WEBVTT

1
00:00:03.520 --> 00:00:05.919
- Welcome to episode 437

2
00:00:05.919 --> 00:00:09.139
- of the Microsoft Cloud IT Pro podcast recorded

3
00:00:09.199 --> 00:00:11.939
- live on 09/18/2026.

4
00:00:12.160 --> 00:00:14.375
- This is a show about Microsoft three sixty

5
00:00:14.375 --> 00:00:16.535
- five and Azure from the perspective of IT

6
00:00:16.535 --> 00:00:18.695
- pros and end users, where we discuss a

7
00:00:18.695 --> 00:00:21.015
- topic for recent news and how it relates

8
00:00:21.015 --> 00:00:21.594
- to you.

9
00:00:21.894 --> 00:00:23.894
- Today, we have a special treat as we're

10
00:00:23.894 --> 00:00:25.434
- joined by one of our sponsors,

11
00:00:25.800 --> 00:00:28.919
- Nasuni, to chat about unstructured data and what

12
00:00:28.919 --> 00:00:31.099
- it takes to get it ready for AI.

13
00:00:31.399 --> 00:00:34.699
- Nick Burling, chief product officer at Nasuni,

14
00:00:35.159 --> 00:00:37.239
- joins us to talk about why so many

15
00:00:37.239 --> 00:00:40.954
- AI projects stall out, the psychic framework Nasuni

16
00:00:41.015 --> 00:00:42.795
- uses to assess AI readiness,

17
00:00:43.094 --> 00:00:46.534
- and how AI activate enforces permissions before the

18
00:00:46.534 --> 00:00:49.114
- query instead of filtering results afterward.

19
00:00:49.575 --> 00:00:52.799
- Nick also covers their recent Drive IQ acquisition,

20
00:00:53.259 --> 00:00:57.179
- how that extends discovery, classification, and tagging beyond

21
00:00:57.179 --> 00:00:57.679
- SharePoint

22
00:00:58.059 --> 00:01:01.020
- alongside the sensitivity labels you already manage in

23
00:01:01.020 --> 00:01:04.640
- Microsoft Purview. He also teases a bigger announcement

24
00:01:04.780 --> 00:01:06.405
- coming on October 6.

25
00:01:08.965 --> 00:01:11.385
- Here we are, another episode of the Microsoft

26
00:01:11.444 --> 00:01:14.884
- Cloud IT Pro podcast. Another Friday afternoon. Back

27
00:01:14.884 --> 00:01:16.724
- again. I'm excited about this one. We have

28
00:01:16.724 --> 00:01:18.325
- a special guest this week. I am too.

29
00:01:18.325 --> 00:01:19.740
- We have a special guest. So if you've

30
00:01:19.740 --> 00:01:21.819
- been listening to the show for the last

31
00:01:21.819 --> 00:01:25.020
- several months, we've had one of our key

32
00:01:25.020 --> 00:01:25.520
- sponsors

33
00:01:25.980 --> 00:01:30.140
- recently is Nasuni. And Nick is actually joining

34
00:01:30.140 --> 00:01:33.194
- us today from Nasuni to talk a little

35
00:01:33.194 --> 00:01:35.774
- bit about them, about some of their products,

36
00:01:35.834 --> 00:01:36.334
- about

37
00:01:36.795 --> 00:01:40.015
- AI and data. So, Nick, do you wanna

38
00:01:40.234 --> 00:01:42.655
- introduce yourself, give us a little bit of

39
00:01:42.795 --> 00:01:43.375
- a overview

40
00:01:43.834 --> 00:01:45.435
- of Nasuni and some of that, and then

41
00:01:45.435 --> 00:01:47.834
- we'll dive into some of the topics for

42
00:01:47.834 --> 00:01:49.969
- today. Fantastic. Yeah. Well, thanks for having me,

43
00:01:49.969 --> 00:01:52.209
- guys. So I'm Nick Burling, the chief product

44
00:01:52.209 --> 00:01:54.770
- officer here at the SUNY, located in our

45
00:01:54.770 --> 00:01:57.490
- Boston area where our headquarters are. Been with

46
00:01:57.490 --> 00:01:59.909
- the company for just over five years now,

47
00:02:00.130 --> 00:02:02.130
- and this is a phenomenal topic for me.

48
00:02:02.450 --> 00:02:03.944
- I was talking with the guys, you know,

49
00:02:03.944 --> 00:02:05.784
- the IT pro audience is one that I

50
00:02:05.784 --> 00:02:07.564
- have served pretty much my whole career

51
00:02:07.865 --> 00:02:08.525
- in technology.

52
00:02:08.905 --> 00:02:11.224
- I actually did a stint at Microsoft where

53
00:02:11.224 --> 00:02:12.985
- I spent a lot of time talking about

54
00:02:12.985 --> 00:02:14.985
- how to make lives better for IT pros.

55
00:02:14.985 --> 00:02:16.044
- So here at Nasuni,

56
00:02:16.360 --> 00:02:19.959
- we're really focused around providing an unstructured data

57
00:02:19.959 --> 00:02:23.479
- foundation for enterprise teams and now increasingly the

58
00:02:23.479 --> 00:02:25.239
- AI that they depend on. So what do

59
00:02:25.239 --> 00:02:26.599
- I mean by that? So if you think

60
00:02:26.599 --> 00:02:27.099
- about

61
00:02:27.400 --> 00:02:29.259
- large engineering, construction,

62
00:02:29.639 --> 00:02:31.794
- architecture firms, you think about manufacturing

63
00:02:32.094 --> 00:02:35.635
- firms, energy firms, their businesses run on unstructured

64
00:02:35.775 --> 00:02:38.115
- data, right? And unstructured data represents,

65
00:02:38.495 --> 00:02:40.495
- depending on how you talk to 80 plus

66
00:02:40.495 --> 00:02:43.639
- percent of an enterprise's data. And this all

67
00:02:43.639 --> 00:02:46.120
- lives in what we call the operational file

68
00:02:46.120 --> 00:02:48.459
- layer, right? So the CAD drawings,

69
00:02:48.759 --> 00:02:51.959
- the geospatial data that energy companies are gathering

70
00:02:51.959 --> 00:02:53.979
- when they're doing ocean floor exploration.

71
00:02:54.360 --> 00:02:56.280
- And these are the critical files that their

72
00:02:56.280 --> 00:02:59.305
- business runs on. And here at Nasuni, we

73
00:02:59.305 --> 00:03:01.865
- serve some of the largest enterprise companies around

74
00:03:01.865 --> 00:03:04.264
- the world. They tend to have petabytes of

75
00:03:04.264 --> 00:03:07.004
- data. They have very distributed organizations,

76
00:03:07.705 --> 00:03:10.105
- and that that data really is kind of

77
00:03:10.105 --> 00:03:12.939
- the lifeblood of their organization. So it's critical

78
00:03:12.939 --> 00:03:15.439
- that they have a way to manage, protect,

79
00:03:15.580 --> 00:03:18.699
- and then activate that data. Manage means I

80
00:03:18.699 --> 00:03:22.000
- may have 20 different offices around the world

81
00:03:22.219 --> 00:03:25.019
- and they're all all these engineers, all these

82
00:03:25.019 --> 00:03:26.560
- architects, etcetera, are

83
00:03:26.895 --> 00:03:29.395
- generating new files every day, working and collaborating

84
00:03:29.534 --> 00:03:31.775
- on files. How do I get my hand

85
00:03:31.775 --> 00:03:34.414
- around that? And traditionally, when Asuni goes and

86
00:03:34.414 --> 00:03:36.495
- talks with customers, they often have silos of

87
00:03:36.495 --> 00:03:38.334
- data all over the place. So how do

88
00:03:38.334 --> 00:03:39.955
- they make sense of that? Manage.

