From Financial Data to AI-Ready Decisions: Power BI, Microsoft Fabric, Semantic Models & AI Agents with Rishi Sapra [MVP]

1
00:00:00,000 --> 00:00:05,000
Welcome to another edition of the MC65FM podcast.

2
00:00:05,000 --> 00:00:09,760
Today we welcome Richie and we want to go behind

3
00:00:09,760 --> 00:00:13,680
dashboard and even beyond Microsoft fabric itself.

4
00:00:13,680 --> 00:00:18,800
Financial organizations already have enormous amount

5
00:00:18,800 --> 00:00:21,900
of data, they have earpiece system, budgets,

6
00:00:21,900 --> 00:00:26,900
forecast, actual cost centers, power BI reports,

7
00:00:27,320 --> 00:00:32,320
ex-surprise and yeah, increasingly using Microsoft fabric.

8
00:00:32,320 --> 00:00:35,480
Now we adding a co-pilot and AI agents

9
00:00:35,480 --> 00:00:38,800
on the top of all of this, but AI agents doesn't,

10
00:00:38,800 --> 00:00:42,880
yeah, automatically understand what revenue

11
00:00:42,880 --> 00:00:47,040
may margin, consume up a probability

12
00:00:47,040 --> 00:00:50,280
or even this quarter means to particular companies.

13
00:00:50,280 --> 00:00:53,360
So today I want to explore the layers

14
00:00:53,360 --> 00:00:57,440
between raw financial data and AI that actually

15
00:00:57,440 --> 00:01:00,760
understand the business and that brings us directly

16
00:01:00,760 --> 00:01:02,840
into the world of the Monteque models,

17
00:01:02,840 --> 00:01:06,960
context layers, ontologies, governments and present driven data.

18
00:01:06,960 --> 00:01:10,480
So, but I think we let's start with you, Richie.

19
00:01:10,480 --> 00:01:14,200
How would you get into the Microsoft data

20
00:01:14,200 --> 00:01:16,520
and analytics ecosystem?

21
00:01:16,520 --> 00:01:21,200
- Well, through the beloved Excel, power pair of power queries.

22
00:01:21,200 --> 00:01:23,480
And when power BI came out in 2015,

23
00:01:23,480 --> 00:01:26,880
I was a financial modeler, KPMG.

24
00:01:26,880 --> 00:01:31,160
So my background before I got into data world really was big four.

25
00:01:31,160 --> 00:01:33,320
So working to the big four Deloitte and KPMG

26
00:01:33,320 --> 00:01:37,080
and then some of the big banks in London, HSPC, Barclays,

27
00:01:37,080 --> 00:01:39,000
doing kind of process automation.

28
00:01:39,000 --> 00:01:40,960
I've got still a fairly technical role

29
00:01:40,960 --> 00:01:43,720
rather than pure accounting, but yeah,

30
00:01:43,720 --> 00:01:45,280
financial modeling was kind of that area

31
00:01:45,280 --> 00:01:46,960
where I really got into kind of the data side

32
00:01:46,960 --> 00:01:49,000
behind all of this and making decisions

33
00:01:49,000 --> 00:01:52,080
and how is data used to drive calculations?

34
00:01:52,080 --> 00:01:53,680
And yeah, power BI came out then

35
00:01:53,680 --> 00:01:56,080
and I saw a massive opportunity.

36
00:01:56,080 --> 00:01:59,560
I mean, I don't know if you do remember power BI in 2015.

37
00:01:59,560 --> 00:02:05,680
It was pretty basic to be polite, I mean, pretty craft.

38
00:02:05,680 --> 00:02:09,520
A little bit less polite.

39
00:02:09,520 --> 00:02:13,480
But even then, I think it just showed that comments

40
00:02:13,480 --> 00:02:14,960
of especially the semantic model stuff

41
00:02:14,960 --> 00:02:17,880
in the power query of making that accessible

42
00:02:17,880 --> 00:02:22,600
beyond kind of reporting being stuck in IT teams

43
00:02:22,600 --> 00:02:28,200
and data and BI teams and having to outsource it all to consultants

44
00:02:28,200 --> 00:02:30,680
because that's been the model for decades.

45
00:02:30,680 --> 00:02:35,240
And I think power BI and Tableau and Kleeck as well,

46
00:02:35,240 --> 00:02:37,960
obviously, kind of really tried to change that model around

47
00:02:37,960 --> 00:02:40,280
and introduce this whole idea of self-service BI.

48
00:02:40,280 --> 00:02:43,960
And I think it's great and I really like the idea

49
00:02:43,960 --> 00:02:46,800
that actually, as a non-technical business user

50
00:02:46,800 --> 00:02:49,440
who's not a developer, you can actually go

51
00:02:49,440 --> 00:02:50,960
and build these things yourself.

52
00:02:50,960 --> 00:02:53,400
You can go in rather than just use Excel,

53
00:02:53,400 --> 00:02:56,680
you can go and build solutions that can sit

54
00:02:56,680 --> 00:02:58,160
within your corporate environment

55
00:02:58,160 --> 00:03:01,200
and connect to your source systems, connect to all of these things.

56
00:03:01,200 --> 00:03:03,840
And this was before Power Quiz in Excel obviously as well, right?

57
00:03:03,840 --> 00:03:06,680
So, well, at least built into Excel.

58
00:03:06,680 --> 00:03:09,480
And it was an add-on in 2010.

59
00:03:09,480 --> 00:03:15,920
So it gave that platform to be able to take Excel users,

60
00:03:15,920 --> 00:03:18,040
I think, into a whole new world.

61
00:03:18,040 --> 00:03:19,360
So that's kind of where I got into it,

62
00:03:19,360 --> 00:03:21,440
through this self-service BI wave.

63
00:03:21,440 --> 00:03:27,160
I think the challenge has been that since 2015, Microsoft,

64
00:03:27,160 --> 00:03:28,600
do you want to take, I don't know,

65
00:03:28,600 --> 00:03:30,600
do you want to take a guess at the number of features

66
00:03:30,600 --> 00:03:33,160
that Microsoft have added to Power BI?

67
00:03:33,160 --> 00:03:34,640
It would include fabric in that as well

68
00:03:34,640 --> 00:03:36,040
because it's probably on the bottom of fabric.

69
00:03:36,040 --> 00:03:37,560
I mean, does it take a guess, right?

70
00:03:37,560 --> 00:03:40,920
Tens of thousands of features and capabilities.

71
00:03:40,920 --> 00:03:44,000
So it's not so easy to use any board.

72
00:03:44,000 --> 00:03:46,880
It's not so easy for those kind of Excel users to kind of pick up.

73
00:03:46,880 --> 00:03:48,880
And now you've not got the in copilot studio.

74
00:03:48,880 --> 00:03:50,800
It's said that you've got semantic model,

75
00:03:50,800 --> 00:03:54,080
that you've got all sorts of technologies and tools.

76
00:03:54,080 --> 00:03:57,760
And I think it just gets overwhelming very quickly.

77
00:03:57,760 --> 00:04:00,800
So really, I think now actually,

78
00:04:00,800 --> 00:04:04,040
we need to think about how we can delegate that solution design

79
00:04:04,040 --> 00:04:07,920
in architecture and planning and thinking about how to build these things

80
00:04:07,920 --> 00:04:09,440
and actually do the building of them,

81
00:04:09,440 --> 00:04:13,880
delegate that to agents so that we can then focus on what we're,

82
00:04:13,880 --> 00:04:16,640
we're good at humans, which is not trying to keep up with all this tech

83
00:04:16,640 --> 00:04:18,000
and know how to do everything,

84
00:04:18,000 --> 00:04:21,320
but actually understand the business and build that context.

85
00:04:21,320 --> 00:04:23,120
This is saying to to to to Gen A.

86
00:04:23,120 --> 00:04:29,600
I yeah, I've seen you often write about topics like

87
00:04:29,600 --> 00:04:33,880
operation models, governance and there's all that driven insights.

88
00:04:33,880 --> 00:04:37,240
What do is it mean in practice?

89
00:04:37,240 --> 00:04:41,440
And why is it becoming more important now with the,

90
00:04:42,800 --> 00:04:47,480
I say the generative AI is entering the analytics tech.

91
00:04:47,480 --> 00:04:49,160
Oh, it's way more important now.

92
00:04:49,160 --> 00:04:52,760
There's, there's this book, I really recommend your audience to read.

93
00:04:52,760 --> 00:04:54,280
Actually, it's called Fair Game.

94
00:04:54,280 --> 00:04:55,920
So I think the website's Fair Game book,

95
00:04:55,920 --> 00:05:00,040
AI by Rob Cully, who's who's actually one of the ex-mokes of pns

96
00:05:00,040 --> 00:05:02,520
in the Excel, Power BI team around this time,

97
00:05:02,520 --> 00:05:04,920
actually when Power BI was coming out.

98
00:05:04,920 --> 00:05:08,320
And in here, he's talking about the two different types of computing,

99
00:05:08,320 --> 00:05:13,080
like CPU, computing and AI computing.

100
00:05:13,080 --> 00:05:14,280
And he's really,

101
00:05:14,280 --> 00:05:16,960
comparing and contrasting actually,

102
00:05:16,960 --> 00:05:20,800
they're both completely different in terms of their strengths and weaknesses.

103
00:05:20,800 --> 00:05:25,000
So, you know, Gen AI is really clever,

104
00:05:25,000 --> 00:05:27,760
but it's also quite clumsy in the sense of,

105
00:05:27,760 --> 00:05:30,120
you wouldn't trust it to really come up with your numbers

106
00:05:30,120 --> 00:05:31,840
that you need to be reliable every month

107
00:05:31,840 --> 00:05:35,120
and grounded in your business processes.

108
00:05:35,120 --> 00:05:38,080
But that's exactly what you need to delegate to the other type of computing,

109
00:05:38,080 --> 00:05:40,760
which is still there, CPU, traditional computing, right?

110
00:05:40,760 --> 00:05:42,880
It's, it's deterministic.

111
00:05:42,880 --> 00:05:46,080
It's, you know, it's great at doing those calculations,

112
00:05:46,080 --> 00:05:47,560
of working, connecting to your sources,

113
00:05:47,560 --> 00:05:49,160
and understanding your data.

114
00:05:49,160 --> 00:05:51,000
That should be the grounding, you know,

115
00:05:51,000 --> 00:05:55,200
it's, I think you called it, reliable, but rigid, right?

116
00:05:55,200 --> 00:05:56,840
And how do you get the best of both worlds?

117
00:05:56,840 --> 00:05:58,320
How do you combine them?

118
00:05:58,320 --> 00:06:01,080
And that's the real essence, I think, of where we are right now.

119
00:06:01,080 --> 00:06:04,920
How do we combine, you know, deterministic,

120
00:06:04,920 --> 00:06:09,880
you know, structured data and logic that calculates our financial numbers,

121
00:06:09,880 --> 00:06:12,600
that calculates, you know, all of our kind of core KPIs

122
00:06:12,600 --> 00:06:16,200
and metrics in a business with a kind of single version of the truth?

123
00:06:16,200 --> 00:06:20,680
How do we combine that with the analysis capability,

124
00:06:20,680 --> 00:06:24,120
but also the cleverness of AI in terms of the flexibility?

125
00:06:24,120 --> 00:06:28,120
Because what ends up happening is you have, you know,

126
00:06:28,120 --> 00:06:31,120
processes that are run in a very rigid fashion.

127
00:06:31,120 --> 00:06:32,720
And so you want to try and automate things

128
00:06:32,720 --> 00:06:37,400
when you end up writing with VBA macros or RPA and how automate those.

129
00:06:37,400 --> 00:06:39,320
And you know, even Power BI reports.

130
00:06:39,320 --> 00:06:43,960
And before you know it, you've got hundreds of these solutions to try and manage.

131
00:06:43,960 --> 00:06:48,880
And you build technical depth, where us with AI,

132
00:06:48,880 --> 00:06:52,280
actually, it can reason over those things in flexible ways.

133
00:06:52,280 --> 00:06:55,160
So maybe you have far fewer artifacts,

134
00:06:55,160 --> 00:06:57,520
and actually they can flex to different scenarios

135
00:06:57,520 --> 00:06:59,520
and different customers and different types of data.

136
00:06:59,520 --> 00:07:02,280
You don't need to set up a separate spreadsheet process

137
00:07:02,280 --> 00:07:04,400
or set for Power BI reports or a set for a thing

138
00:07:04,400 --> 00:07:09,000
that every single type of variance that you have around the type of data

139
00:07:09,000 --> 00:07:10,600
and the type of process that it works with.

140
00:07:10,600 --> 00:07:15,560
So you combine, how do you combine the data and the flexibility of the AI?

141
00:07:15,560 --> 00:07:16,960
I think that's where we are now.

142
00:07:16,960 --> 00:07:18,720
And that's what's excited about it.

143
00:07:18,720 --> 00:07:20,800
So yeah, yeah.

144
00:07:20,800 --> 00:07:23,080
But let's start with the financial problem.

145
00:07:23,080 --> 00:07:28,040
So let's imagine I'm the CEO of a large company,

146
00:07:28,040 --> 00:07:29,320
I hope, but not for them.

147
00:07:29,320 --> 00:07:31,280
I get to give also their bankrupt.

148
00:07:31,280 --> 00:07:35,400
OK, I have SAP or another ERP system.

149
00:07:35,400 --> 00:07:36,480
I have Power BI.

