by Microsoft, and you just focus on the agents design piece.

792
00:44:47,960 --> 00:44:57,640
And what I also think, there are a lot of our stories out there about final,

793
00:44:58,200 --> 00:45:05,400
because we don't pay this only the $30 product for co-pilot. We also pay for tokens.

794
00:45:05,400 --> 00:45:16,200
So, what's your tip on final, so that the CFO don't get a heart attack?

795
00:45:16,200 --> 00:45:26,600
Yeah, that's interesting, and it is, if you look at it, well, all these organizations,

796
00:45:27,160 --> 00:45:34,840
Microsoft, Google, OpenAI, they're really, because it's such a new space,

797
00:45:34,840 --> 00:45:42,440
each, it's not so new, because we pay that for analysis, yes. But they try to do what's called

798
00:45:42,440 --> 00:45:48,760
like land grabbing, and they try to just take that market share. And because of that,

799
00:45:48,760 --> 00:45:55,800
even look at the whole idea of tokens, well, they've really subsidized a lot of the actual cost,

800
00:45:55,800 --> 00:46:02,920
and they've just made it look cheap. And in some cases, now we're seeing a form of change of mind,

801
00:46:02,920 --> 00:46:07,640
where they're like, oh, well, this is actually not so stupid. Now we have to start letting the

802
00:46:07,640 --> 00:46:14,760
users be some of the real costs. The way I see this is, I feel like overtime is still going to get

803
00:46:14,760 --> 00:46:24,440
better, because like every technology, it's, over time, it gets cheaper, and as we speak,

804
00:46:24,440 --> 00:46:33,240
the alludes of research going around different areas of this whole LLM's, and this new AI

805
00:46:33,240 --> 00:46:42,200
implementation, right? So, loads of research are going around beta centers, around sustainability,

806
00:46:42,200 --> 00:46:48,040
and things like that. So I feel like overtime, it's going to get better, and it won't be as

807
00:46:48,040 --> 00:46:55,640
expensive as it is now, but the fact is right now is really expensive. In a lot of ways, the,

808
00:46:55,640 --> 00:47:02,840
well, Microsoft in particular tries to rule things up into like a license, right? Like a subscription

809
00:47:02,840 --> 00:47:08,440
cost. Right now, it still have this. So I feel like, let's enjoy it as much as we can.

810
00:47:08,440 --> 00:47:17,160
Until they want to start changing things, because now even look at GitHub, Copilot. Now,

811
00:47:17,160 --> 00:47:21,480
I feel like that's really, really a lot of horror stories that come in and that people

812
00:47:21,480 --> 00:47:29,640
that say that, well, how could I use my subscription for a month in a day? What do I do for the rest of the

813
00:47:29,640 --> 00:47:36,360
month? Right. So that, it wasn't like that a few weeks ago, right? Because a few weeks ago,

814
00:47:36,360 --> 00:47:42,200
Microsoft was still happy to be able to cost, but now they're like, well, let's, we can continue to

815
00:47:42,200 --> 00:47:48,520
do this. And that's where you see Microsoft saying, well, we can give the free Copilot to organizations

816
00:47:48,520 --> 00:47:54,760
less than a certain amount of members, but once they have more than, I can't remember what

817
00:47:54,760 --> 00:47:58,520
they know about it right now, but once they have more than that number, then they don't get the

818
00:47:58,520 --> 00:48:06,440
free Copilot. That's because there's actual real cost behind the scenes. They try to cover it up

819
00:48:06,440 --> 00:48:14,920
and so that you can subscribe and get that, you know, get the license, the latest and the shiny license.

820
00:48:14,920 --> 00:48:22,920
But yeah, the fact is the real cost are beginning to show, but my confidence is over time,

821
00:48:22,920 --> 00:48:29,560
those real costs are actually going to go down and who gets point where the AI and injecting

822
00:48:29,560 --> 00:48:36,200
use cases become much more sustainable and much more affordable. And I actually believe there's a

823
00:48:36,200 --> 00:48:44,280
feature, a feature of local, local LLM's where you have a LLM on your machine that is actually doing

824
00:48:44,280 --> 00:48:51,240
a lot of heavy lifting right now. I did install some local LLM's and I'm doing some like image

825
00:48:51,240 --> 00:48:58,840
processing categorization on my computer offline without the, well, was it anything? Maybe not,

826
00:48:58,840 --> 00:49:03,960
because yeah, it's really local on my computer. So I believe there is going to be a future of local

827
00:49:03,960 --> 00:49:10,600
large language modules, but yeah, everything is still dependent on the research that is being done

828
00:49:10,600 --> 00:49:17,720
all around the world right now. I think another option, I don't know if it's possible in Copilot,

829
00:49:17,720 --> 00:49:25,400
studio, but on AI Foundry, it's one, it's used local models. Yeah, also I don't need the biggest

830
00:49:25,400 --> 00:49:32,200
model for everything. So I can say, that's the thing. Yeah, actually, I don't know if there's also

831
00:49:32,200 --> 00:49:40,200
possible in Copilot studio to change the models. Well, so you can change models though, you're

832
00:49:40,200 --> 00:49:47,560
definitely not using local models because that is, yeah, all web, web is a web solution basically.