89
00:03:40.290 --> 00:03:42.050
- Protect, right? I mean, this is the critical

90
00:03:42.050 --> 00:03:43.889
- data that runs the business. It's gotta be

91
00:03:43.889 --> 00:03:46.530
- protected. Ransomware, we all know is rampant out

92
00:03:46.530 --> 00:03:49.409
- there. Natural disasters happen. So do I have

93
00:03:49.409 --> 00:03:51.330
- a way of protecting that data and then

94
00:03:51.330 --> 00:03:52.469
- rapidly recovering

95
00:03:52.805 --> 00:03:54.564
- if there is an issue? And then the

96
00:03:54.564 --> 00:03:56.724
- last one is really around activate, and we

97
00:03:56.724 --> 00:03:58.564
- think about that in two ways. One is

98
00:03:58.564 --> 00:04:00.805
- for teams. So it doesn't do me any

99
00:04:00.805 --> 00:04:02.724
- good if I'm an engineer working in the

100
00:04:02.724 --> 00:04:04.724
- field and I can't access the files that

101
00:04:04.724 --> 00:04:05.979
- I need to be able to do my

102
00:04:05.979 --> 00:04:08.060
- job, and it doesn't do my teammates any

103
00:04:08.060 --> 00:04:10.060
- good if they can't collaborate on that data

104
00:04:10.060 --> 00:04:11.740
- safely and in a way where we're not

105
00:04:11.740 --> 00:04:14.540
- overriding each other's work. So there's activating for

106
00:04:14.540 --> 00:04:17.199
- teams, but then there's now activating for AI,

107
00:04:17.259 --> 00:04:20.000
- which introduces a whole other set of challenges.

108
00:04:20.365 --> 00:04:22.604
- And really security and governance are kind of

109
00:04:22.604 --> 00:04:24.125
- the big ones that we hear a lot

110
00:04:24.125 --> 00:04:26.925
- from customers. Do I know who's using AI

111
00:04:26.925 --> 00:04:29.104
- and what data they're accessing? Do I know

112
00:04:29.245 --> 00:04:31.644
- that certain types of data should or shouldn't

113
00:04:31.644 --> 00:04:33.485
- be made available to AI even if that

114
00:04:33.485 --> 00:04:36.290
- individual has access to that? Am I able

115
00:04:36.290 --> 00:04:39.810
- to understand how effectively organizations are using AI?

116
00:04:39.810 --> 00:04:41.910
- So we talk a lot about token utilization.

117
00:04:42.050 --> 00:04:43.889
- I'm sure you guys live with this in

118
00:04:43.889 --> 00:04:46.129
- your own use of AI every day. Being

119
00:04:46.129 --> 00:04:48.290
- able to optimize how people are actually able

120
00:04:48.290 --> 00:04:51.384
- to use AI with their proprietary data because

121
00:04:51.384 --> 00:04:53.324
- that's gold in the AI area.

122
00:04:53.785 --> 00:04:55.305
- So that's kind of a top level view.

123
00:04:55.305 --> 00:04:58.025
- Again, we work with large enterprise organizations across

124
00:04:58.025 --> 00:04:59.944
- the world and Microsoft is one of our

125
00:04:59.944 --> 00:05:02.345
- closest partners. We do a tremendous amount of

126
00:05:02.345 --> 00:05:05.399
- business because we provide a way to orchestrate

127
00:05:05.460 --> 00:05:07.460
- storage that lives on top of Azure Blob

128
00:05:07.460 --> 00:05:09.300
- Storage, for example. And so we have a

129
00:05:09.300 --> 00:05:11.220
- great partnership. And, yeah, happy to tell you

130
00:05:11.220 --> 00:05:12.980
- more about kinda what we're seeing in the

131
00:05:12.980 --> 00:05:14.600
- market. Yeah. I think it's an interesting

132
00:05:14.980 --> 00:05:16.040
- space. So

133
00:05:16.954 --> 00:05:19.115
- everybody goes on this journey. They kinda start

134
00:05:19.115 --> 00:05:20.735
- maybe just throwing

135
00:05:21.194 --> 00:05:24.254
- MCPs at things and connecting to data stores.

136
00:05:24.314 --> 00:05:26.154
- And you you you sit there, and it's

137
00:05:26.154 --> 00:05:28.254
- like the early days where LLMs were magical.

138
00:05:28.314 --> 00:05:29.915
- Hey. This machine can talk to me. Can

139
00:05:29.915 --> 00:05:31.769
- it can it really talk to me? That's

140
00:05:32.250 --> 00:05:34.430
- a conversation for another day. But it actually

141
00:05:35.050 --> 00:05:35.550
- becomes

142
00:05:36.329 --> 00:05:36.829
- really

143
00:05:37.209 --> 00:05:37.709
- hard

144
00:05:38.169 --> 00:05:38.669
- to

145
00:05:39.289 --> 00:05:41.129
- to when you think about it for, like,

146
00:05:41.129 --> 00:05:44.810
- how to use AI safely, especially with things

147
00:05:44.810 --> 00:05:45.789
- that are unstructured

148
00:05:46.169 --> 00:05:49.264
- data stores, like object storage, even, like, your

149
00:05:49.264 --> 00:05:52.064
- generic SMB file share. Right? You've had all

150
00:05:52.064 --> 00:05:54.384
- these departmental drives and things sitting out there

151
00:05:54.384 --> 00:05:57.345
- where ACLs have just kinda Oh, we lost

152
00:05:57.345 --> 00:05:59.824
- track. Time, and nobody knows what's in there

153
00:05:59.824 --> 00:06:01.410
- or what's going on. So could you maybe

154
00:06:01.410 --> 00:06:03.490
- talk a little bit about why it's hard

155
00:06:03.490 --> 00:06:06.769
- for folks to use AI safely today? Yeah.

156
00:06:06.769 --> 00:06:09.250
- Absolutely. So I started I alluded a little

157
00:06:09.250 --> 00:06:11.250
- bit to it. But, typically, when we're first

158
00:06:11.250 --> 00:06:13.490
- getting introduced to a customer or a prospect,

159
00:06:13.490 --> 00:06:15.884
- I should say, it's a pretty consistent story

160
00:06:15.884 --> 00:06:17.964
- just like you said. They've had these little

161
00:06:18.044 --> 00:06:21.004
- these pockets of data that grow over time.