150
00:07:36,480 --> 00:07:38,720
I have Excel everywhere.

151
00:07:38,720 --> 00:07:41,960
Maybe we are implementing Microsoft Fabric.

152
00:07:41,960 --> 00:07:44,680
And now someone walks into my office and says,

153
00:07:44,680 --> 00:07:45,520
great news.

154
00:07:45,520 --> 00:07:48,160
We have by an agent that can answer,

155
00:07:48,160 --> 00:07:50,080
glass to the body of financial support.

156
00:07:50,080 --> 00:07:51,800
Super power.

157
00:07:51,800 --> 00:07:56,160
So what's the first question you will ask?

158
00:07:59,080 --> 00:08:01,920
How are documented are your processes?

159
00:08:01,920 --> 00:08:04,200
I asked if that was Christian or at odds.

160
00:08:04,200 --> 00:08:09,960
And for my organization, it's not, is that knowledge exists in people's heads?

161
00:08:09,960 --> 00:08:13,120
And it's embedded within the logic in their Excel formulas

162
00:08:13,120 --> 00:08:16,880
or in their Power BI reports or in their source systems.

163
00:08:16,880 --> 00:08:19,880
Maybe some of that notchics in the ERP, some of that notchics in Excel,

164
00:08:19,880 --> 00:08:24,880
some of that notchics in Power BI, some of that notchics in the SharePoint files

165
00:08:24,880 --> 00:08:28,280
that ultimately come to it, or at least some of the context is in there.

166
00:08:28,280 --> 00:08:30,840
So everything is scattered all over the place.

167
00:08:30,840 --> 00:08:34,840
And what we've done up until now is kind of build and say there's point solutions

168
00:08:34,840 --> 00:08:36,640
to deal with every specific thing, right?

169
00:08:36,640 --> 00:08:37,960
We need to do this financial analysis.

170
00:08:37,960 --> 00:08:42,560
We need to produce this report each month and let's go and build a solution

171
00:08:42,560 --> 00:08:45,720
to that connected sub data, model it, structure it,

172
00:08:45,720 --> 00:08:49,160
put our logic in there and then have that as a Power BI report,

173
00:08:49,160 --> 00:08:51,360
have that in Excel spreadsheet.

174
00:08:51,360 --> 00:08:55,760
OK, then you do that for every single process, every single time.

175
00:08:55,760 --> 00:09:01,120
So now let's take a step back and say, look, if we were doing this right from scratch,

176
00:09:01,120 --> 00:09:06,640
let's forget we have any company to get re-hide, let's take a moment, right?

177
00:09:06,640 --> 00:09:11,000
We just have our systems, we have our technical infrastructure is,

178
00:09:11,000 --> 00:09:14,680
we have that all this and it, but how would we redesign this and the grounds up?

179
00:09:14,680 --> 00:09:15,800
How would we think about this?

180
00:09:15,800 --> 00:09:18,360
And it all starts with that business context.

181
00:09:18,360 --> 00:09:20,880
It all starts with that thinking, OK, what are we got?

182
00:09:20,880 --> 00:09:22,040
What processes do we need to run?

183
00:09:22,040 --> 00:09:23,080
What outputs do we need to do?

184
00:09:23,080 --> 00:09:25,400
Which decisions do we need to make?

185
00:09:25,400 --> 00:09:32,080
Once you start there, what decisions do you need to make every month, every week, every day?

186
00:09:32,080 --> 00:09:37,520
Once you start with the decisions you need to make and the business questions that you need answered,

187
00:09:37,520 --> 00:09:45,360
then you go from there into logic and into data and then you pass that to Gen AI

188
00:09:45,360 --> 00:09:49,600
because once it gets an understanding of the business insights and it's got an understanding

189
00:09:49,600 --> 00:09:53,440
of your data and this could just be some CSV files that you give Gen AI.

190
00:09:53,440 --> 00:09:56,120
It can go everywhere from there.

191
00:09:56,120 --> 00:10:01,560
So really I'd just, I'd want to understand how well documented are your processes,

192
00:10:01,560 --> 00:10:08,280
but actually my IG, the little bit to say how well documented are your business insights?

193
00:10:08,280 --> 00:10:10,840
Are your decisions are your personas?

194
00:10:10,840 --> 00:10:16,720
Like because what Gen AI really needs is to understand who the real people in the business are

195
00:10:16,720 --> 00:10:20,640
and what they're interested, what their pain points are, what keeps them awake at night.

196
00:10:20,640 --> 00:10:23,160
That's what Gen AI really needs.

197
00:10:23,160 --> 00:10:30,240
So I've given like a brain dump of that and I've just actually, I'm doing an event next week

198
00:10:30,240 --> 00:10:35,480
and I'm going to show an example at the end where we've got, it's for a retail company.

199
00:10:35,480 --> 00:10:41,120
It is one in the UK called Mums and Puppas and I met somebody who works there at a conference

200
00:10:41,120 --> 00:10:45,920
and I bought all my kids stuff from there, so I never company.

201
00:10:45,920 --> 00:10:51,120
I've just downloaded their financial statements from company's house

202
00:10:51,120 --> 00:10:55,920
and I asked Claude to generate a general ledger like I gave it a zero, a child of accounts.

203
00:10:55,920 --> 00:10:59,760
And so it generate me a whole general ledger for this and then it can be a general ledger

204
00:10:59,760 --> 00:11:01,240
that tied up to their accounts.

205
00:11:01,240 --> 00:11:04,720
So actually it's probably fairly realistic and then it gave me all the dimensions that

206
00:11:04,720 --> 00:11:10,160
they need staff, you know, what, they can use stores and real stores, suppliers, real suppliers

207
00:11:10,160 --> 00:11:13,760
that they have because it knows their stores, it has their suppliers with their website.

208
00:11:13,760 --> 00:11:17,240
So again, the order is the data, but now I have this data set.

209
00:11:17,240 --> 00:11:23,080
And what I did is passed it into Gen AI, like just the CSV files and I passed it like a page

210
00:11:23,080 --> 00:11:27,080
or two of kind of descriptions of the kind of things that someone would be interested

211
00:11:27,080 --> 00:11:32,440
into analyzes data, you know, the kind of, and it was quite broad, you know, it was data

212
00:11:32,440 --> 00:11:37,080
on, you know, inventory management and, you know, star feed rate as even or, you know,

213
00:11:37,080 --> 00:11:40,880
things like that, like just quite broad, but actually, you know, these are the kind of things

214
00:11:40,880 --> 00:11:45,080
that need to make, these are decisions, this person needs to make on a daily basis.

215
00:11:45,080 --> 00:11:49,880
And it went through a workflow that I've got kind of a genetic work though that basically

216
00:11:49,880 --> 00:11:53,800
built this gopet template, that summarized all the phase decisions and then it said, okay,

217
00:11:53,800 --> 00:11:58,320
what for each, for each area of analysis, who's the audience, what's their pain points,

218
00:11:58,320 --> 00:12:00,480
what's the business questions they need to answer?

219
00:12:00,480 --> 00:12:04,160
And then what's the kind of KPIs that logic that would help to answer those questions

220
00:12:04,160 --> 00:12:07,280
and what would you need to view and filter by it or the ones of them?

221
00:12:07,280 --> 00:12:11,320
And, you know, essentially what you do is to conceptually design a semantic model at

222
00:12:11,320 --> 00:12:12,320
a stage, right?

223
00:12:12,320 --> 00:12:13,320
You don't need to call it that.

224
00:12:13,320 --> 00:12:15,160
That's essentially what you do.

225
00:12:15,160 --> 00:12:19,800
And then you then generate, I take that input and then designs your semantic model of that

226
00:12:19,800 --> 00:12:23,640
designs your facts and dimensions as that and designs all the measures and gives you all

227
00:12:23,640 --> 00:12:26,160
of those to approve and say, yeah, this is good.

228
00:12:26,160 --> 00:12:27,680
Then it goes, goes, it builds it.

229
00:12:27,680 --> 00:12:31,880
It goes, it builds you a whole power of your semantic model using power of your MCB server.

230
00:12:31,880 --> 00:12:32,880
Okay, right?

231
00:12:32,880 --> 00:12:37,440
Now you, I've got, you know, 50, 60 measures all aligns to my business decisions, all align

232
00:12:37,440 --> 00:12:42,480
to my, you know, the pain points and all connected or working, right?

233
00:12:42,480 --> 00:12:47,440
This is, so this is way, way ahead of where we were with how VI before this, right?

234
00:12:47,440 --> 00:12:51,160
Where we were just kind of that queries could help you with personal productivity.

235
00:12:51,160 --> 00:12:55,360
Now we're taking business insights into data and it's building you false semantic models

236
00:12:55,360 --> 00:13:01,000
with measures and then it went to build a reports and then it went to build ontologies and

237
00:13:01,000 --> 00:13:04,280
platinum level lake houses because it was made well actually the semantic model can't

238
00:13:04,280 --> 00:13:08,600
answer all those questions that you had because it's aggregated data, it's, you know,

239
00:13:08,600 --> 00:13:13,080
tapio based, it doesn't tell you really why what's happening in your business.

240
00:13:13,080 --> 00:13:16,960
So then you know, it's just a gap analysis to say, well, what could we not answer this

241
00:13:16,960 --> 00:13:20,720
semantic model that we needed to know for our business decisions and then designed, you

242
00:13:20,720 --> 00:13:24,640
know, maybe that data's coming, maybe it's for me, our PCM, but maybe it's also from SharePoint

243
00:13:24,640 --> 00:13:28,400
to other places and it builds you that whole kind of platinum layer that's going to

244
00:13:28,400 --> 00:13:33,680
ask for the, the why they allow you to drill down and then it puts it all into an ontology

245
00:13:33,680 --> 00:13:37,640
and then the data agents, now the data agents connected to your semantic model to your,

246
00:13:37,640 --> 00:13:43,680
you know, to your lake houses, to your ontology, and then you put that just as one click into

247
00:13:43,680 --> 00:13:46,360
co-pilot studio and then you can ask it in teams.

248
00:13:46,360 --> 00:13:50,920
So someone could go into teams and ask any question or ask, run any process or anything

249
00:13:50,920 --> 00:13:54,360
because you've got workflows in co-pilot studio, you've got all of these things now.

250
00:13:54,360 --> 00:14:00,360
So that's the path that I think makes the most sense is, is public data agents into co-pilot

251
00:14:00,360 --> 00:14:01,360
studio.

252
00:14:01,360 --> 00:14:04,680
So coming back to your question, there's a log with the answer to the question, what

253
00:14:04,680 --> 00:14:10,720
will I ask the CFO is, yeah, how well documented have you got this, you know, context or can

254
00:14:10,720 --> 00:14:12,120
you document this context?

255
00:14:12,120 --> 00:14:17,360
You know, maybe we need to record some calls with his team or her team to go through their

256
00:14:17,360 --> 00:14:21,280
processes and then just pass those transcripts into Gen AI and get it to produce these subways

257
00:14:21,280 --> 00:14:22,280
of what people need.

258
00:14:22,280 --> 00:14:24,960
And it's not really more complicated than that, right?

259
00:14:24,960 --> 00:14:25,960
It really isn't.

260
00:14:25,960 --> 00:14:28,080
You give us a data, you give us some context.

261
00:14:28,080 --> 00:14:30,760
Hey, everything else could be built for you.

262
00:14:30,760 --> 00:14:39,720
Yeah, I think let's just think about something simple, like revenue or so.

263
00:14:39,720 --> 00:14:47,880
So there are different roles, I think, the CFO from the company, the regional edge of

264
00:14:47,880 --> 00:14:50,960
the sales, the marketing guy and so on.

265
00:14:50,960 --> 00:14:55,920
And they all ask what was our revenue last quarter?

266
00:14:55,920 --> 00:15:00,680
And normally they have their different systems and get their different answers.

267
00:15:00,680 --> 00:15:09,800
I think you, we can have the, yeah, the same KPI in different ways.

268
00:15:09,800 --> 00:15:16,280
Or short, we have one KPI that's the universal truth.

269
00:15:16,280 --> 00:15:17,280
Well, exactly.

270
00:15:17,280 --> 00:15:22,800
So that's where semantic models really come in because they are your single source of truth

271
00:15:22,800 --> 00:15:25,000
for those numbers, right?

272
00:15:25,000 --> 00:15:27,520
So you've got your governed enterprise level semantic models.

273
00:15:27,520 --> 00:15:30,040
Obviously, you've got the semantic models that people build off the side of the desk as

274
00:15:30,040 --> 00:15:35,800
well, that's great for a peerless, easy experimentation learning all the rest of that, right?

275
00:15:35,800 --> 00:15:40,720
But once it gets certified and you know, pull it up into the enterprise world, that's where

276
00:15:40,720 --> 00:15:43,240
then your semantic models are, that's what's the truth.

277
00:15:43,240 --> 00:15:47,880
And so when this is why you need that as a grandied for your agents, so you don't have this

278
00:15:47,880 --> 00:15:51,640
challenge where people are coming in with their own different versions of spreadsheets,

279
00:15:51,640 --> 00:15:53,040
cuts taking a different time.

280
00:15:53,040 --> 00:15:57,160
So therefore you're getting different numbers, some include these adjustments, some include

281
00:15:57,160 --> 00:15:58,160
these adjustments.

282
00:15:58,160 --> 00:16:01,280
You know, you'll always end up with different numbers and pick up.

283
00:16:01,280 --> 00:16:05,400
And yeah, there might be different numbers that you need to bring up to pendium on the context.