833
00:49:47,560 --> 00:49:54,920
So you won't be using the local model there, though you can probably architect something

834
00:49:55,320 --> 00:50:00,280
that leverages a local model if that's really what you want to do and just have endpoints that talk

835
00:50:00,280 --> 00:50:06,440
to each other. But out of the box, Copilot studio, you can select models, you can switch models,

836
00:50:06,440 --> 00:50:12,040
you have just looked into one model, you can change models and one of our workflows, you could actually

837
00:50:12,040 --> 00:50:19,240
have a model from Azure AI Foundry and you can still leverage that within your Copilot studio agent,

838
00:50:20,120 --> 00:50:26,280
but you won't be able to reach local models out of the box and Copilot studio. Yeah,

839
00:50:26,280 --> 00:50:34,200
that's cool. So now we can join the quickfire round. I asked you question and you give me a

840
00:50:34,200 --> 00:50:38,280
shorter answer. What's the answer with your mind? So power app, all the way to

841
00:50:38,280 --> 00:50:47,480
power, activate, okay, Canvas app or model driven apps.

842
00:50:48,120 --> 00:50:54,760
Canvas apps. Copilot studio or traditional chatbot development?

843
00:50:54,760 --> 00:51:01,000
Copilot studio, please. One Microsoft technology people should pay more attention to.

844
00:51:01,000 --> 00:51:13,400
Copilot studio. Where Microsoft comes to you tomorrow and say you get unlimited money,

845
00:51:13,400 --> 00:51:25,080
people and so on. What will you develop? Microsoft. I will develop something that makes me

846
00:51:25,080 --> 00:51:33,400
money and I don't have to work again. It's good. What is one overrated trend in tech?

847
00:51:35,000 --> 00:51:47,320
One what? One overrated trend, actually, in tech. This is tough overrated, trained in tech. AI

848
00:51:47,320 --> 00:51:57,800
content creation. And what was the best career advice you ever received?

849
00:52:01,400 --> 00:52:16,840
The best career advice I've ever received is really about politics and is more like play the game.

850
00:52:16,840 --> 00:52:22,280
Play the career game, I think that's the best career advice I've ever received.

851
00:52:24,760 --> 00:52:32,920
Is there a book, resource, podcast or so on? You recommend to everyone who reads especially in technology?

852
00:52:32,920 --> 00:52:38,200
So I'll say stay tuned to the power platform, DeepLife podcast.

853
00:52:38,200 --> 00:52:49,960
I also put this in the show notes. If you've ever heard of it, if you weren't working tech,

854
00:52:49,960 --> 00:53:01,000
what would you be doing? I will be so even knowing tech, I will be doing video production and telling stories.

855
00:53:01,000 --> 00:53:12,200
Yeah, thank you. So my last question is when we think about this episode and there are people

856
00:53:12,200 --> 00:53:17,720
they are starting with power platform in the day. What point or what does the key point they

857
00:53:17,720 --> 00:53:30,360
should take from this discussion today? So yeah, I will say, do not just focus on the technical

858
00:53:30,360 --> 00:53:36,040
implementation, sorry, I'll take that again. I'll say you do not just focus on the technical

859
00:53:36,040 --> 00:53:46,680
implementation because in many cases, that's just like 60% or even 50% of the story. There is so many

860
00:53:46,680 --> 00:53:54,360
other beats around governance and adoption. They are very important because you could do all the

861
00:53:54,360 --> 00:54:01,080
technical beats, but it still feels because you've learned to stop around governance and adoption.

862
00:54:01,080 --> 00:54:07,320
Yeah, so then I say thank you for joining me today on the M65FM podcast. It was,

863
00:54:07,320 --> 00:54:13,400
yeah, I really enjoyed this conversation because it brought a practical realistic perspective

864
00:54:13,400 --> 00:54:20,600
to topics that are often surrounded by hype. Whatever we are talking about, power platform,

865
00:54:20,600 --> 00:54:27,640
co-pilot, enterprise architecture or AI adoption, one message stood out clearly. Successful

866
00:54:27,640 --> 00:54:33,080
transformation is never just about technology, it's about people, governance, business outcomes,

867
00:54:33,080 --> 00:54:41,080
and building solutions that can scale. So yeah, thank you. This was so awesome. Thank you for your time.

868
00:54:41,720 --> 00:54:45,640
Yes, thank you so much, Marco. I really enjoyed every beat of these sad. Yeah,

869
00:54:45,640 --> 00:54:53,640
look at her to see the recording. Goodbye, tall. Cheers. Bye.