162
00:06:21.004 --> 00:06:22.925
- It can be living on Windows file servers

163
00:06:22.925 --> 00:06:24.764
- in some part of the organization, could be

164
00:06:24.764 --> 00:06:27.004
- backed by NetApp arrays and others. They could

165
00:06:27.004 --> 00:06:30.030
- have data up in Azure blob storage, in

166
00:06:30.030 --> 00:06:33.170
- native object storage. And so first and foremost

167
00:06:33.230 --> 00:06:35.790
- is just getting their hands around the data

168
00:06:35.790 --> 00:06:37.870
- that they have. So we think about sort

169
00:06:37.870 --> 00:06:40.350
- of unifying your view and control over the

170
00:06:40.350 --> 00:06:42.355
- data is one of the first challenges. Cause

171
00:06:42.355 --> 00:06:43.714
- if you don't know what data you have,

172
00:06:43.714 --> 00:06:45.395
- then how are you gonna know what data

173
00:06:45.395 --> 00:06:48.355
- could be made ready and optimized for use

174
00:06:48.355 --> 00:06:50.835
- with AI? And, but also included in that

175
00:06:50.835 --> 00:06:53.955
- is what you don't want is individuals that

176
00:06:53.955 --> 00:06:56.194
- may have a small pocket of data in

177
00:06:56.194 --> 00:06:59.160
- one location off, you know, exposing data to

178
00:06:59.160 --> 00:07:02.220
- public LLMs, etcetera. And so, so that unification

179
00:07:02.360 --> 00:07:04.600
- and management piece is a really critical part.

180
00:07:04.600 --> 00:07:07.000
- Unstructured data is the fastest growing type of

181
00:07:07.000 --> 00:07:08.620
- data across all organizations.

182
00:07:09.240 --> 00:07:10.379
- And so just getting

183
00:07:10.680 --> 00:07:13.394
- IT pros in particular, like IT infrastructure

184
00:07:13.694 --> 00:07:15.454
- leaders who we deal with on a daily

185
00:07:15.454 --> 00:07:17.615
- basis, just struggle to get their hands around

186
00:07:17.615 --> 00:07:19.634
- the data that they have within the organization.

187
00:07:20.014 --> 00:07:22.595
- So when you start there, okay, maybe now

188
00:07:22.654 --> 00:07:25.454
- I've I understand that I have 200 terabytes

189
00:07:25.454 --> 00:07:27.750
- here, a 100 terabytes there, blah, blah, blah.

190
00:07:27.829 --> 00:07:29.110
- It's not enough just to know where the

191
00:07:29.110 --> 00:07:30.709
- data lives. And this is one of the

192
00:07:30.709 --> 00:07:32.949
- big areas. In fact, one of the things

193
00:07:32.949 --> 00:07:35.110
- that we're really excited about is we actually

194
00:07:35.110 --> 00:07:38.089
- just announced an acquisition specifically going after

195
00:07:38.389 --> 00:07:40.230
- this challenge of how do I get better

196
00:07:40.230 --> 00:07:42.569
- insight into what is in the data itself?

197
00:07:42.664 --> 00:07:44.904
- And so being able to actually scan the

198
00:07:44.904 --> 00:07:46.524
- data, to be able to understand,

199
00:07:46.985 --> 00:07:48.925
- is there personally identifiable

200
00:07:49.384 --> 00:07:52.345
- information in there? Are there critical file types

201
00:07:52.345 --> 00:07:54.345
- to deal with government entities, etcetera, or are

202
00:07:54.345 --> 00:07:56.584
- they looking at ITAR data? So being able

203
00:07:56.584 --> 00:07:59.269
- to understand the data and then classify that

204
00:07:59.269 --> 00:08:00.810
- data and then ideally

205
00:08:01.189 --> 00:08:03.110
- tag that data in a way in which

206
00:08:03.110 --> 00:08:05.189
- you now have some control over it. So

207
00:08:05.189 --> 00:08:07.349
- it was interesting, we had a couple of

208
00:08:07.349 --> 00:08:09.430
- our customer advisory board meetings. We do one

209
00:08:09.430 --> 00:08:11.175
- in Europe and one year in The US

210
00:08:11.254 --> 00:08:13.814
- back in the spring, the dominant theme and

211
00:08:13.814 --> 00:08:17.014
- we're talking to VP of infrastructure types, chief

212
00:08:17.014 --> 00:08:18.875
- security officer types, CIOs.

213
00:08:19.414 --> 00:08:22.474
- The dominant theme was governance, was

214
00:08:22.854 --> 00:08:26.074
- we can't deploy AI broadly across our organization

215
00:08:26.134 --> 00:08:28.850
- if we don't have visibility into the data,

216
00:08:28.850 --> 00:08:30.930
- who has access to it certainly, but also

217
00:08:30.930 --> 00:08:32.769
- what is in the data, and then a

218
00:08:32.769 --> 00:08:34.470
- mechanism to actually control

219
00:08:35.090 --> 00:08:37.330
- how people are leveraging that data. Right? Because

220
00:08:37.330 --> 00:08:40.065
- just because someone has access inside of active

221
00:08:40.065 --> 00:08:41.825
- directory, as far as they're concerned to a

222
00:08:41.825 --> 00:08:43.584
- set of data again, doesn't mean you want

223
00:08:43.584 --> 00:08:46.065
- them using it with Copilot, with Chat GPT,

224
00:08:46.065 --> 00:08:48.784
- with Claude, etcetera. That's the number one challenge

225
00:08:48.784 --> 00:08:51.024
- that we hear across the board. And then

226
00:08:51.024 --> 00:08:51.924
- you get into,

227
00:08:52.384 --> 00:08:54.370
- okay, well, now I have visibility.

228
00:08:54.990 --> 00:08:57.149
- How do I start to think about optimal

229
00:08:57.149 --> 00:08:59.549
- use of that data with AI? You guys

230
00:08:59.549 --> 00:09:01.710
- are, I'm sure, very familiar with the term

231
00:09:01.710 --> 00:09:04.830
- of context in AI. Right? So if you

232
00:09:04.830 --> 00:09:06.509
- go and we saw lots of customers try

233
00:09:06.509 --> 00:09:08.675
- and do this. Hey, I have a volume

234
00:09:08.975 --> 00:09:12.735
- of our enterprise file data and there's millions

235
00:09:12.735 --> 00:09:14.654
- of files in there. You're not gonna go

236
00:09:14.654 --> 00:09:15.875
- just throw Copilot

237
00:09:16.175 --> 00:09:18.254
- at a massive volume of all that data.

238
00:09:18.254 --> 00:09:20.095
- You'll blow through your context window in a

239
00:09:20.095 --> 00:09:23.810
- heartbeat, you'll get hallucinations, etcetera. And so once

240
00:09:23.810 --> 00:09:25.490
- you think about it from a security and

241
00:09:25.490 --> 00:09:28.050
- risk reduction perspective around governance, then you start

242
00:09:28.050 --> 00:09:30.129
- saying, well, how do I optimize how I'm

243
00:09:30.129 --> 00:09:33.170
- deploying AI services for our customers? And so

244
00:09:33.170 --> 00:09:36.050
- good example, we had one local customer here

245
00:09:36.050 --> 00:09:38.544
- on the Boston area. Their legal department wanted

246
00:09:38.544 --> 00:09:41.424
- to use hundreds of thousands of prior legal

247
00:09:41.424 --> 00:09:41.924
- documents,

248
00:09:42.304 --> 00:09:45.105
- very classic sort of rag institutional knowledge use

249
00:09:45.105 --> 00:09:47.684
- case. I wanna go generate new contracts automatically.