284
00:16:05,400 --> 00:16:08,520
So again, when someone's asking revenue, is it a sales manager?

285
00:16:08,520 --> 00:16:13,080
He's asking revenue and just should that include, you know, commissions or should it not include

286
00:16:13,080 --> 00:16:15,320
commissions like the ones that does depend.

287
00:16:15,320 --> 00:16:16,720
And that's okay.

288
00:16:16,720 --> 00:16:19,280
But what you need to do is tell it when it depends.

289
00:16:19,280 --> 00:16:23,760
And then you have skills that you encode into agents to tell it these kind of instructions

290
00:16:23,760 --> 00:16:27,520
to give it this kind of process to say, follow this methodology.

291
00:16:27,520 --> 00:16:31,920
If it's this person asking or if it's this context, go down this semantic model and maybe look

292
00:16:31,920 --> 00:16:34,160
at this and this is the kind of data that we're working on that.

293
00:16:34,160 --> 00:16:39,200
If it's this scenario, maybe go down this route and you're coding those instructions into

294
00:16:39,200 --> 00:16:40,400
a skill.

295
00:16:40,400 --> 00:16:46,120
So your data is coming from a ground source like fabric and semantic models, your instructions

296
00:16:46,120 --> 00:16:49,640
on how to work with those semantic models and how to interpret those semantic models and

297
00:16:49,640 --> 00:16:53,120
how to read the data is coming in skills.

298
00:16:53,120 --> 00:16:55,880
And that's like how we work as humans, right?

299
00:16:55,880 --> 00:16:58,960
That's, we have skills, we have abilities to do things.

300
00:16:58,960 --> 00:17:03,160
So I think we really need to start thinking about how we apply the human type roles and

301
00:17:03,160 --> 00:17:09,080
human type skills to agentech agents and treat them like digital employees, like performance

302
00:17:09,080 --> 00:17:14,440
management, give them, give them, give them clear guardrails, give them clear instructions,

303
00:17:14,440 --> 00:17:18,400
give them clear performance metrics in terms of what goes like, so give them access to data

304
00:17:18,400 --> 00:17:19,400
and tools.

305
00:17:19,400 --> 00:17:23,600
And then let them come up with, with the wide answers.

306
00:17:23,600 --> 00:17:32,440
Yeah, I like this, this context, context layer stuff, but I think where should this business

307
00:17:32,440 --> 00:17:39,080
context actually live, should it inside the power BI, inside the semantic model or in the

308
00:17:39,080 --> 00:17:44,440
meta data, like peer view, or whether in the on the, notology?

309
00:17:44,440 --> 00:17:46,240
Yeah, why is it mixed?

310
00:17:46,240 --> 00:17:50,520
It's a mix and it does, you know, it depends on what the context is needed.

311
00:17:50,520 --> 00:17:57,240
So as I say, semantic models are a brilliant source for AI because it has those calculations

312
00:17:57,240 --> 00:18:00,600
and, you know, the logic that really represents your core business.

313
00:18:00,600 --> 00:18:03,320
But it's not the early source.

314
00:18:03,320 --> 00:18:09,520
So when I, when I talk about the organizational brain, there's five layers of this organizational

315
00:18:09,520 --> 00:18:10,520
brain.

316
00:18:10,520 --> 00:18:13,720
So the first is data and that's your kind of source system, your schemers, you know, where

317
00:18:13,720 --> 00:18:17,520
your data is coming from and how it looks like, what it looks like in your, in your certified

318
00:18:17,520 --> 00:18:20,120
kind of hub workspaces if you like, maybe, yeah.

319
00:18:20,120 --> 00:18:22,120
And then you've got logic.

320
00:18:22,120 --> 00:18:27,000
This is your, again, your certified C core, Python, DAX, the logic that it needs to be able

321
00:18:27,000 --> 00:18:30,080
to transport, Aitor and work in a different ways.

322
00:18:30,080 --> 00:18:31,400
You then got tools.

323
00:18:31,400 --> 00:18:36,280
So tools are just any tool and it, you know, could be any MCP servers to connect into your

324
00:18:36,280 --> 00:18:37,280
systems.

325
00:18:37,280 --> 00:18:41,680
It could be a semantic model is a tool because it could be query, that's a tool.

326
00:18:41,680 --> 00:18:43,160
And the tool is a tool.

327
00:18:43,160 --> 00:18:45,520
So it could be tools by any tools that you've got.

328
00:18:45,520 --> 00:18:49,720
Then you've got skills and that's both utility skills or the skills that Microsoft make

329
00:18:49,720 --> 00:18:52,960
available on how to work with tools and platforms and things like that.

330
00:18:52,960 --> 00:18:57,000
And you know, copilot studio has got a whole bunch of skills and agents and plugins and

331
00:18:57,000 --> 00:18:58,000
all of that.

332
00:18:58,000 --> 00:19:00,360
But also your custom skills.

333
00:19:00,360 --> 00:19:04,560
So your business processes get turned into custom skills that live in this layer.

334
00:19:04,560 --> 00:19:10,480
And then you've got governance, which is things like, you know, your access, commissions,

335
00:19:10,480 --> 00:19:13,200
your policies, all the things that needs to do.

336
00:19:13,200 --> 00:19:17,200
Now with those kind of anything that you can think of can be fought across those five

337
00:19:17,200 --> 00:19:18,200
layers.

338
00:19:18,200 --> 00:19:21,920
So you take a business process and it has some data that it connects to it, has some tools

339
00:19:21,920 --> 00:19:23,400
that it needs to connect into those.

340
00:19:23,400 --> 00:19:28,080
It has some skills on how to work with those tools, how to interpret that data, how to calculate

341
00:19:28,080 --> 00:19:30,440
the right things or where to go.

342
00:19:30,440 --> 00:19:36,320
And plus it has, you know, governance, right, in terms of how it needs to be accessed.

343
00:19:36,320 --> 00:19:38,400
And that's logic, obviously.

344
00:19:38,400 --> 00:19:42,720
So I think it's a combination of all of those five layers that work together.

345
00:19:42,720 --> 00:19:44,840
And that's what provides the overall grounding.

346
00:19:44,840 --> 00:19:47,040
So this is something you could build is something like Foundry IQ.

347
00:19:47,040 --> 00:19:51,200
I think that if you have you come across it, have you worked with that tool?

348
00:19:51,200 --> 00:19:52,200
Yeah.

349
00:19:52,200 --> 00:19:53,200
Yeah.

350
00:19:53,200 --> 00:19:54,200
Yeah.

351
00:19:54,200 --> 00:19:58,440
That's what we fought for years ago.

352
00:19:58,440 --> 00:20:02,040
I did find that IQ exists four years ago.

353
00:20:02,040 --> 00:20:03,040
I think.

354
00:20:03,040 --> 00:20:04,040
Really?

355
00:20:04,040 --> 00:20:05,040
Oh, okay.

356
00:20:05,040 --> 00:20:06,040
Let me think.

357
00:20:06,040 --> 00:20:08,040
I'm not a lot of great.

358
00:20:08,040 --> 00:20:09,040
I'm not so.

359
00:20:09,040 --> 00:20:14,320
I mean, it's just a bug search index over over over your knowledge sources, right?

360
00:20:14,320 --> 00:20:20,440
So in a sense, it's been around for a lot of, I said a long time lag hasn't been around

361
00:20:20,440 --> 00:20:22,840
for that long, but yeah, it's been around for a while.

362
00:20:22,840 --> 00:20:24,800
But I think it could sexually it's there.

363
00:20:24,800 --> 00:20:28,280
And I think it's all just plumbing this stuff together.

364
00:20:28,280 --> 00:20:30,960
And again, Gen AI can help you do that.

365
00:20:30,960 --> 00:20:36,560
So I'm trying to look at how can we take existing processes like existing Excel spreadsheets

366
00:20:36,560 --> 00:20:43,080
and convert that into a custom skill in the organizational grain because it can take

367
00:20:43,080 --> 00:20:48,040
an Excel spreadsheet and it can look at the data, look at things that are static data and

368
00:20:48,040 --> 00:20:51,920
whether they're as inputs, you can look at your Excel formulas, which is the logic and

369
00:20:51,920 --> 00:20:56,320
it can look at your outputs, which is whether the results land.

370
00:20:56,320 --> 00:21:01,480
And it can then kind of work with you to understand that process that's happening and then

371
00:21:01,480 --> 00:21:06,000
convert that into fabric lake houses, convert that into semantic models, convert that into

372
00:21:06,000 --> 00:21:11,080
reports and ontologies and data agents and co-pilot studio agents and things like that

373
00:21:11,080 --> 00:21:13,120
in a way that makes sense.

374
00:21:13,120 --> 00:21:16,960
And you don't know how to know how to build all of these things individually yourself.

375
00:21:16,960 --> 00:21:19,160
You don't have to know all these tools inside out.

376
00:21:19,160 --> 00:21:23,480
You don't have to spend five years to become a data engineer and then an AI engineer who

377
00:21:23,480 --> 00:21:28,360
knows boundary and all of this because the agents can do all of that stuff for you as long

378
00:21:28,360 --> 00:21:31,600
as you give it the right context and as long as you get it to communicating the right

379
00:21:31,600 --> 00:21:32,600
way with you.

380
00:21:32,600 --> 00:21:35,680
And I asked the thing I'm struggling with the most is that if I'm honest at the moment,

381
00:21:35,680 --> 00:21:38,360
AI doesn't communicate very well.

382
00:21:38,360 --> 00:21:41,040
It likes to just jump ahead to do everything and tell you, hey, I've done it.

383
00:21:41,040 --> 00:21:42,040
All.

384
00:21:42,040 --> 00:21:45,600
And you know, when you try to get it to the development stuff, hey, should I run this

385
00:21:45,600 --> 00:21:46,600
the end of environment?

386
00:21:46,600 --> 00:21:48,600
Should I run this bash command?

387
00:21:48,600 --> 00:21:50,800
And you're like, I have no idea what you're talking about.

388
00:21:50,800 --> 00:21:52,280
I didn't know what the bash command is.

389
00:21:52,280 --> 00:21:55,520
You haven't told me that you're writing the bash command of what it's going to do.

390
00:21:55,520 --> 00:21:57,280
All you're doing is asking me for approval.

391
00:21:57,280 --> 00:21:59,440
It's like, come on.

392
00:21:59,440 --> 00:22:00,440
Yeah, really.

393
00:22:00,440 --> 00:22:04,240
Like, I know you want to show off that you clever, but we've gone past that stage.

394
00:22:04,240 --> 00:22:07,960
You know, I know I want you to communicate with me like a human, right?

395
00:22:07,960 --> 00:22:10,920
I want you to communicate and tell me what you're doing and tell me how I get

396
00:22:10,920 --> 00:22:12,600
what you need from me.

397
00:22:12,600 --> 00:22:15,600
That's where AI needs to come scale right back down.

398
00:22:15,600 --> 00:22:17,080
And that's not about intelligence.

399
00:22:17,080 --> 00:22:19,400
You know, ashtras not going to fix that.

400
00:22:19,400 --> 00:22:21,240
Vables not going to fix that.

401
00:22:21,240 --> 00:22:22,640
It doesn't matter how intelligent it gets.

402
00:22:22,640 --> 00:22:25,400
In fact, the more intelligent it gets, the worse it gets.

403
00:22:25,400 --> 00:22:30,000
The worse it communicates because it's now things it doesn't need to listen to you anymore.

404
00:22:30,000 --> 00:22:31,920
It can go off into everything itself.

405
00:22:31,920 --> 00:22:33,600
So that's the biggest risk, I see with AI.

406
00:22:33,600 --> 00:22:37,040
And that's the biggest thing I think I'd need to solve to be able to get this whole

407
00:22:37,040 --> 00:22:39,560
solution working.

408
00:22:39,560 --> 00:22:46,160
I think, yeah, for the people from Power BI, they know the data model, the semantic model,

409
00:22:46,160 --> 00:22:52,720
and now with fabric and AI, we have something new, I say, the onotology.

410
00:22:52,720 --> 00:22:56,360
How will you explain these three things?

411
00:22:56,360 --> 00:23:04,840
Are the data models a semantic model in the actuality to someone with new and, yeah.

412
00:23:04,840 --> 00:23:05,840
Okay, yeah.

413
00:23:05,840 --> 00:23:07,840
Let me give you a try.

414
00:23:07,840 --> 00:23:13,080
If it's not clear, let me know and I'll read you the information.

415
00:23:13,080 --> 00:23:17,680
So data model is basically just a structure of your tables.

416
00:23:17,680 --> 00:23:24,800
So if you look at transactional systems like at ERP, CRM, those tables are all scattered

417
00:23:24,800 --> 00:23:25,800
around.

418
00:23:25,800 --> 00:23:29,200
If you look at the say something like SAP's, hundreds of tables that have just got all

419
00:23:29,200 --> 00:23:30,920
different connections.

420
00:23:30,920 --> 00:23:37,360
And that data structure of tables works really well for running a transactional system.

421
00:23:37,360 --> 00:23:42,360
Because you've got no redundancy, everything's in only one place, there's no repetition

422
00:23:42,360 --> 00:23:43,360
anywhere.

423
00:23:43,360 --> 00:23:47,720
So everything's just connected and it's very good for reading the writing single records.

424
00:23:47,720 --> 00:23:53,040
It's an absolute nightmare if you try and use that data model in Power BI.