250
00:09:48.065 --> 00:09:49.024
- 300,000

251
00:09:49.024 --> 00:09:51.105
- files in this set of directories. They tried

252
00:09:51.105 --> 00:09:52.649
- to index that using

253
00:09:52.950 --> 00:09:55.669
- the graph connector going up into Microsoft semantic

254
00:09:55.669 --> 00:09:57.909
- index and then throw Copilot at it, and

255
00:09:57.909 --> 00:09:59.509
- it did not go so well. And so

256
00:09:59.509 --> 00:10:01.350
- this is where you now start thinking, like

257
00:10:01.350 --> 00:10:04.070
- we talk about activating data for AI, that

258
00:10:04.070 --> 00:10:05.850
- curation that you can do upfront,

259
00:10:06.345 --> 00:10:08.664
- Understanding, okay, we got the legal team. We

260
00:10:08.664 --> 00:10:10.985
- know that there are sets of folders within

261
00:10:10.985 --> 00:10:13.625
- that volume, a 100,000 files. How do we

262
00:10:13.625 --> 00:10:15.865
- help them understand which of those files are

263
00:10:15.865 --> 00:10:17.945
- gonna be useful for their scenarios so they're

264
00:10:17.945 --> 00:10:19.945
- not blowing through their context windows and they're

265
00:10:19.945 --> 00:10:22.009
- not getting AI slop as a result of

266
00:10:22.009 --> 00:10:23.450
- it. So these are some of the challenge

267
00:10:23.450 --> 00:10:25.610
- we hear. Governance is the biggest one, I'd

268
00:10:25.610 --> 00:10:27.850
- say, right up front and then leading through

269
00:10:27.850 --> 00:10:30.089
- to how do I start showing the business

270
00:10:30.089 --> 00:10:32.669
- that we're actually getting a return on leveraging

271
00:10:32.730 --> 00:10:35.129
- AI with all of this unstructured. Yeah. So

272
00:10:35.129 --> 00:10:38.075
- is there a process that you guys have

273
00:10:38.075 --> 00:10:40.075
- found too that works really well? Because it's

274
00:10:40.075 --> 00:10:42.315
- interesting to hear you talk about that because

275
00:10:42.315 --> 00:10:44.654
- I have a long history in SharePoint

276
00:10:45.035 --> 00:10:48.075
- and viewing Purview and having those same types

277
00:10:48.075 --> 00:10:51.610
- of conversations, but it's SharePoint sensitivity labels and

278
00:10:51.909 --> 00:10:54.569
- working through it that way. It sounds like

279
00:10:55.029 --> 00:10:56.949
- you've done a bunch of that from that

280
00:10:56.949 --> 00:10:59.589
- unstructured data side of it and really helping

281
00:10:59.589 --> 00:11:00.089
- customers

282
00:11:00.789 --> 00:11:03.044
- work through that process. And curious if you

283
00:11:03.044 --> 00:11:06.424
- have, like, tips or tricks or ways that

284
00:11:06.644 --> 00:11:07.865
- you help people

285
00:11:08.165 --> 00:11:11.924
- start thinking about that with that particular data

286
00:11:11.924 --> 00:11:13.684
- to get to that point where they can

287
00:11:13.684 --> 00:11:15.389
- start using it with AI. You got it.

288
00:11:15.470 --> 00:11:17.470
- And so we think about and we've spent

289
00:11:17.470 --> 00:11:19.389
- a lot of time, like, in with customers,

290
00:11:19.389 --> 00:11:21.309
- and we see there are different folks in

291
00:11:21.309 --> 00:11:23.149
- the industry that sort of have different frameworks

292
00:11:23.149 --> 00:11:24.750
- of trying to sort of think about what

293
00:11:24.750 --> 00:11:27.309
- is being ready for AI actually mean. We've

294
00:11:27.309 --> 00:11:29.549
- kind of developed our own, which we like

295
00:11:29.629 --> 00:11:32.184
- we call it PSYCHIC. So it's a long

296
00:11:32.184 --> 00:11:34.424
- acronym, but bear with me. So the first

297
00:11:34.424 --> 00:11:35.725
- one's easy, proprietary.

298
00:11:36.105 --> 00:11:38.345
- Right? Everyone you gotta be using your data.

299
00:11:38.345 --> 00:11:40.985
- That's gold. Right? Everybody can choose a different

300
00:11:40.985 --> 00:11:41.884
- model nowadays.

301
00:11:42.345 --> 00:11:45.004
- Everyone typically has access to the same models.

302
00:11:45.170 --> 00:11:47.330
- The differentiation is gonna be you're working with

303
00:11:47.330 --> 00:11:50.930
- your organization's proprietary data. Great. Like, everyone starts

304
00:11:50.930 --> 00:11:52.769
- there. The s is secure, and this is

305
00:11:52.769 --> 00:11:55.250
- the governed piece. So, like, don't go down

306
00:11:55.250 --> 00:11:57.330
- a path of trying to think about how

307
00:11:57.330 --> 00:11:59.125
- you make data AI ready until you feel

308
00:11:59.125 --> 00:12:00.644
- like you you have your hands around it

309
00:12:00.644 --> 00:12:02.664
- from a security and from a governance

310
00:12:03.285 --> 00:12:05.285
- perspective. Yours, which you might say, well, what's

311
00:12:05.285 --> 00:12:07.544
- the difference between yours and proprietary?

312
00:12:07.924 --> 00:12:09.924
- Yours is, do you have a process in

313
00:12:09.924 --> 00:12:11.524
- place that makes sure that your data is

314
00:12:11.524 --> 00:12:14.850
- never leaving your security perimeter? You're not creating

315
00:12:14.910 --> 00:12:16.750
- copies that are now put up in somebody

316
00:12:16.750 --> 00:12:17.250
- else's

317
00:12:17.710 --> 00:12:21.070
- cloud environment. You're not exposing your proprietary data.

318
00:12:21.070 --> 00:12:23.170
- And there have been lots of cases of

319
00:12:23.470 --> 00:12:24.690
- developers and organizations

320
00:12:25.070 --> 00:12:27.889
- unknowingly exposing source code out to public LLMs,

321
00:12:27.950 --> 00:12:29.325
- right? And now that's in the public domain.

322
00:12:29.404 --> 00:12:31.085
- Domain. So that kind of thing. So yours

323
00:12:31.085 --> 00:12:33.485
- is very important. Then we think about being

324
00:12:33.485 --> 00:12:34.304
- able to

325
00:12:34.924 --> 00:12:35.424
- have

326
00:12:35.725 --> 00:12:38.465
- current data, and that may sound silly, but

327
00:12:38.524 --> 00:12:41.085
- current data, meaning this is the living, breathing

328
00:12:41.085 --> 00:12:42.845
- file system that your business runs on. Do

329
00:12:42.845 --> 00:12:44.820
- you have pipelines in place, or do you

330
00:12:44.820 --> 00:12:46.660
- have processes to make sure you're always getting

331
00:12:46.660 --> 00:12:49.779
- the freshest data into whatever solution you're using?

332
00:12:49.779 --> 00:12:51.480
- Because if you're making business decisions

333
00:12:52.019 --> 00:12:54.440
- powered by AI, whether that's just an individual

334
00:12:54.500 --> 00:12:57.264
- using ChattyPT, Copilot, etcetera, or you're all the

335
00:12:57.264 --> 00:12:59.585
- way into a multi step agentic workflow built

336
00:12:59.585 --> 00:13:01.904
- in Foundry, etcetera. Like you've got to have

337
00:13:01.904 --> 00:13:04.144
- the latest data and you'll sometimes see this

338
00:13:04.144 --> 00:13:05.825
- of, oh, just throw a copy of that

339
00:13:05.825 --> 00:13:07.664
- into your data lake and you're good. Like,

340
00:13:07.664 --> 00:13:09.539
- no, it should be coming off of like

341
00:13:09.620 --> 00:13:12.339
- the real conveyor belt of work that's being

342
00:13:12.339 --> 00:13:15.139
- done around that data. The I in psychic

343
00:13:15.139 --> 00:13:17.779
- is indexed. So this is really important. And

344
00:13:17.779 --> 00:13:20.179
- that is back to my example of the

345
00:13:20.179 --> 00:13:22.740
- legal department trying to analyze 300,000

346
00:13:22.740 --> 00:13:24.804
- documents with Copilot. If you don't have an

347
00:13:24.804 --> 00:13:26.644
- index that sits over top of that, you're

348
00:13:26.644 --> 00:13:29.065
- going to blow your tokens doing the research,

349
00:13:29.284 --> 00:13:32.325
- not research but research problem, right? And so

350
00:13:32.325 --> 00:13:34.565
- an index and this is a classic RAG

351
00:13:34.565 --> 00:13:36.565
- model, right, on the SharePoint case? Yeah, you

352
00:13:36.565 --> 00:13:38.459
- get the benefit of SharePoint. But what about

353
00:13:38.459 --> 00:13:40.620
- all that data that's not on SharePoint, right?

354
00:13:40.620 --> 00:13:42.940
- Is there an index that sits behind it

355
00:13:42.940 --> 00:13:45.259
- so that AI is reasoning over the most

356
00:13:45.259 --> 00:13:47.100
- relevant files, not just trying to find the

357
00:13:47.100 --> 00:13:49.579
- relevant files in the first place? H stands

358
00:13:49.579 --> 00:13:51.820
- for hybrid. So this is an interesting one.

359
00:13:51.820 --> 00:13:55.175
- So in our world, unstructured data management has

360
00:13:55.175 --> 00:13:56.455
- to be hybrid. And what I mean by

361
00:13:56.455 --> 00:13:57.975
- that is you wanna be able to take

362
00:13:57.975 --> 00:14:00.634
- advantage of the cloud for all of the

363
00:14:00.695 --> 00:14:03.335
- scale, the cost efficiencies you get of object

364
00:14:03.335 --> 00:14:05.735
- storage, but your people are doing work all

365
00:14:05.735 --> 00:14:07.720
- around the world out at the edge. Right?

366
00:14:07.720 --> 00:14:09.000
- And so you have to have a solution

367
00:14:09.000 --> 00:14:11.480
- that can make that data available wherever those

368
00:14:11.480 --> 00:14:13.639
- people are. AI is no different and we're

369
00:14:13.639 --> 00:14:16.220
- seeing it in spades today. Right? Whether it's

370
00:14:16.360 --> 00:14:18.220
- challenges getting access to GPUs,

371
00:14:18.679 --> 00:14:20.985
- whether it's cost of compute running in the

372
00:14:20.985 --> 00:14:22.985
- cloud and maybe you decided you wanna actually

373
00:14:22.985 --> 00:14:24.745
- do some proof of concepts in the cloud,

374
00:14:24.745 --> 00:14:27.065
- but then actually run your models on prem

375
00:14:27.065 --> 00:14:28.985
- and we see this with our customers. So

376
00:14:28.985 --> 00:14:31.065
- you gotta have something that accounts for the

377
00:14:31.065 --> 00:14:33.705
- fact that unstructured data that's hybrid and AI

378
00:14:33.705 --> 00:14:35.450
- needs to be hybrid as well. And then

379
00:14:35.450 --> 00:14:37.210
- the last one is an interesting one, which

380
00:14:37.210 --> 00:14:39.850
- is what we call continuous. So an interesting

381
00:14:39.850 --> 00:14:41.210
- thing, you can think of it sort of

382
00:14:41.210 --> 00:14:43.370
- like the iceberg analogy, right? The tip of

383
00:14:43.370 --> 00:14:45.289
- the iceberg is your current data, all the

384
00:14:45.289 --> 00:14:47.529
- active working data, all of the design files

385
00:14:47.529 --> 00:14:49.574
- people are actively working on, but there are

386
00:14:49.574 --> 00:14:52.294
- millions and millions of versions of that data

387
00:14:52.294 --> 00:14:54.695
- that live in a platform like ours that

388
00:14:54.695 --> 00:14:57.414
- keeps unlimited versions of that data. So think

389
00:14:57.414 --> 00:14:59.654
- about that from an AI perspective, right? Where

390
00:14:59.654 --> 00:15:02.294
- instead of just giving AI visibility just to

391
00:15:02.294 --> 00:15:04.389
- what's above the surface, if it could tap

392
00:15:04.389 --> 00:15:06.230
- into all that richness that lives in the

393
00:15:06.230 --> 00:15:08.309
- history, what could that do in terms of

394
00:15:08.309 --> 00:15:10.790
- whether you're actually doing model training, right, where

395
00:15:10.790 --> 00:15:13.750
- getting access to data is so critical, or

396
00:15:13.750 --> 00:15:16.470
- whether you're looking at, like, more sophisticated scenarios

397
00:15:16.470 --> 00:15:18.389
- where you wanna trace the lineage of a

398
00:15:18.389 --> 00:15:18.889
- project

399
00:15:19.365 --> 00:15:21.845
- and that lives in that file history. So

400
00:15:21.845 --> 00:15:23.684
- that's a sort of tongue and cheek around,

401
00:15:23.684 --> 00:15:26.004
- yeah, it's psychic, but but we we actually

402
00:15:26.004 --> 00:15:28.325
- have found that is a really great way

403
00:15:28.325 --> 00:15:30.485
- to engage with customers to help them start

404
00:15:30.485 --> 00:15:32.990
- thinking about their own environment. And we actually

405
00:15:32.990 --> 00:15:34.829
- use sort of an assessment with them to

406
00:15:34.829 --> 00:15:36.429
- kind of sit down and just talk about,

407
00:15:36.429 --> 00:15:39.230
- hey, let's evaluate where you sit here. And

408
00:15:39.230 --> 00:15:41.649
- it actually gives particularly like our IT infrastructure

409
00:15:41.789 --> 00:15:43.389
- leaders a way to go kind of talk

410
00:15:43.389 --> 00:15:45.105
- to the rest of the business too, like

411
00:15:45.345 --> 00:15:48.225
- security teams, compliance teams, etcetera. Whole bunch of

412
00:15:48.225 --> 00:15:50.945
- moving parts there. Yes. Agentic side of it

413
00:15:50.945 --> 00:15:52.865
- is one of the things that always kinda

414
00:15:52.865 --> 00:15:54.784
- comes back for me. Like, there's always humans

415
00:15:54.784 --> 00:15:56.544
- in the loop, but when you move on

416
00:15:56.544 --> 00:15:58.004
- to more agentic workflows

417
00:15:58.464 --> 00:15:58.964
- and

418
00:15:59.424 --> 00:16:00.485
- you go from

419
00:16:01.019 --> 00:16:03.580
- arguably, like, deterministic and maybe a little less

420
00:16:03.580 --> 00:16:05.620
- problematic to maybe a little more problematic and

421
00:16:05.620 --> 00:16:08.000
- a little less deterministic, kinda having those guardrails

422
00:16:08.060 --> 00:16:10.540
- in place, getting things on. I see and

423
00:16:10.540 --> 00:16:12.220
- I experienced this myself. Like, a lot of

424
00:16:12.220 --> 00:16:14.264
- things that you start out with with agentic

425
00:16:14.264 --> 00:16:16.264
- workflows tend to kinda die on the vine.