425
00:23:53,040 --> 00:23:58,080
Because suddenly, to do a simple DAX query, it has to do 50 joins in the back between all

426
00:23:58,080 --> 00:24:02,400
these tables and try and understand how they all relate and your DAX becomes stupidly complex

427
00:24:02,400 --> 00:24:05,680
because it has to then pick out all these tables at different places.

428
00:24:05,680 --> 00:24:10,360
So you need to denormalize and actually turn it from an OLTP or a transaction processor

429
00:24:10,360 --> 00:24:16,240
model, data model into a low lap, tabular model, which is just dimensions and facts.

430
00:24:16,240 --> 00:24:18,160
And there is with that repetition and redundancy.

431
00:24:18,160 --> 00:24:19,800
And that's OK.

432
00:24:19,800 --> 00:24:23,520
Because it's just a structure that Power BI could work within the Kineco get open and

433
00:24:23,520 --> 00:24:24,520
analyze.

434
00:24:24,520 --> 00:24:25,520
So that's data model.

435
00:24:25,520 --> 00:24:28,320
And you know, you've got different types of data model, but that's your physical data

436
00:24:28,320 --> 00:24:29,320
model.

437
00:24:29,320 --> 00:24:30,760
And you know, you do it through Kimble.

438
00:24:30,760 --> 00:24:33,880
So Kimble is the kind of methodology, what if you data model?

439
00:24:33,880 --> 00:24:34,880
Right.

440
00:24:34,880 --> 00:24:36,760
And then he's been said semantic model.

441
00:24:36,760 --> 00:24:41,360
And again, obviously data model is called a semantic model, but OLTP, the structure is for

442
00:24:41,360 --> 00:24:42,360
semantic model.

443
00:24:42,360 --> 00:24:46,720
That's the data structure you need or is ideally what's to do.

444
00:24:46,720 --> 00:24:50,880
And then you have your measures, your logic and very little security permissions and all

445
00:24:50,880 --> 00:24:53,440
these other things that you can add into a semantic model.

446
00:24:53,440 --> 00:24:56,280
So now this becomes a kind of asset.

447
00:24:56,280 --> 00:25:01,040
And I think this is the real interesting way to think about these things as assets, right?

448
00:25:01,040 --> 00:25:06,120
You've got a data model, which is foundational, kind of asset, you've got semantic model,

449
00:25:06,120 --> 00:25:10,240
which now contains logic and stuff, which is another type of mail assets.

450
00:25:10,240 --> 00:25:12,480
And then you've got, as I said before, right?

451
00:25:12,480 --> 00:25:17,640
You've got semantic models are great for those KPIs and measures, but AI needs more than

452
00:25:17,640 --> 00:25:18,640
that.

453
00:25:18,640 --> 00:25:23,480
AI needs to understand not just what the numbers were and what you were porting was over

454
00:25:23,480 --> 00:25:25,480
time, but why?

455
00:25:25,480 --> 00:25:29,640
And actually, it's not, it could tell you that your cache was lower this month, but

456
00:25:29,640 --> 00:25:33,280
in order to know why your cache is lower this month, it probably needs to go into your customer

457
00:25:33,280 --> 00:25:37,120
contracts and identify that, you know, the terms have changed for one of your big suppliers

458
00:25:37,120 --> 00:25:41,760
and, you know, actually, you've had to pay out earlier than you expected or when your customers

459
00:25:41,760 --> 00:25:45,080
and they, you know, they're not paying the same period or something.

460
00:25:45,080 --> 00:25:48,400
It's not going to know that from the Power BI report because Power BI doesn't have that

461
00:25:48,400 --> 00:25:49,400
level of granularity.

462
00:25:49,400 --> 00:25:54,360
You wouldn't put that net of granularity or data into your semantic model.

463
00:25:54,360 --> 00:25:57,200
So that's where then the idea of nontology comes in.

464
00:25:57,200 --> 00:25:59,360
nontology can connect to lake houses.

465
00:25:59,360 --> 00:26:04,360
It can use semantic models to start endpoint, but your relationships are also less rigid.

466
00:26:04,360 --> 00:26:07,560
It's not a single table to table, single relationship.

467
00:26:07,560 --> 00:26:09,360
It's verb based relationships.

468
00:26:09,360 --> 00:26:14,040
So it's saying, well, actually, you've got stores which sell products, which have employees,

469
00:26:14,040 --> 00:26:17,080
which, you know, are located in a region.

470
00:26:17,080 --> 00:26:20,440
But then all of these things can hold true at the same time.

471
00:26:20,440 --> 00:26:25,480
And it's up to AI to then use this kind of nerve based language to almost understand

472
00:26:25,480 --> 00:26:26,800
the business.

473
00:26:26,800 --> 00:26:31,840
So business processes and entities are actually really well represented in nontology, especially

474
00:26:31,840 --> 00:26:34,040
for AI, which needs context.

475
00:26:34,040 --> 00:26:38,640
And it could be a little bit restrictive to just use a semantic model.

476
00:26:38,640 --> 00:26:43,360
So semantic models great for your KPI logic data models, and see what your foundation

477
00:26:43,360 --> 00:26:48,240
of going from transactional to Kimball style tap into models.

478
00:26:48,240 --> 00:26:52,440
And then you got ontologies that can have a different lens in your business.

479
00:26:52,440 --> 00:26:57,400
And actually can be connected to what grandly the lake house data or different data, real

480
00:26:57,400 --> 00:26:58,760
time data even, right?

481
00:26:58,760 --> 00:27:02,280
But then go to a reason over your business in that way.

482
00:27:02,280 --> 00:27:10,280
Yeah, I think a little bit about the model, do is the model need to understand concept

483
00:27:10,280 --> 00:27:19,000
like product, customer region, cost center, revenue, map margin, and especially the relationships

484
00:27:19,000 --> 00:27:20,000
between them.

485
00:27:20,000 --> 00:27:21,000
And how?

486
00:27:21,000 --> 00:27:22,000
Exactly.

487
00:27:22,000 --> 00:27:27,280
So we've got the logic based relationships, if you like, which is what your semantic model

488
00:27:27,280 --> 00:27:28,280
has.

489
00:27:28,280 --> 00:27:32,760
It says, take this field, well, it's to this field and to do this calculation, you take the

490
00:27:32,760 --> 00:27:35,600
number from this field and divide it by the number from this field.

491
00:27:35,600 --> 00:27:36,600
Okay, great.

492
00:27:36,600 --> 00:27:40,280
That's the layer you need to be able to calculate business logic because you don't want it

493
00:27:40,280 --> 00:27:41,280
to be lucid.

494
00:27:41,280 --> 00:27:44,880
That's, you need that to be quite tight, right?

495
00:27:44,880 --> 00:27:48,880
Now when you go into AI and asking for context and asking why things happen, and all

496
00:27:48,880 --> 00:27:50,360
of that, you needed to be lucid.

497
00:27:50,360 --> 00:27:51,360
You actually do need it.

498
00:27:51,360 --> 00:27:54,880
You want to be able to have your KPIs and metrics and know, but then you start asking,

499
00:27:54,880 --> 00:27:55,880
well, why?

500
00:27:55,880 --> 00:27:56,880
Why is my revenue doubtment?

501
00:27:56,880 --> 00:28:00,840
Why is, you know, why is it down in this region or what's happened here?

502
00:28:00,840 --> 00:28:05,720
And that's where it needs to start to understand those entities in a slightly different way,

503
00:28:05,720 --> 00:28:10,680
in a less rigid way, in a way that can start to explore, you know, the different relationships

504
00:28:10,680 --> 00:28:14,200
that might exist and you know, should exist between anything.

505
00:28:14,200 --> 00:28:18,200
So, you know, in the semantic model, you would know, it'd be very hard to write a formula

506
00:28:18,200 --> 00:28:21,560
that does all of these different things and has all these relationships, but an ontology

507
00:28:21,560 --> 00:28:23,640
is better suited to that.

508
00:28:23,640 --> 00:28:27,280
So, I think this update is not a single short answer.

509
00:28:27,280 --> 00:28:31,880
I think it's a combination of these things, but these context also needs to live together.

510
00:28:31,880 --> 00:28:34,960
This is a status, skills, tools, governance.

511
00:28:34,960 --> 00:28:37,440
All of these things need to kind of coexist.

512
00:28:37,440 --> 00:28:42,520
And you kind of need to start building this before you really can take advantage of AI

513
00:28:42,520 --> 00:28:44,520
beyond personal productivity.

514
00:28:44,520 --> 00:28:49,000
And if you want to use AI beyond kind of, you know, sub-rises, email, or even, you know,

515
00:28:49,000 --> 00:28:52,200
generate, you know, especially generate, you know, this presentation.

516
00:28:52,200 --> 00:28:56,600
Even if you start to turn into skills, you still have the challenge of multiple skills.

517
00:28:56,600 --> 00:29:01,040
You know, two different people have built the same kind of skill, different, different

518
00:29:01,040 --> 00:29:05,600
context, different data, different ideas, you know, how do you manage that?

519
00:29:05,600 --> 00:29:07,080
You still get technical debt.

520
00:29:07,080 --> 00:29:08,840
So, it's not just about skills.

521
00:29:08,840 --> 00:29:13,040
It's then about the broader context there is the goal of the land as well.

522
00:29:13,040 --> 00:29:14,040
Yeah.

523
00:29:14,040 --> 00:29:20,640
I think a little bit about there was before the co-pilot joins the Power BI.

524
00:29:20,640 --> 00:29:22,680
There was also a function.

525
00:29:22,680 --> 00:29:24,440
You can ask questions in Power BI.

526
00:29:24,440 --> 00:29:25,440
Q&A, yeah.

527
00:29:25,440 --> 00:29:26,440
Q&A, yeah, the Q&A function.

528
00:29:26,440 --> 00:29:34,760
I think it's a bit difficult to say for co-pilot.

529
00:29:34,760 --> 00:29:38,320
What would you say from this function to a co-pilot function?

530
00:29:38,320 --> 00:29:40,320
What is the big picture?

531
00:29:40,320 --> 00:29:44,520
They're model in each other.

532
00:29:44,520 --> 00:29:49,340
Q&A feature actually had this thing called a linguistic schema, which if it's their

533
00:29:49,340 --> 00:29:53,360
existence, their existence in Tindall, Power BI, semantic model.

534
00:29:53,360 --> 00:29:54,600
You can see the linguistic schema.

535
00:29:54,600 --> 00:29:57,400
The linguistic schema is exactly what we're talking about with our technologies.

536
00:29:57,400 --> 00:29:58,400
It was based relationships.

537
00:29:58,400 --> 00:30:03,280
And things like, ah, in fact, layer on top of your semantic model that kind of tells

538
00:30:03,280 --> 00:30:08,160
things how entities relate or when someone asks through this field, how does that bind

539
00:30:08,160 --> 00:30:09,160
into data?

540
00:30:09,160 --> 00:30:11,600
Like, it was very fitting.

541
00:30:11,600 --> 00:30:14,480
It wasn't a great experience to build it and it was very hidden.

542
00:30:14,480 --> 00:30:16,000
Like, put it into this linguistic schema.

543
00:30:16,000 --> 00:30:20,640
There's no way to extract that out and manage it at a corporate level and things like that.

544
00:30:20,640 --> 00:30:23,080
So that's the problem that ontology solves.

545
00:30:23,080 --> 00:30:26,080
You couldn't put AI, I don't think, on top of the linguistic schema, Tindall.

546
00:30:26,080 --> 00:30:27,080
I don't know.

547
00:30:27,080 --> 00:30:30,560
I mean, you could try it if you've still got any other models of an old Power BI model.

548
00:30:30,560 --> 00:30:31,560
That has a Q&A.

549
00:30:31,560 --> 00:30:33,040
Then even just try and see if AI can leverage it.

550
00:30:33,040 --> 00:30:34,040
I know.

551
00:30:34,040 --> 00:30:35,040
I've never tried it.

552
00:30:35,040 --> 00:30:41,880
I imagine ontology is going to be a far better solution to something like that.

553
00:30:41,880 --> 00:30:52,200
So, we have also, or, yeah, where do you say Microsoft fabric fit into this architecture?

554
00:30:52,200 --> 00:30:53,680
What did you think?

555
00:30:53,680 --> 00:30:58,320
What's the benefit of using Microsoft fabric?

556
00:30:58,320 --> 00:31:02,520
Well, it's your core ground, all your data and logic.

557
00:31:02,520 --> 00:31:08,480
It's no small part of plays in that tiny ecosystem.

558
00:31:08,480 --> 00:31:16,360
Even a small company that has a couple of systems and a dozen or two employees.

559
00:31:16,360 --> 00:31:23,880
Even then, they would benefit, I think, massively from spending $300 a month or less on fabric,

560
00:31:23,880 --> 00:31:27,880
getting all of that data into a single place, into a single name house that's governed,

561
00:31:27,880 --> 00:31:28,880
managed it.

562
00:31:28,880 --> 00:31:29,880
You can run SQL queries against it.

563
00:31:29,880 --> 00:31:31,360
You could put out the AI against it.

564
00:31:31,360 --> 00:31:35,080
You can put data agents against the AI.

565
00:31:35,080 --> 00:31:40,120
Just having that data logic foundation is so key.

566
00:31:40,120 --> 00:31:44,280
And the fact that they've then made it so easy to literally say, "Okay, add it as an agent

567
00:31:44,280 --> 00:31:49,080
into Copilot Studio and use the GitHub Copilot hardest now and Copilot Studio."

568
00:31:49,080 --> 00:31:52,000
To be able to reason over it as well.

569
00:31:52,000 --> 00:31:55,400
And then expose that into teams, or something like that.

570
00:31:55,400 --> 00:32:00,600
The fact that they've made that integration, and don't get me well, Microsoft have made

571
00:32:00,600 --> 00:32:05,560
so many mess ups with integrations and stuff and different teams working at different

572
00:32:05,560 --> 00:32:07,680
silo products that have not really talked into each other.

573
00:32:07,680 --> 00:32:13,280
But I mean, the data agent talking to Copilot Studio, I mean, that's genius, whoever came

574
00:32:13,280 --> 00:32:15,080
up with that.

575
00:32:15,080 --> 00:32:24,160
Yeah, I think a little bit makes it, when I have fabric, and fabric also have a semantic model,

576
00:32:24,160 --> 00:32:31,160
makes it then also sense to have two semantic models, one living in Power BI and one living

577
00:32:31,160 --> 00:32:32,160
in fabric.

578
00:32:32,160 --> 00:32:35,160
No, Power BI is part of fabric.

579
00:32:35,160 --> 00:32:36,160
There is no difference.

580
00:32:36,160 --> 00:32:37,160
There is no difference.

581
00:32:37,160 --> 00:32:39,160
Power BI is just one of the workloads in fabric now.

582
00:32:39,160 --> 00:32:41,160
It's not too separable.

583
00:32:41,160 --> 00:32:42,160
Okay, okay.

584
00:32:42,160 --> 00:32:52,920
I think what another part of what I find really interesting in fabric is it's the one

585
00:32:52,920 --> 00:32:54,920
like stuff.

586
00:32:54,920 --> 00:33:02,520
So I think we have years before we have worked and then we have this, we have to make all,

587
00:33:02,520 --> 00:33:03,920
yeah, all data structure.

588
00:33:03,920 --> 00:33:09,600
And now we have also this option to work with unstructured data with the work like, what

589
00:33:09,600 --> 00:33:14,160
did you think about the benefits, what's the biggest benefit?

590
00:33:14,160 --> 00:33:17,600
That's what I'm saying, even for a small company, right?

591
00:33:17,600 --> 00:33:20,640
Like it makes sense for them to bring their data into a lake house.

592
00:33:20,640 --> 00:33:23,800
And you can synchronize, I didn't, you know, with one lake, you can even synchronize it with

593
00:33:23,800 --> 00:33:24,800
your machine, right?

594
00:33:24,800 --> 00:33:25,800
Like one drive.

595
00:33:25,800 --> 00:33:29,600
See, you download the one lake file explorer and you see the Windows Explorer and you can

596
00:33:29,600 --> 00:33:33,720
just drag and drop files into there or save files into there directly from Excel or anything

597
00:33:33,720 --> 00:33:34,720
else.

598
00:33:34,720 --> 00:33:38,000
And then those get synchronized to the file section of the lake house.

599
00:33:38,000 --> 00:33:41,760
And then you can have things that file off there to actually convert that into delta or

600
00:33:41,760 --> 00:33:46,240
actually if they're just playing CSVs and, you know, it's a bit picky on it, but, you know,

601
00:33:46,240 --> 00:33:51,800
if it were structural works, it could convert them automatically as shortcuts into delta tables.

602
00:33:51,800 --> 00:33:55,840
And then suddenly, once they're in delta tables, once they're in delta format, everything

603
00:33:55,840 --> 00:33:57,160
houses just open to it.

604
00:33:57,160 --> 00:34:01,560
It's then that, you know, AI is open to it, sequels open to it, you know, real time intelligence

605
00:34:01,560 --> 00:34:06,360
cake, you're also open to it, you know, every work related fabric and every tool that could

606
00:34:06,360 --> 00:34:12,200
clearly use a sequel that takes a sequel endpoint, suddenly has this data available to it.

607
00:34:12,200 --> 00:34:15,640
And that's that's the power, including Excel, by the way, obviously.

608
00:34:15,640 --> 00:34:17,640
And that's how one lake like.

609
00:34:17,640 --> 00:34:21,400
Yeah, you live living in, in, in, in, in, right?

610
00:34:21,400 --> 00:34:22,400
Yeah.

611
00:34:22,400 --> 00:34:30,640
I think that, that, that, there are, there are so many lakes, we have actually one lake,

612
00:34:30,640 --> 00:34:33,080
a lake house and so on.

613
00:34:33,080 --> 00:34:35,880
It's, it's, feel a little bit like Scotland, right?

614
00:34:35,880 --> 00:34:36,880
But,

615
00:34:36,880 --> 00:34:37,880
I'm great.

616
00:34:37,880 --> 00:34:39,880
Actually, I'm going for Scottish Summit.

617
00:34:39,880 --> 00:34:44,720
In a, in a, in a few weeks, I've had you been once for like, carbonate.

618
00:34:44,720 --> 00:34:47,000
Actually, don't be really interested trip, but yes, you're right.

619
00:34:47,000 --> 00:34:52,840
There's, I think the scene you're in in the landscape and the lakes are, are, are, are supposed

620
00:34:52,840 --> 00:34:54,840
to be amazing in Scotland.

621
00:34:54,840 --> 00:34:55,840
Yeah.

622
00:34:55,840 --> 00:35:00,680
I'm a whiskey drinker as well, which helps, you know, and kind of, and it books the time to

623
00:35:00,680 --> 00:35:02,680
do whiskey tour after that.

624
00:35:02,680 --> 00:35:04,680
Yeah, yeah.

625
00:35:04,680 --> 00:35:08,520
So, some really good distil, you know, far worse, far worse.

626
00:35:08,520 --> 00:35:09,520
Yeah.

627
00:35:09,520 --> 00:35:11,000
It's a bit, it's a bit, I've got two and a half year olds.

628
00:35:11,000 --> 00:35:13,000
It's a bit harder with that, really, top.

629
00:35:13,000 --> 00:35:14,000
Yeah.

630
00:35:14,000 --> 00:35:17,160
I don't think, I don't think she's allowed in these distilries.

631
00:35:17,160 --> 00:35:21,240
And I'm going to, we'll, we'll opt out, Scotland would be.

632
00:35:21,240 --> 00:35:26,280
Yeah, but, but what, what, what's the, what's the, the direct lake?

633
00:35:26,280 --> 00:35:27,280
Direct lake.

634
00:35:27,280 --> 00:35:32,400
So, I mean, direct lake is, um, essentially queer, so, so tabular models.

635
00:35:32,400 --> 00:35:38,840
So it's analysis services, tabular model, um, which is obviously what, how pivot is, um,

636
00:35:38,840 --> 00:35:43,640
and it's, um, you know, your, your semantic models have been in tabular models.

637
00:35:43,640 --> 00:35:50,560
And now, the analysis service engine, what that does is basically takes your data, um,

638
00:35:50,560 --> 00:35:56,320
compresses it massively, you know, like 10 times or so, um, you know, and indexes, indexes

639
00:35:56,320 --> 00:35:58,280
stuff indexes your data, essentially.

640
00:35:58,280 --> 00:36:04,920
So, you know, every row, if you've got, you know, a row with 10 million rows of data,

641
00:36:04,920 --> 00:36:06,920
but it's just one and zero flag.

642
00:36:06,920 --> 00:36:07,920
It's like one bite.

643
00:36:07,920 --> 00:36:10,000
It compresses it to one bite as a dictionary, right?

644
00:36:10,000 --> 00:36:11,320
So one or zero, right?

645
00:36:11,320 --> 00:36:14,240
And so it's commonly, commonly made like based, yep.

646
00:36:14,240 --> 00:36:17,840
So actually the amount of compression you can achieve and therefore your performance,

647
00:36:17,840 --> 00:36:24,920
you can achieve a semantic model is, um, based on that compression technology, um, vertepac,

648
00:36:24,920 --> 00:36:29,480
which, um, determines how, um, how much it could compress your data and therefore how

649
00:36:29,480 --> 00:36:32,480
quickly it can, uh, analyze it and form over it.

650
00:36:32,480 --> 00:36:36,040
So that's vertepac that's been there, you know, sits in that list of services and, and,

651
00:36:36,040 --> 00:36:40,280
obviously, been enriched and, you know, with DAX and everything else that sits on top of

652
00:36:40,280 --> 00:36:41,280
that.

653
00:36:41,280 --> 00:36:49,200
So when Databricks came out with Delta for not, um, you know, I think Microsoft flipped

654
00:36:49,200 --> 00:36:51,360
about and said, well, how got a second?

655
00:36:51,360 --> 00:36:54,880
This is another compression technology that's calling them as well.

656
00:36:54,880 --> 00:37:00,760
Um, could we actually just build the entire vertepac engine into park against park a

657
00:37:00,760 --> 00:37:01,760
files?

658
00:37:01,760 --> 00:37:04,720
And then we don't need to actually do all the transformations.

659
00:37:04,720 --> 00:37:09,360
You don't need to go and convert your data like with the tabular vertepac engine and

660
00:37:09,360 --> 00:37:10,360
analysis services.

661
00:37:10,360 --> 00:37:13,680
And now the services is its own layer, right?

662
00:37:13,680 --> 00:37:19,800
So you need to convert data from whatever spreadsheets, CRM systems, whatever that data is

663
00:37:19,800 --> 00:37:24,680
coming into, you need to convert, bring it refresh your semantic model and bring your data

664
00:37:24,680 --> 00:37:25,680
into there.

665
00:37:25,680 --> 00:37:29,000
You could you die it query, but it does, that's, that's horrible.

666
00:37:29,000 --> 00:37:30,000
Okay.

667
00:37:30,000 --> 00:37:33,400
But if you're not, if you do an import mode, you need to refresh your semantic model,

668
00:37:33,400 --> 00:37:37,480
which will take all the data, bring it compression and bring it into the semantic model.

669
00:37:37,480 --> 00:37:42,720
But actually, if your data coming into this layer is in Delta format already in park a,

670
00:37:42,720 --> 00:37:46,840
which is park a files with a Delta log, then you don't need to do that anymore.

671
00:37:46,840 --> 00:37:51,880
You don't need to go and transform your data because Kaubi could read directly against

672
00:37:51,880 --> 00:37:58,080
those park a files and actually still has the same similar level of column, like compression

673
00:37:58,080 --> 00:37:59,840
and therefore performance.

674
00:37:59,840 --> 00:38:01,320
So actually that was the key thing.

675
00:38:01,320 --> 00:38:07,080
The key thing, the key reason why you needed to get your data into a import mode, vertipack

676
00:38:07,080 --> 00:38:08,600
semantic model was performance.

677
00:38:08,600 --> 00:38:13,280
But if you can get that performance directly off park a files, then all you need to do is get

678
00:38:13,280 --> 00:38:16,280
your Delta files, that's all you need to do is get your data to that format.

679
00:38:16,280 --> 00:38:17,640
So that's what diet query is.

680
00:38:17,640 --> 00:38:24,160
It's basically replicated vertipack into park a files to get against park a files instead.

681
00:38:24,160 --> 00:38:25,160
Yeah.

682
00:38:25,160 --> 00:38:27,140
And we have the next lake, Delta Lake.

683
00:38:27,140 --> 00:38:30,720
How can it, I need Delta Lake Delta Delta Delta Delta Delta, Delta just a format of one

684
00:38:30,720 --> 00:38:31,720
or eight or more.

685
00:38:31,720 --> 00:38:32,720
Yeah.

686
00:38:32,720 --> 00:38:33,720
The call, let's go ahead.

687
00:38:33,720 --> 00:38:39,080
Yeah, I read on your profile and you often also have a topic, it's a little bit different.

688
00:38:39,080 --> 00:38:42,120
It's the persona driven insights.

689
00:38:42,120 --> 00:38:47,480
Is this a different to or what's the difference to self service BI?

690
00:38:47,480 --> 00:38:50,600
I mean, it's the same, it's the same thing.

691
00:38:50,600 --> 00:38:55,440
I think, or self service BI or self service AI, it's also as AI as this idea that I'm kind

692
00:38:55,440 --> 00:39:01,600
of trying to kind of build thought leadership around, which is actually like self service BI

693
00:39:01,600 --> 00:39:06,040
in it, as they didn't really deliver as well as should have done because of the complexity

694
00:39:06,040 --> 00:39:07,800
of these tools as grown.

695
00:39:07,800 --> 00:39:13,720
Actually, can we bring that right back down to actually having an individual and they

696
00:39:13,720 --> 00:39:17,120
need to know a bit about data, they need to look good data looks like a day's quality,

697
00:39:17,120 --> 00:39:20,320
but you know, they don't need to spend years or the same before to become data engineers

698
00:39:20,320 --> 00:39:21,840
and bioblists.

699
00:39:21,840 --> 00:39:28,840
Can we enable them to actually work with tools like power BI and fabric and co-pilot studio

700
00:39:28,840 --> 00:39:36,040
and Excel to actually build the artifacts they need and but have agents do that where they

701
00:39:36,040 --> 00:39:37,040
provide a context.

702
00:39:37,040 --> 00:39:39,040
So that's the idea of self service AI.

703
00:39:39,040 --> 00:39:46,520
Now you said you're asking how that relates to relates to what we just talking about.

704
00:39:46,520 --> 00:39:47,520
Yeah.

705
00:39:47,520 --> 00:39:48,520
Yeah.