426
00:16:16.264 --> 00:16:17.865
- You get far enough along with them, and

427
00:16:17.865 --> 00:16:19.785
- then they just kinda stop. And you had

428
00:16:19.785 --> 00:16:21.465
- shared, that's kind of as as we were

429
00:16:21.465 --> 00:16:22.985
- chatting about having you come on the show,

430
00:16:22.985 --> 00:16:24.585
- a little bit of research that Nasuni has

431
00:16:24.585 --> 00:16:25.865
- done. Like, there there was a stat in

432
00:16:25.865 --> 00:16:27.629
- there that really stood out to me that

433
00:16:27.789 --> 00:16:30.750
- 90% of the orgs are piloting AI agents,

434
00:16:30.750 --> 00:16:33.250
- but only 43 call those projects successful.

435
00:16:33.710 --> 00:16:35.710
- So that leaves this massive gap around some

436
00:16:35.710 --> 00:16:37.149
- of these areas that you talked about. It

437
00:16:37.149 --> 00:16:39.470
- could be a data problem, data quality, garbage

438
00:16:39.470 --> 00:16:42.004
- in, garbage out. Could be a governance problem,

439
00:16:42.144 --> 00:16:44.945
- a security, like, versioning, how many versions of

440
00:16:44.945 --> 00:16:46.625
- this do I have and where it lives

441
00:16:46.625 --> 00:16:48.945
- and that whole sprawl. So so I'm curious,

442
00:16:48.945 --> 00:16:51.105
- like, kinda tying this back to Nasuni and

443
00:16:51.105 --> 00:16:53.524
- what you all do, how does something

444
00:16:53.825 --> 00:16:54.325
- like

445
00:16:54.879 --> 00:16:57.519
- AI activate come in and let an agent

446
00:16:57.519 --> 00:16:58.019
- do

447
00:16:58.399 --> 00:16:59.379
- more or

448
00:16:59.759 --> 00:17:02.079
- have kind of better governance, things like that,

449
00:17:02.079 --> 00:17:04.000
- than it would over just a traditional file

450
00:17:04.000 --> 00:17:06.640
- share, whether it was SMB, Nasuni managed, and

451
00:17:06.640 --> 00:17:08.579
- anything like that? So first and foremost,

452
00:17:08.994 --> 00:17:10.855
- when we talk when I talk with customers,

453
00:17:11.154 --> 00:17:12.054
- part of that

454
00:17:12.355 --> 00:17:15.634
- disparity between the 90% that are experimenting versus

455
00:17:15.634 --> 00:17:17.714
- those that are actually seeing success, some of

456
00:17:17.714 --> 00:17:19.954
- it is the confidence to deploy it in

457
00:17:19.954 --> 00:17:21.095
- a meaningful enough

458
00:17:21.394 --> 00:17:21.894
- scenario

459
00:17:22.289 --> 00:17:23.970
- because they don't necessarily feel that they have

460
00:17:23.970 --> 00:17:27.009
- the governance frameworks in place. So, ultimately, those

461
00:17:27.009 --> 00:17:29.269
- might not be as business impactful

462
00:17:29.650 --> 00:17:31.809
- in terms of those agentic deployments, and so

463
00:17:31.809 --> 00:17:33.090
- they kind of die on the vine as

464
00:17:33.090 --> 00:17:35.490
- you said. Back to your question specifically around

465
00:17:35.490 --> 00:17:38.144
- AI activate, so a couple key things. Number

466
00:17:38.144 --> 00:17:40.625
- one are the the first kind of parts

467
00:17:40.625 --> 00:17:43.144
- of our psychic framework, proprietary obviously, it's all

468
00:17:43.265 --> 00:17:45.585
- it's your data. But number two is the

469
00:17:45.585 --> 00:17:48.164
- secure side. And so knowing that

470
00:17:48.625 --> 00:17:50.005
- security is enforced

471
00:17:50.490 --> 00:17:52.910
- pre what we call pre query, right? So

472
00:17:52.970 --> 00:17:55.069
- if I sit down and I wanna use

473
00:17:55.369 --> 00:17:57.930
- Copilot to interrogate a set of projects that

474
00:17:57.930 --> 00:17:59.849
- I've been working on and I wanna interact

475
00:17:59.849 --> 00:18:02.250
- with it over what could be a massive

476
00:18:02.250 --> 00:18:04.505
- volume that has two petabytes in it. We

477
00:18:04.505 --> 00:18:06.265
- have customers that have, you know, multi petabyte

478
00:18:06.265 --> 00:18:07.724
- volumes. First and foremost,

479
00:18:08.184 --> 00:18:10.184
- we know or the customer who's using AIA

480
00:18:10.184 --> 00:18:12.984
- Activate understands that I will never be able

481
00:18:12.984 --> 00:18:13.484
- to

482
00:18:13.944 --> 00:18:17.065
- search over, interact with, interrogate data that I

483
00:18:17.065 --> 00:18:19.919
- don't explicitly have permissions to touch. And that's

484
00:18:19.919 --> 00:18:21.539
- done before we ever

485
00:18:22.079 --> 00:18:24.339
- go search across an index of the data

486
00:18:24.399 --> 00:18:26.480
- because sometimes they'll see what people do, they'll

487
00:18:26.480 --> 00:18:28.000
- go do a big search, they'll get the

488
00:18:28.000 --> 00:18:30.399
- results and then they'll filter afterwards for the

489
00:18:30.399 --> 00:18:32.355
- permissions that is not a secure model. So

490
00:18:32.355 --> 00:18:33.734
- first and foremost is security.

491
00:18:34.035 --> 00:18:36.835
- Second is around the index piece. So I

492
00:18:36.835 --> 00:18:39.795
- already mentioned it, but AyaActivate is a true

493
00:18:39.795 --> 00:18:41.974
- SaaS service, it's all our infrastructure

494
00:18:42.515 --> 00:18:44.595
- and we're able to ensure a customer that

495
00:18:44.595 --> 00:18:46.210
- all they have to do is scope whatever

496
00:18:46.210 --> 00:18:47.809
- they're looking at. So we'll go back to

497
00:18:47.809 --> 00:18:50.130
- my example of the legal department. They're now

498
00:18:50.130 --> 00:18:52.289
- able to go pick the volume that they

499
00:18:52.289 --> 00:18:55.089
- wanna work on and then generate. So it

500
00:18:55.089 --> 00:18:57.889
- is back. You mentioned MCP. So it is

501
00:18:57.889 --> 00:19:00.715
- basically exposed through an MCP. So literally anything

502
00:19:00.715 --> 00:19:02.134
- that can talk to an MCP

503
00:19:02.515 --> 00:19:04.755
- can now talk to AI activate and now

504
00:19:04.755 --> 00:19:07.075
- leverage this data knowing that the permissions are

505
00:19:07.075 --> 00:19:09.235
- enforced against it, whether that's an individual or

506
00:19:09.235 --> 00:19:11.575
- an agent operating on behalf of an individual.