706
00:39:48,520 --> 00:39:51,720
So the idea then is actually, so the idea is that people can

707
00:39:51,720 --> 00:39:56,320
build with if agents have the right context of those developer agents, if you like, have

708
00:39:56,320 --> 00:40:03,320
the right context, then they can sit on top of that and grounding and could build you,

709
00:40:03,320 --> 00:40:06,720
somatic models, could build you reports, could build you ontologies, could build you data

710
00:40:06,720 --> 00:40:08,120
agents.

711
00:40:08,120 --> 00:40:11,320
What it needs to understand is the business context of a sonar driven insight.

712
00:40:11,320 --> 00:40:15,280
So say you ask about the sonar driven insights and that's exactly where it comes down to

713
00:40:15,280 --> 00:40:20,720
because what it really needs is that level of who is this for what business decisions

714
00:40:20,720 --> 00:40:22,480
are coming right back to where we started, right?

715
00:40:22,480 --> 00:40:24,440
What's the business decisions they need to make?

716
00:40:24,440 --> 00:40:30,680
What's the key questions they need to answer from data and the ones that knows that and

717
00:40:30,680 --> 00:40:33,520
it has a sample of data, everything else it can do for you.

718
00:40:33,520 --> 00:40:37,040
So, the sonar driven insights are your key starting point.

719
00:40:37,040 --> 00:40:40,840
So coming back to that question for the CFO, it really is actually, how well have you

720
00:40:40,840 --> 00:40:44,640
defined your sonar driven insights if we're going to ask the question in a nice six

721
00:40:44,640 --> 00:40:45,640
thing to it?

722
00:40:45,640 --> 00:40:48,240
Right, how well have you defined those presented driven insights?

723
00:40:48,240 --> 00:40:52,920
If you define those presented driven insights as well, especially in the context of data

724
00:40:52,920 --> 00:40:57,360
that you have available, then GDIA has what it is.

725
00:40:57,360 --> 00:41:01,840
And no, no, we have bring a eye inside the system.

726
00:41:01,840 --> 00:41:09,440
So, and I think, while for years, I think, Colin, just Pauma, I think, like, was a guru in

727
00:41:09,440 --> 00:41:10,440
the...

728
00:41:10,440 --> 00:41:12,480
Yes, I think, yeah, we tell you data, yeah, I love that.

729
00:41:12,480 --> 00:41:14,480
Was that a story telling?

730
00:41:14,480 --> 00:41:15,480
Yes.

731
00:41:15,480 --> 00:41:21,560
I think you have also, right, the book, storytelling with data, I think, yeah, yeah, yeah,

732
00:41:21,560 --> 00:41:24,440
okay, I write, great.

733
00:41:24,440 --> 00:41:31,800
But how did I change the way we do storytelling with data?

734
00:41:31,800 --> 00:41:39,520
Oh, it's changed it massively, but with exactly the same principles, which is brilliant,

735
00:41:39,520 --> 00:41:40,520
right?

736
00:41:40,520 --> 00:41:42,520
Because, so, you know, storytelling with data is also...

737
00:41:42,520 --> 00:41:48,080
So, I did a talk in secret, actually, storytelling with data, and I talked about the different

738
00:41:48,080 --> 00:41:49,080
types of reports, right?

739
00:41:49,080 --> 00:41:52,560
So, Colin, this is Bob and Affleck actually has an example where she talks about the different

740
00:41:52,560 --> 00:41:54,920
types of reporting, and it...

741
00:41:54,920 --> 00:41:57,920
She has the analogy of oysters and pearls.

742
00:41:57,920 --> 00:42:02,480
So, you could build a kind of exploratory report in dashboard, which is kind of for you,

743
00:42:02,480 --> 00:42:04,160
is your scratch pad, if you like.

744
00:42:04,160 --> 00:42:07,760
And that's where you're going hunting for pearls, you're hunting for insights, you're looking

745
00:42:07,760 --> 00:42:09,280
at what's important.

746
00:42:09,280 --> 00:42:14,840
You've then got the kind of... second example, which is kind of where you're building curated

747
00:42:14,840 --> 00:42:16,840
kind of insights, right?

748
00:42:16,840 --> 00:42:18,840
So, an exploratory journey, right?

749
00:42:18,840 --> 00:42:20,160
So, you're building kind of...

750
00:42:20,160 --> 00:42:22,000
It's like an essay style, kind of, right?

751
00:42:22,000 --> 00:42:26,800
So, you write in arguments, and an argument is basically a statement with context, because

752
00:42:26,800 --> 00:42:30,240
just to say, we had $2 million revenue last year, okay?

753
00:42:30,240 --> 00:42:31,240
Is that good or bad?

754
00:42:31,240 --> 00:42:36,560
So, what makes that useful is the context, how do you compare it to prior year?

755
00:42:36,560 --> 00:42:37,560
How do you compare it to budget?

756
00:42:37,560 --> 00:42:38,560
Or something like that?

757
00:42:38,560 --> 00:42:42,040
Or, you know, it's argument of context, and you're structuring it with this context to be

758
00:42:42,040 --> 00:42:45,640
able to put an end, and you've visualized it with this context, and that's really, really

759
00:42:45,640 --> 00:42:46,640
key.

760
00:42:46,640 --> 00:42:49,960
And then the third time, and that's a vast analogy, just to...

761
00:42:49,960 --> 00:42:53,960
You pointed people to oysters that contain pearls, right?

762
00:42:53,960 --> 00:42:56,560
So, that's the idea of storytelling now.

763
00:42:56,560 --> 00:42:59,640
You basically say, look, you know, I've done the hard work of figuring out kind of where

764
00:42:59,640 --> 00:43:00,640
things are interesting.

765
00:43:00,640 --> 00:43:02,840
You know, go look at this oyster, exactly.

766
00:43:02,840 --> 00:43:03,840
Tens this plug.

767
00:43:03,840 --> 00:43:05,400
Go look at this oyster, because that takes this plug.

768
00:43:05,400 --> 00:43:07,400
And, you know, you've got this breakdown, this visualization.

769
00:43:07,400 --> 00:43:09,840
And here's the context for you to evaluate whether that's good or bad.

770
00:43:09,840 --> 00:43:11,720
That's the key part of storytelling.

771
00:43:11,720 --> 00:43:17,440
And your third one is basically executive reporting where you're just giving them the numbers.

772
00:43:17,440 --> 00:43:19,720
And they trust that everything behind that is right.

773
00:43:19,720 --> 00:43:23,980
They're not beginning to go digging and do drill downs, and, you know, do all of this power

774
00:43:23,980 --> 00:43:24,980
be our exploratory work?

775
00:43:24,980 --> 00:43:27,140
They might ask Jenny, "Are there some questions now?"

776
00:43:27,140 --> 00:43:30,440
But, you know, they kind of trust that those numbers are fine, and that's basically the

777
00:43:30,440 --> 00:43:32,400
analogy of giving the pearls, technically.

778
00:43:32,400 --> 00:43:35,760
So, they're going into guaranteed oysters, or telling people, "Well, are you going

779
00:43:35,760 --> 00:43:37,760
to just give them the pearls?"

780
00:43:37,760 --> 00:43:39,640
So, I think that's the idea of kind of storytelling.

781
00:43:39,640 --> 00:43:42,600
And then I actually, at the end of that talk, I really talked about AI.

782
00:43:42,600 --> 00:43:46,320
And I said, "Well, AI is the best storyteller we've ever come across."

783
00:43:46,320 --> 00:43:47,320
Right?

784
00:43:47,320 --> 00:43:52,080
But, it gives you the same generic, plausible sounds and answers to everyone.

785
00:43:52,080 --> 00:43:53,080
Right?

786
00:43:53,080 --> 00:43:57,400
Now, actually, what you need is, is personalize intelligence.

787
00:43:57,400 --> 00:44:03,320
What you need is to know, you know, in addition, it would be great if AI actually learnt

788
00:44:03,320 --> 00:44:05,800
really about what you needed, what you want.

789
00:44:05,800 --> 00:44:09,160
And it tries to do that today by looking at all your chat history, and looking at everything

790
00:44:09,160 --> 00:44:12,560
you've uploaded, and looking at all of this, and storing it in a storage layer.

791
00:44:12,560 --> 00:44:16,680
But, actually, it's controlling how to use that context.

792
00:44:16,680 --> 00:44:20,920
So, this is where you have things like system instructions you can put in, and they help,

793
00:44:20,920 --> 00:44:25,000
but this is where your context is not just your organizational context, and it's not

794
00:44:25,000 --> 00:44:28,000
even just your team's context is your context.

795
00:44:28,000 --> 00:44:29,000
How do you work?

796
00:44:29,000 --> 00:44:30,000
What are you interested in?

797
00:44:30,000 --> 00:44:31,840
What's the business questions you want answered?

798
00:44:31,840 --> 00:44:36,160
So, this is where you get persona-driven insights, and then at the end you can tell the right

799
00:44:36,160 --> 00:44:41,120
stories for that, because, as we said at the beginning, where you really need to start is

800
00:44:41,120 --> 00:44:47,960
understanding those specific personas, those specific insights, and then AI can produce

801
00:44:47,960 --> 00:44:52,760
the right semantic models, the right reports, to tell the right stories for the right data.

802
00:44:52,760 --> 00:45:00,360
So, the changes, I think, before we have the dashboard, to tell me the story, the revenue

803
00:45:00,360 --> 00:45:07,560
decreased by 7%, and then I have to figure out why it's decreased, and I have to click through

804
00:45:07,560 --> 00:45:11,840
my, I don't know, throw down, and so on.

805
00:45:11,840 --> 00:45:22,160
And now, AI can say, okay, revenue decreased, 7% primarily, because the customer during increased

806
00:45:22,160 --> 00:45:28,080
the Germany, and then it's persona-driven.

807
00:45:28,080 --> 00:45:37,760
So, I say, the CEO, it says, because you have a Merco, and he is really bad to manage, and

808
00:45:37,760 --> 00:45:43,640
to me, it says, "Werco, you have to go and do this, this, this is the..."

809
00:45:43,640 --> 00:45:46,440
But this is what comes back to context, right?

810
00:45:46,440 --> 00:45:51,040
So, again, it's the context of the insights you need and what matters and how to interpret

811
00:45:51,040 --> 00:45:52,040
it.

812
00:45:52,040 --> 00:45:55,280
I give you an example, and I show this example again in SQL bits as well.

813
00:45:55,280 --> 00:45:58,000
So, say, "Costing Conversier", right?

814
00:45:58,000 --> 00:46:04,120
It's a ratio, example, and it has, so it's your operating cost divided by your revenue.

815
00:46:04,120 --> 00:46:05,120
Yeah.

816
00:46:05,120 --> 00:46:10,480
And, you know, we can calculate that quite easily from financial statements and draw balance.

817
00:46:10,480 --> 00:46:13,680
And we said, "Okay, so you've got a cost in Conversier."

818
00:46:13,680 --> 00:46:16,000
And now, how do we visualize that?

819
00:46:16,000 --> 00:46:17,960
How do you tell the story about cost in Conversier?

820
00:46:17,960 --> 00:46:21,920
So, you can show it broken down by division, and I showed you the bar charts, right?

821
00:46:21,920 --> 00:46:28,120
And I said, "Costing Conversier" was highest in A and B, and then division C and D had lower

822
00:46:28,120 --> 00:46:32,400
cost in Conversier's, and cost in Conversier's high as B bad, obviously.

823
00:46:32,400 --> 00:46:34,400
So, I said, "Okay, so what does the CFO do here?"

824
00:46:34,400 --> 00:46:37,400
Well, they go and call up their heads of division A and B and go, "Well, what the hell are you

825
00:46:37,400 --> 00:46:38,400
doing?

826
00:46:38,400 --> 00:46:39,400
Why is your cost in Conversier?

827
00:46:39,400 --> 00:46:40,400
Don't say much, right?

828
00:46:40,400 --> 00:46:43,480
Why are you spending so much when the benefit of the revenue here?"

829
00:46:43,480 --> 00:46:47,880
But I said, actually, let's now, and because the challenge overall is the cost in Conversier

830
00:46:47,880 --> 00:46:48,880
is high, right?

831
00:46:48,880 --> 00:46:49,880
This year.

832
00:46:49,880 --> 00:46:53,960
So then I said, "Actually, let's turn that bar chart into a scatter plot."

833
00:46:53,960 --> 00:47:00,560
So, now let's put cost in Conversier on Ralaxes and revenue on the other axes, because

834
00:47:00,560 --> 00:47:04,880
then you saw A and B actually, once they had a high cost in Conversier, they were actually

835
00:47:04,880 --> 00:47:07,040
relatively small units, divisions.

836
00:47:07,040 --> 00:47:12,560
So, actually, C and D had slightly lower cost in Conversier's, but higher revenues.

837
00:47:12,560 --> 00:47:17,640
Therefore, they're the ones that contributed probably more to the cost in Conversier being

838
00:47:17,640 --> 00:47:23,960
high overall, because revenue, because that's your denominator, right?

839
00:47:23,960 --> 00:47:28,880
So actually, maybe we don't need to call that A and B. Maybe we need to call that C and D,

840
00:47:28,880 --> 00:47:32,560
because they're the ones that, you know, a small change in their cost in Conversier is

841
00:47:32,560 --> 00:47:36,000
going to make a massive difference because of the revenue they contribute to it at all.

842
00:47:36,000 --> 00:47:39,960
So, actually, like, it's that context, which you need to know.

843
00:47:39,960 --> 00:47:43,000
Now, you need to know that context in order to know how to tell the right story, and

844
00:47:43,000 --> 00:47:47,720
to know that actually a bar chart by itself is a given enough context, maybe a scatter block

845
00:47:47,720 --> 00:47:49,320
with the revenue will work better.

846
00:47:49,320 --> 00:47:54,560
So that's your kind of visualization head of, how do I establish the context?

847
00:47:54,560 --> 00:47:58,000
Now, Gen A, I, yeah, this is where the intelligence comes in.

848
00:47:58,000 --> 00:47:59,000
I don't know.

849
00:47:59,000 --> 00:48:02,320
Maybe we need to give us a data and understand if it does do that kind of analysis for you,

850
00:48:02,320 --> 00:48:05,360
but at least given the right skills and instructions, it will.

851
00:48:05,360 --> 00:48:09,040
So, if you give it the kind of instructions to say, look, when you analyse these kind of

852
00:48:09,040 --> 00:48:13,720
ratios, you know, make sure you analyse the context of them and you can be quite genuine about

853
00:48:13,720 --> 00:48:14,720
this.

854
00:48:14,720 --> 00:48:17,480
You can say these kind of principles that we're talking about now, and you can bake some

855
00:48:17,480 --> 00:48:23,400
of those principles into your skills to interpret Power BI semantic models and faults.

856
00:48:23,400 --> 00:48:27,040
And then it will give you those kind of insights automatically, and it will tell you why it

857
00:48:27,040 --> 00:48:29,200
will tell you you'll drill into detail itself.

858
00:48:29,200 --> 00:48:33,280
You don't need to go and right click on a bar chart and drill down into detail to see

859
00:48:33,280 --> 00:48:35,360
what's going on and try and explore.

860
00:48:35,360 --> 00:48:37,280
It will do all of that for you.

861
00:48:37,280 --> 00:48:41,760
But by default, it will just probably say to you, oh, the cost of the conversation is this

862
00:48:41,760 --> 00:48:42,760
right?

863
00:48:42,760 --> 00:48:46,200
And yes, and then it will try and find some other reasons and those reasons might not be

864
00:48:46,200 --> 00:48:47,840
important and it might not be relevant.

865
00:48:47,840 --> 00:48:51,360
So you need to kind of guide it as to where personalised intelligence comes in.

866
00:48:51,360 --> 00:48:53,680
You need to guide it into what to look for.

867
00:48:53,680 --> 00:48:54,680
What's important.

868
00:48:54,680 --> 00:48:58,720
Like that example, we were just talking about actually think about what's important

869
00:48:58,720 --> 00:48:59,720
here.

870
00:48:59,720 --> 00:49:00,720
You're, you're denominated important.

871
00:49:00,720 --> 00:49:01,720
You're revenue important.

872
00:49:01,720 --> 00:49:03,920
I don't just want to focus on ones with the highest metrics.

873
00:49:03,920 --> 00:49:09,760
I want to focus on ones that are contributing most to my bottom line, which might actually

874
00:49:09,760 --> 00:49:13,520
just be the biggest divisions.

875
00:49:13,520 --> 00:49:18,720
So we've all become to the better static.

876
00:49:18,720 --> 00:49:21,840
We have, we have to do a little bit transition.

877
00:49:21,840 --> 00:49:24,000
So we'll be with some groups.

878
00:49:24,000 --> 00:49:30,200
So the people are, oh, we have, as I say, the co-pilot, AI, Tobic, is a little bit, yeah,

879
00:49:30,200 --> 00:49:31,200
last year.

880
00:49:31,200 --> 00:49:35,080
Now we're living in the agent.

881
00:49:35,080 --> 00:49:41,440
Agent, well, yeah, how would do is, where, what was the impact in Power BI and, and

882
00:49:41,440 --> 00:49:43,600
fabric when agents come?

883
00:49:43,600 --> 00:49:50,240
Well, it's, I think, I think Power BI and Fabric are a core grounding for agents, right?

884
00:49:50,240 --> 00:49:55,800
So it's a knowledge source with skills on how to interpret a data and logic, right?

885
00:49:55,800 --> 00:49:57,520
So that's, that's your core thing.

886
00:49:57,520 --> 00:50:03,760
So agents, you know, if, if we didn't have Power BI and fabric and we just had agents, or

887
00:50:03,760 --> 00:50:09,280
we didn't have a data platform and, you know, BI, I don't know, I think it's pretty, it's

888
00:50:09,280 --> 00:50:13,200
pretty risky already with how that and grand it is, it's even worse, right?

889
00:50:13,200 --> 00:50:15,520
So that's, that's one thing.

890
00:50:15,520 --> 00:50:20,240
But I think let's, let's come back a bit as well to what we mean by agents because actually

891
00:50:20,240 --> 00:50:22,760
there's different types of agents.

892
00:50:22,760 --> 00:50:29,080
So Copilot Studio, when up until fairly recently, I think maybe a few weeks or maybe a month

893
00:50:29,080 --> 00:50:35,160
or two, I'm not sure, Copilot Studio is essentially just Power Virtual Agents, right?

894
00:50:35,160 --> 00:50:41,800
So you built an agent in Copilot Studio, but all that agent was, was Topics, and the

895
00:50:41,800 --> 00:50:44,720
Topic Hand Knowledge Sources aligns to it.

896
00:50:44,720 --> 00:50:46,280
And then you, a chatbot.

897
00:50:46,280 --> 00:50:50,240
So you asked the chatbot an agent question, it tries to fit it into one of the topics and

898
00:50:50,240 --> 00:50:54,240
then use the knowledge sources, or if it can't answer with that, it just falls back to a

899
00:50:54,240 --> 00:50:55,240
full-back topic.

900
00:50:55,240 --> 00:50:56,240
Okay?

901
00:50:56,240 --> 00:50:57,240
That's all right.

902
00:50:57,240 --> 00:51:01,040
It was, it was the bad, but then people were saying, "Hang on a second, I'm not comparing

903
00:51:01,040 --> 00:51:05,040
this to Claude, I'm not Claude and Excel, where it's going through this chain of what

904
00:51:05,040 --> 00:51:08,600
we's in need and it's got, you know, it's calculated things that's producing reports,

905
00:51:08,600 --> 00:51:09,600
that's producing these things."

906
00:51:09,600 --> 00:51:14,440
So that is such a different level to a chatbot with Topics and Knowledge.

907
00:51:14,440 --> 00:51:16,080
So actually there are different types of agents.

908
00:51:16,080 --> 00:51:20,080
Now, Copilot Studio with a GitHub Copilot harness has now got this, and it's going

909
00:51:20,080 --> 00:51:23,000
to have this second type of agent standard.

910
00:51:23,000 --> 00:51:24,000
Right?

911
00:51:24,000 --> 00:51:25,680
Then you experience a Copilot Studio.

912
00:51:25,680 --> 00:51:32,480
I did a trade-in course on it last week with a Moxot product team in London.

913
00:51:32,480 --> 00:51:36,960
And I mean, I have looked at Copilot Studio for a few months, at least, but I was blown

914
00:51:36,960 --> 00:51:39,880
away by what the GitHub Copilot harness could do.

915
00:51:39,880 --> 00:51:41,680
It's got, you know, an agent sandbox.

916
00:51:41,680 --> 00:51:46,360
It's got, now you can add tools and skills and, you know, MCP connectors and all of this

917
00:51:46,360 --> 00:51:48,000
stuff into a single agent.

918
00:51:48,000 --> 00:51:50,960
And then you can combine them to get into the workloads and all of this.

919
00:51:50,960 --> 00:51:58,040
So I think when you say how agents have worked, I mean, that level of agents is really powerful.

920
00:51:58,040 --> 00:52:04,800
And yes, you have to be more careful with those types of agents because it's far more dangerous.

921
00:52:04,800 --> 00:52:05,960
Right?

922
00:52:05,960 --> 00:52:11,160
And this is why I think Microsoft actually probably did a safe thing, you know, a rich

923
00:52:11,160 --> 00:52:17,200
legal copilot where they kind of restricted copilot to be within these kind of topic, knowledge-based

924
00:52:17,200 --> 00:52:23,120
domains and, you know, it was very limited in terms of what you can do, but it was also

925
00:52:23,120 --> 00:52:24,800
less chance of hallucination, right?

926
00:52:24,800 --> 00:52:28,840
If you've got very defined topics and knowledge sources for those, your chance of hallucination

927
00:52:28,840 --> 00:52:30,240
are a lot lower.

928
00:52:30,240 --> 00:52:35,560
If you've got Claude and agents that, you know, all of this stuff now going off and thinking

929
00:52:35,560 --> 00:52:39,240
for themselves and going for a chain of thoughts and saying, hey, I can look at this or

930
00:52:39,240 --> 00:52:41,040
maybe it's to do with this or maybe it's to do with that.

931
00:52:41,040 --> 00:52:44,440
Or actually, here's analysis out there, you didn't know since then, but here you go.

932
00:52:44,440 --> 00:52:45,440
There's lots of fun.

933
00:52:45,440 --> 00:52:46,440
This is this.

934
00:52:46,440 --> 00:52:51,560
Now, suddenly, your risk level goes out in terms of your risk of hallucination goes up massively

935
00:52:51,560 --> 00:52:55,160
and your token cost goes up because it's charging you for all of this fault process as well,

936
00:52:55,160 --> 00:52:57,160
by the way, absolutely.

937
00:52:57,160 --> 00:53:01,720
But, you know, your everything's, everything's not becoming riskier and more expensive than

938
00:53:01,720 --> 00:53:02,720
agents.

939
00:53:02,720 --> 00:53:09,040
So, now that makes it even more important than it was before for your agents to be grounded

940
00:53:09,040 --> 00:53:13,640
in solid data foundation and solid logic foundation with semantic models of like houses and things

941
00:53:13,640 --> 00:53:14,640
like that.

942
00:53:14,640 --> 00:53:16,400
And so then it's not hallucinating.

943
00:53:16,400 --> 00:53:19,960
It's it's the and you need skills and instructions to tell it.

944
00:53:19,960 --> 00:53:20,960
Don't invent numbers.

945
00:53:20,960 --> 00:53:21,960
Don't go looking in places.

946
00:53:21,960 --> 00:53:23,960
I haven't told you to go and look.

947
00:53:23,960 --> 00:53:27,320
Don't go come up with bits of analysis that I haven't asked you to do.

948
00:53:27,320 --> 00:53:30,040
Like you need to be and the as the models get more and more intelligent, you need to be

949
00:53:30,040 --> 00:53:36,280
strict and strict and stricter is the instruction and tell it to use or take to foundation,

950
00:53:36,280 --> 00:53:38,840
your core logic foundation, your semantic models and like.

951
00:53:38,840 --> 00:53:45,840
And then you need to be a little bit more intelligent and you need to be a little bit more intelligent

952
00:53:45,840 --> 00:53:52,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

953
00:53:52,840 --> 00:53:56,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

954
00:53:56,840 --> 00:54:03,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

955
00:54:03,840 --> 00:54:07,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

956
00:54:07,840 --> 00:54:14,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

957
00:54:14,840 --> 00:54:21,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

958
00:54:21,840 --> 00:54:26,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

959
00:54:26,840 --> 00:54:31,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

960
00:54:31,840 --> 00:54:36,840
and you need to be a little bit more intelligent and you need to be a little bit more intelligent

961
00:54:36,840 --> 00:54:41,480
is where Ella Lemous, a judge comes in and you know critics and and again this stuff is built

962
00:54:41,480 --> 00:54:46,440
into copilot studio now. So, this is where you know, we're not having to reinvent the wheel

963
00:54:46,440 --> 00:54:51,000
and actually it's much more aligned to how we'd manage employees rather than this

964
00:54:51,000 --> 00:54:57,000
chatbot agent which we had before which was just topics. Yep, Now we come to the med

965
00:54:57,000 --> 00:55:03,080
path. I think everyone who would have like need these crew agents, man yeah, And zwar

966
00:55:03,080 --> 00:55:14,200
wants to yeah, nobody wants the governance meeting actually. The governance meeting.

967
00:55:14,200 --> 00:55:25,240
No, no, no, no. Yeah, how did we make governance great again? I do make cover this great again.

968
00:55:25,240 --> 00:55:29,720
Oh, maybe that needs to be a cap. Make governance great again. M-G-G-A.

969
00:55:31,720 --> 00:55:35,000
That should be your new tackle. How do they come from this great again?

970
00:55:35,000 --> 00:55:40,920
Go for this. So, you know what I was talking about is organizational brain.

971
00:55:40,920 --> 00:55:47,480
I've got six agents who are the core kind of team that sit on top of this brain as obviously

972
00:55:47,480 --> 00:55:52,760
you can define your own agents skills as well. And one of those is a governance manager.

973
00:55:52,760 --> 00:55:58,120
And I've got a little persona based avatar videos for each of these agents personas that talk

974
00:55:58,120 --> 00:56:02,920
about their tools, their skills, what they do. The first one with Gabrielle is the governance manager.