507
00:19:11.795 --> 00:19:12.295
- So

508
00:19:12.609 --> 00:19:13.970
- right out of the gate, you've got the

509
00:19:13.970 --> 00:19:16.230
- security in place, you've got the permissions enforcement

510
00:19:16.289 --> 00:19:18.690
- in place, and you've got the benefit of

511
00:19:18.690 --> 00:19:20.609
- the index, which again means you're gonna get

512
00:19:20.609 --> 00:19:22.369
- much, much better results than if you were

513
00:19:22.369 --> 00:19:25.009
- just throwing an LLM at a massive volume

514
00:19:25.009 --> 00:19:26.795
- of data. And the MCP part is really

515
00:19:26.795 --> 00:19:28.875
- important. Not only is it standards based, like,

516
00:19:28.875 --> 00:19:31.275
- we're a platform. Right? So I don't wanna

517
00:19:31.275 --> 00:19:32.714
- be in the game of trying to go

518
00:19:32.714 --> 00:19:34.335
- build hundreds of

519
00:19:34.634 --> 00:19:36.335
- vertical specific agents.

520
00:19:36.634 --> 00:19:38.555
- The world changes too fast. That's not our

521
00:19:38.555 --> 00:19:39.295
- core competency.

522
00:19:39.799 --> 00:19:41.240
- We want our customers to be able to

523
00:19:41.240 --> 00:19:43.720
- use best of breed AI services to plug

524
00:19:43.720 --> 00:19:45.659
- into our platform. And so that MCP

525
00:19:46.599 --> 00:19:48.359
- piece is really important. And then we built

526
00:19:48.359 --> 00:19:51.079
- some intelligence into there in terms of how

527
00:19:51.079 --> 00:19:53.019
- to take the prompts that come through

528
00:19:53.424 --> 00:19:55.904
- and optimize that query against that back end

529
00:19:55.904 --> 00:19:58.244
- index to get the best results, but also

530
00:19:58.545 --> 00:20:00.704
- token utilization and token efficiency is a big

531
00:20:00.704 --> 00:20:02.464
- thing. So making sure that it's gonna be

532
00:20:02.464 --> 00:20:05.184
- as performance so the customer again is spending

533
00:20:05.184 --> 00:20:07.105
- more time reasoning over the data versus just

534
00:20:07.105 --> 00:20:08.309
- trying to find the data. Now the last

535
00:20:08.309 --> 00:20:09.750
- bit that I'll say and I mentioned the

536
00:20:09.750 --> 00:20:11.849
- acquisition we made was a company called DriveIQ

537
00:20:12.470 --> 00:20:12.970
- and

538
00:20:13.349 --> 00:20:15.109
- this is the next bit for us which

539
00:20:15.109 --> 00:20:18.490
- is layering on that additional layer of governance

540
00:20:18.549 --> 00:20:20.549
- on top of it. Because permission enforcement is

541
00:20:20.549 --> 00:20:22.255
- great, But now we wanna go that next

542
00:20:22.255 --> 00:20:24.255
- layer, right? And so now a customer can

543
00:20:24.255 --> 00:20:26.414
- work with us and say, all those legal

544
00:20:26.414 --> 00:20:28.755
- documents, I'm gonna go in, I'm gonna discover,

545
00:20:28.974 --> 00:20:31.535
- classify that, tag it. I can even redact

546
00:20:31.535 --> 00:20:34.849
- individual files. So there might be some particular

547
00:20:35.070 --> 00:20:37.150
- field or a customer name or something that

548
00:20:37.150 --> 00:20:38.910
- you don't want that's in a legal contract

549
00:20:38.910 --> 00:20:41.230
- to ever be exposed to AI. Great. You

550
00:20:41.230 --> 00:20:43.630
- can redact the file and still leverage that

551
00:20:43.630 --> 00:20:45.470
- file on all the context it has with

552
00:20:45.470 --> 00:20:47.654
- AI, but in a governed fashion. So that's

553
00:20:47.654 --> 00:20:49.735
- the big piece. So AI activate is all

554
00:20:49.735 --> 00:20:51.275
- about that secure

555
00:20:51.575 --> 00:20:55.095
- index data, standards based connectivity for AI solutions

556
00:20:55.095 --> 00:20:57.434
- like Copilot. It plugs right into Foundry.

557
00:20:57.815 --> 00:20:59.909
- We have one of our customers is building

558
00:20:59.909 --> 00:21:03.109
- a multi step agentic workflow. They're in civil

559
00:21:03.109 --> 00:21:05.769
- engineering firm, multi step agentic workflow

560
00:21:06.069 --> 00:21:09.190
- where they're combining structured data with another agent

561
00:21:09.190 --> 00:21:11.589
- in the flow with the unstructured data retrieved

562
00:21:11.589 --> 00:21:14.085
- through AI activated. So we've got all that,

563
00:21:14.085 --> 00:21:15.845
- and then now we're gonna be layering in

564
00:21:15.845 --> 00:21:18.964
- the DriveIQ piece for that additional governance layer

565
00:21:18.964 --> 00:21:20.424
- for our protection and oversight.

566
00:21:22.164 --> 00:21:24.565
- If you are managing file storage today, you've

567
00:21:24.565 --> 00:21:27.684
- probably felt the pressure. Refresh cycles are harder

568
00:21:27.684 --> 00:21:30.740
- to plan, costs and complexity are increasing, especially

569
00:21:31.039 --> 00:21:33.299
- as file workloads scale across regions.

570
00:21:33.680 --> 00:21:37.039
- That's why Microsoft customers trust Nasuni. As an

571
00:21:37.039 --> 00:21:38.580
- Azure certified software,

572
00:21:38.960 --> 00:21:41.920
- Nasuni provides a global file system designed for

573
00:21:41.920 --> 00:21:43.299
- enterprise scale environments,

574
00:21:43.684 --> 00:21:47.125
- keeping file data secure, governed, and accessible across

575
00:21:47.125 --> 00:21:49.464
- Azure environments without added infrastructure

576
00:21:49.765 --> 00:21:50.265
- complexity.

577
00:21:50.644 --> 00:21:51.625
- It's the foundation

578
00:21:51.924 --> 00:21:54.964
- to keep your unstructured data stable even as

579
00:21:54.964 --> 00:21:55.704
- your infrastructure

580
00:21:56.005 --> 00:21:58.359
- evolves. To see how much you could save

581
00:21:58.359 --> 00:22:00.779
- while keeping your initiatives on track, visit

582
00:22:01.080 --> 00:22:04.680
- nasuni.com/tco.

583
00:22:04.680 --> 00:22:09.339
- That's nasuni.com/tco.