975
00:56:02,920 --> 00:56:10,280
She's like, I know governance has a reputation of where good ideas go to die, but let me let me

976
00:56:10,280 --> 00:56:15,480
reframe that for you, right? And it talks about governance in terms of kind of actually not

977
00:56:15,480 --> 00:56:21,320
being that a nabler, being that a nabler to be able to trust AI. So I think governance is great

978
00:56:21,320 --> 00:56:27,240
again. And I think it's been brought right to the top of the agenda for almost every company

979
00:56:27,240 --> 00:56:32,840
because as soon as these you started to use agents in AI, especially as become more intelligent,

980
00:56:32,840 --> 00:56:38,120
and it's no longer restricted to this kind of topic based chatbot, how do we make sure that we're

981
00:56:38,120 --> 00:56:43,640
getting reliable results? And that's governance in the nutshell. And governance is, you know,

982
00:56:43,640 --> 00:56:49,080
and also security, right? How do we make sure agents, not just employees, but now also agents,

983
00:56:49,080 --> 00:56:54,280
are accessing the right data using the right tools with the right skills, with the right,

984
00:56:55,400 --> 00:57:00,280
you know, access permissions with the right security policies or the right infrastructure.

985
00:57:00,280 --> 00:57:05,880
And actually, if you build all of that into a context layer, then suddenly whenever agents do

986
00:57:05,880 --> 00:57:11,080
and you force it to use this governance policies, like you tell it where to use it and you force

987
00:57:11,080 --> 00:57:15,240
it to use it with contracts, like again, like humans, right? We have employment contracts.

988
00:57:15,240 --> 00:57:20,280
And my employment contract doesn't say I need to go and use this skill in here, or it doesn't say,

989
00:57:20,280 --> 00:57:24,840
I need to go and make sure I apply this governance policy because the government contracts are not

990
00:57:24,840 --> 00:57:29,960
attended for that. But agent employment contracts are intended for that. So we write employment

991
00:57:29,960 --> 00:57:35,560
contracts for our agents that tell it, we need to use this governance policies. And then it will,

992
00:57:35,560 --> 00:57:40,200
that's how governance is automatically enforced. So for me, governance is security. How do we make

993
00:57:40,200 --> 00:57:47,800
governance great again? No, single word I'd use for that is automation, right? I think governance

994
00:57:47,800 --> 00:57:53,320
is be one of those things that's being too important to just write into 100 page documents

995
00:57:53,320 --> 00:57:58,760
that no one reads. Governance has been too important and security has been too important to have people

996
00:57:58,760 --> 00:58:04,040
manually apply road level security into their power BI to mantifoddles with. It's been too important

997
00:58:04,040 --> 00:58:09,880
to have, you know, to, to try to stick to these all these manual and things and too important to

998
00:58:09,880 --> 00:58:15,880
let it live in people's heads as the majority of its statusors. So how do we make governance great again?

999
00:58:15,880 --> 00:58:19,960
We put it into our organization or brain. We make it a foundation if our organization or brain

1000
00:58:19,960 --> 00:58:25,000
and we put our agents on top of that and that in an automated way, right? Using some of that

1001
00:58:25,000 --> 00:58:31,400
of Andrew Acky. So that's how we make governance great again. Okay, good. And will you say who

1002
00:58:31,400 --> 00:58:37,720
should own own the governance or is it different from the semantic model to KPI definition

1003
00:58:37,720 --> 00:58:44,440
or should we have one over? So I think the way I kind of look at it is that this organizational

1004
00:58:44,440 --> 00:58:51,640
brain is obviously your source of truth, your data for your logic, for your tools, the skills,

1005
00:58:51,640 --> 00:58:59,240
for your governance. Right. Now who owns that? Well, I think in terms of populating that brain,

1006
00:58:59,240 --> 00:59:05,320
I think every business user who runs processes, which is all us,

1007
00:59:05,320 --> 00:59:11,480
needs to convert their processes into a custom skill in that brain that has the right governance

1008
00:59:11,480 --> 00:59:18,120
supply that has the right tools, skills, data logic. Right. Now I think the operating model that I see

1009
00:59:18,120 --> 00:59:23,480
that really operating in is is people could take an exosperitiate and use my tool or other kind of

1010
00:59:23,480 --> 00:59:27,800
processes to convert that into a custom skill that could be part of this brain and then managed

1011
00:59:27,800 --> 00:59:32,440
by a skills manager in that brain. But I think initially that should probably be

1012
00:59:32,440 --> 00:59:38,040
an unsertified level. Right. And maybe not part of the real brain that everyone uses, right? But

1013
00:59:38,040 --> 00:59:44,040
maybe part of just a local kind of, okay, like whatever, right, that you run in yourself or one for

1014
00:59:44,040 --> 00:59:48,120
your team. And then it should go through some kind of certification process like we do with some

1015
00:59:48,120 --> 00:59:52,680
antiponels like we do with these other things, whether it's okay. Yeah. This makes sense. Let's make

1016
00:59:52,680 --> 00:59:57,400
this why accessible to the wider organization. And that's where it then becomes part of the brain.

1017
00:59:57,400 --> 01:00:01,080
And that's where you make sure that everything because governance is applied properly and all

1018
01:00:01,080 --> 01:00:06,120
these things will apply properly. At the moment, this is still very hard to do. But I'm going to be,

1019
01:00:06,120 --> 01:00:11,400
be realistic here, right? Even if you have found your IQ or something, it's not a magic bullet

1020
01:00:11,400 --> 01:00:15,560
to this. And most of those problems are not technology problems. They're people who process

1021
01:00:15,560 --> 01:00:20,760
problems. So how do we get the operating model right around this? And that's the bigger problem

1022
01:00:20,760 --> 01:00:28,680
we've got. So, yeah, the operating morning is also an interesting topic. I think, yeah,

1023
01:00:28,680 --> 01:00:33,720
we have for the sea level, we have the the bus word, the center of excellence.

1024
01:00:33,720 --> 01:00:41,720
So, they might, but what does a good operation model for fabric and AI looks like from your perspective?

1025
01:00:41,720 --> 01:00:46,440
So, center of excellence is absolutely key for that. And when I talk about center of excellence,

1026
01:00:46,440 --> 01:00:51,240
I think it's one of those buzzwords, if you like, that people have different ideas about what it means.

1027
01:00:51,240 --> 01:00:56,840
But for me, a center of excellence is just the enablements team, your core enablements team.

1028
01:00:56,840 --> 01:01:03,320
They're the team that allow you to do self service. They're a team that allow you to do this.

1029
01:01:03,320 --> 01:01:09,400
So, the fabric team are going to be responsible essentially for building all your hub workspaces.

1030
01:01:09,400 --> 01:01:13,720
So, this is a hub and spoke operating model essentially, right? So, you've got your certified

1031
01:01:13,720 --> 01:01:19,800
data sets, data and logic in certain workspaces that are all certified and those are all automated.

1032
01:01:19,800 --> 01:01:24,440
Everything goes through a kind of central function for that. And it's probably, you know, maybe

1033
01:01:24,440 --> 01:01:29,000
code first and that or even, or still AI, but you know, maybe, maybe more code first and

1034
01:01:29,000 --> 01:01:35,320
dev-obsolated, that's your kind of certified layer, your hub layer. And then spokes, which is the

1035
01:01:35,320 --> 01:01:41,160
kind of business user teams, kind of connect into the hubs and build their own versions of logic and

1036
01:01:41,160 --> 01:01:46,040
data tools and asset gets wants to be used more widely and get certified back into the hub.

1037
01:01:46,040 --> 01:01:52,920
Okay, for the process. So, that kind of operating model, I think, works very well for the BI,

1038
01:01:52,920 --> 01:01:58,280
that I would speak, I would speak model. And kind of the sets of excellence is responsible for

1039
01:01:58,280 --> 01:02:04,360
making sure there's a clopier owners for each hub workspace and a data set and for each spokes as

1040
01:02:04,360 --> 01:02:08,600
well and making spokes that maybe help you then move them from there to there or making sure they

1041
01:02:08,600 --> 01:02:13,000
could build stuff in the right way and use the hub data sets in the right way. So, I think that's

1042
01:02:13,000 --> 01:02:19,240
the right kind of model of like for BI and for AI, it becomes even more important to kind of do

1043
01:02:19,240 --> 01:02:23,400
some of this. So, I'm saying, I think with the organizational grain, now your assets are not just

1044
01:02:23,400 --> 01:02:30,520
data assets or, you know, spreadsheets, your assets and our skills and your assets are now,

1045
01:02:30,520 --> 01:02:34,520
you know, all of those things that look for part of your organizational brain, they're all assets,

1046
01:02:34,520 --> 01:02:40,920
your governance policies, your assets, your tools, your assets. So, how do we combine these assets

1047
01:02:40,920 --> 01:02:46,600
together into a catalog that agents consider top of and humans as well, right? Obviously,

1048
01:02:46,600 --> 01:02:49,960
there's nothing stopping the human going, querying that organizational brain with van dreiky,

1049
01:02:49,960 --> 01:02:56,040
right? It's Q and A things as well. So, I think that's the kind of model that works well. So, this is a

1050
01:02:56,040 --> 01:03:00,760
people will build them in skills. I showed you in a context a couple of months ago, I showed how to

1051
01:03:00,760 --> 01:03:05,080
build a custom skill for an Excel process in Excel using Cloud for Excel. So, just lightly

1052
01:03:05,080 --> 01:03:10,760
implementation plan, great looks good, turn this into a skill, open up a new set of data, run the skill,

1053
01:03:10,760 --> 01:03:15,160
create your cow, subgeneral sheets like a VBA macro wood, but you know, now it's driven from natural

1054
01:03:15,160 --> 01:03:20,440
language. So, now, how do we manage those assets? That's the key and that's the off-ageable

1055
01:03:20,440 --> 01:03:24,840
that's needed right then. Yeah, then, oh, yeah, we're running all the time. You have to go to the

1056
01:03:24,840 --> 01:03:32,280
building. Oh, good. I'm a real story, but I have to wrap it up, I make it short. So, centralised

1057
01:03:32,280 --> 01:03:40,200
or federated governance? A mixture, right? So, Zen, I think the hop and spoke is a mix of centralised

1058
01:03:40,200 --> 01:03:47,320
and federated. One enterprise, I think, model or then you domain models. So, well,

1059
01:03:47,320 --> 01:03:54,360
again, a mix, right? Like, I think you need the semantic models to represent your business logic.

1060
01:03:54,360 --> 01:03:59,800
Now, where that, and maybe a different layers, right? So, you know, maybe at the silver layer,

1061
01:03:59,800 --> 01:04:04,840
things are more centralised and more kind of enterprise level, and then maybe your kind of

1062
01:04:04,840 --> 01:04:09,000
gold layers are some of the more business layers. And then again, maybe platinum layers,

1063
01:04:09,000 --> 01:04:12,040
we need to start to think about it. So, maybe when standardising a bit more the gold, then maybe

1064
01:04:12,040 --> 01:04:16,280
people could build their own things on platinum, and that's where agents can kind of then work and

1065
01:04:16,280 --> 01:04:21,400
work across as well. And I think that's really interesting for the rest of the world. What's the best

1066
01:04:21,400 --> 01:04:25,880
food in England? The best food in England. You know, the most, do you know, the most popular food

1067
01:04:25,880 --> 01:04:32,120
in England? You'll take a guess at what it is. The most popular dish. Yeah, a foodie chip? No,

1068
01:04:32,840 --> 01:04:42,200
haggers? Maybe a scones. I don't know. Chicken teacup,

1069
01:04:42,200 --> 01:04:48,280
thunder? Huh? Chicken teacup, a bissala. Oh, okay, yeah, really British.

1070
01:04:48,280 --> 01:04:55,640
Obviously the British style or the British version of it. So, I mean, I don't have asked the best

1071
01:04:55,640 --> 01:05:00,600
food in England, but it's the most popular dish apparently. But yeah, maybe fish and chips if

1072
01:05:01,160 --> 01:05:05,320
there's something coming to the record, they want to try something maybe fish and chips. Yeah,

1073
01:05:05,320 --> 01:05:09,080
and you've come to London to give me a shout because I'm a, there's a nice relationship,

1074
01:05:09,080 --> 01:05:14,360
please, maybe I'll give it to the same house in London for over 30 years. And probably, yeah,

1075
01:05:14,360 --> 01:05:22,040
who shall I invite next and what, see what person should I ask? I think we definitely dive deeper

1076
01:05:22,040 --> 01:05:28,360
into this Copilot studio agents and the copilot harness and stuff like that because I just,

1077
01:05:28,360 --> 01:05:32,520
I saw saying I wasn't really aware of it until last week and it's just such a huge area and I think,

1078
01:05:32,520 --> 01:05:37,160
how do we integrate this? How do we build the right context layers with the, with the

1079
01:05:37,160 --> 01:05:42,840
copilot harness and how do we make it safe? Because again, great, you've got the capability

1080
01:05:42,840 --> 01:05:47,400
that these agents are going to take you for a ride if you're not, if you haven't got the right guard

1081
01:05:47,400 --> 01:05:54,440
12, the right skills, the right tools. Yeah, then I say, which is thank you for for being here

1082
01:05:54,440 --> 01:06:00,200
in my podcast or it's more of the new plan. Yeah, we have a really cool overview, I think,

1083
01:06:00,200 --> 01:06:07,960
now from, yeah, from the start to the end. And I hope we have another session with G-Drive

1084
01:06:07,960 --> 01:06:15,640
and the other topics. And yeah, then thank you for being here and, yeah, have a good meeting.

1085
01:06:15,640 --> 01:06:17,640
No, you too, okay, thanks very much.

1086
01:06:17,640 --> 01:06:18,360
Exist?

1087
01:06:18,360 --> 01:06:28,360
[BLANK_AUDIO]