584
00:22:10.680 --> 00:22:13.394
- When you start thinking about, like, that index

585
00:22:13.454 --> 00:22:15.875
- and the DriveIQ stuff and sensitivity,

586
00:22:16.575 --> 00:22:18.575
- obviously, you have, like, Purview and you have

587
00:22:18.575 --> 00:22:20.515
- DLP over there. Is there,

588
00:22:20.894 --> 00:22:22.194
- like, an integration

589
00:22:22.494 --> 00:22:25.454
- there where it's respecting some of the same

590
00:22:25.454 --> 00:22:28.109
- labels or syncing up with Purview so you

591
00:22:28.109 --> 00:22:30.909
- can have the labels? Like, how does your

592
00:22:30.909 --> 00:22:32.289
- solution with DriveIQ,

593
00:22:32.750 --> 00:22:35.470
- AI activate, and some of that kinda work,

594
00:22:35.470 --> 00:22:36.929
- it sounds like, in partnership

595
00:22:37.230 --> 00:22:37.730
- with

596
00:22:38.044 --> 00:22:40.284
- some of the controls Microsoft already has there?

597
00:22:40.284 --> 00:22:42.605
- There's very much a better together story here.

598
00:22:42.605 --> 00:22:44.525
- A lot of the customers that DriveIQ worked

599
00:22:44.525 --> 00:22:45.664
- with prior to the acquisition

600
00:22:46.125 --> 00:22:48.684
- leverage Purview as well. Right? Because it's top

601
00:22:48.684 --> 00:22:50.684
- notch in terms of the m 30 three

602
00:22:50.684 --> 00:22:51.904
- sixty five estate.

603
00:22:52.230 --> 00:22:53.829
- And but you wanna have something where you're

604
00:22:53.829 --> 00:22:55.829
- able to leverage that work. And so, yeah,

605
00:22:55.829 --> 00:22:57.750
- it can talk and it actually implements the

606
00:22:57.750 --> 00:22:58.569
- same sensitivity

607
00:22:58.869 --> 00:23:02.150
- labeling scheme. And so there's compatibility across there,

608
00:23:02.150 --> 00:23:05.750
- but it extends obviously beyond the Microsoft estate

609
00:23:05.750 --> 00:23:07.855
- where Purview is best to this whole other

610
00:23:07.855 --> 00:23:10.414
- world of all the other types of unstructured

611
00:23:10.414 --> 00:23:12.835
- data that typically don't live outside of SharePoint.

612
00:23:12.974 --> 00:23:15.075
- And so that's the better together piece here.

613
00:23:15.294 --> 00:23:17.855
- And we have, again, we have customers that

614
00:23:17.855 --> 00:23:20.815
- have petabytes in SharePoint today, but then tens

615
00:23:20.815 --> 00:23:23.529
- and twenties of petabytes of other data that

616
00:23:23.529 --> 00:23:25.849
- doesn't live in SharePoint. And so DriveIQ can

617
00:23:25.849 --> 00:23:27.849
- really help to bridge that. And there's also

618
00:23:27.849 --> 00:23:31.049
- great integration. DriveIQ can integrate with cloud SaaS

619
00:23:31.049 --> 00:23:34.329
- services like SharePoint, like Google Drive, etcetera. And

620
00:23:34.329 --> 00:23:36.345
- so it can help. Sometimes you get customers

621
00:23:36.345 --> 00:23:38.664
- that run into challenges with overages because of

622
00:23:38.664 --> 00:23:40.345
- all the data they have put in SharePoint.

623
00:23:40.345 --> 00:23:42.845
- And so it can help migrate data onto

624
00:23:42.904 --> 00:23:45.065
- Nasuni to help archive that, for example. So

625
00:23:45.065 --> 00:23:47.144
- you have a gold master copy, but it

626
00:23:47.144 --> 00:23:49.570
- leaves behind what are essentially stubs. So So

627
00:23:49.570 --> 00:23:51.170
- that if for some reason someone did need

628
00:23:51.170 --> 00:23:53.250
- to go react access that file through SharePoint,

629
00:23:53.250 --> 00:23:54.769
- great. You can just rehydrate it right back

630
00:23:54.769 --> 00:23:57.650
- into SharePoint. So both from a kind of

631
00:23:57.650 --> 00:24:00.369
- data integration and migration perspective, but then also

632
00:24:00.369 --> 00:24:03.090
- in terms of consistency of leveraging that labeling

633
00:24:03.090 --> 00:24:05.095
- work that Purview is doing, there's a really

634
00:24:05.095 --> 00:24:07.255
- nice synergy there between the two. Okay. And

635
00:24:07.255 --> 00:24:08.934
- then when you're trying to, like, filter out

636
00:24:08.934 --> 00:24:11.335
- that content, are you then going and creating

637
00:24:11.335 --> 00:24:13.034
- rules and, like, DriveIQ

638
00:24:13.335 --> 00:24:15.194
- that says just don't ever surface

639
00:24:15.575 --> 00:24:18.454
- this customer name or Social Security numbers or

640
00:24:18.454 --> 00:24:19.595
- credit card numbers

641
00:24:20.500 --> 00:24:23.460
- through that MCP server? Today, so that we

642
00:24:23.460 --> 00:24:25.000
- literally just closed the acquisition.

643
00:24:25.299 --> 00:24:27.140
- So one of the first pieces that we're

644
00:24:27.140 --> 00:24:29.059
- gonna be working on from an integration perspective

645
00:24:29.059 --> 00:24:31.480
- is exactly that. So that act AI activate

646
00:24:31.619 --> 00:24:32.119
- can

647
00:24:32.505 --> 00:24:34.744
- query in real time and understand here are

648
00:24:34.744 --> 00:24:36.664
- the policies that have been set up, and

649
00:24:36.664 --> 00:24:38.585
- then there are what are called action sets.

650
00:24:38.585 --> 00:24:40.265
- And so they can be so it's not

651
00:24:40.265 --> 00:24:42.585
- just a passive scanning solution. I already mentioned

652
00:24:42.585 --> 00:24:44.664
- that it can redact, but it can also

653
00:24:44.664 --> 00:24:47.279
- take action. But also in this scenario, yeah,

654
00:24:47.359 --> 00:24:49.440
- you might have a policy that says, no

655
00:24:49.440 --> 00:24:51.840
- files with this tag can ever be exposed

656
00:24:51.840 --> 00:24:54.320
- to Copilot. Then great. We can enforce that

657
00:24:54.320 --> 00:24:56.559
- at query time as part of AI activate

658
00:24:56.559 --> 00:24:58.160
- as well. That's one of the first scenarios

659
00:24:58.160 --> 00:24:59.680
- that we're going after. I got one of

660
00:24:59.680 --> 00:25:01.734
- the really powerful things. Right? You bring I

661
00:25:01.734 --> 00:25:03.974
- I mean, MCP in this case, like, helps

662
00:25:03.974 --> 00:25:05.894
- you to be kind of that proxy layer

663
00:25:05.894 --> 00:25:07.275
- with a common set of tools.

664
00:25:07.654 --> 00:25:09.595
- The nice thing about agents,

665
00:25:10.214 --> 00:25:11.734
- for the most part is they're just text

666
00:25:11.734 --> 00:25:13.414
- files. Right? There are definitions that are out

667
00:25:13.414 --> 00:25:15.869
- there seeing carry them, evaluate them, and test

668
00:25:15.869 --> 00:25:18.369
- them against various models with kinda that consistent

669
00:25:18.910 --> 00:25:21.630
- interface and layer that sits out there. So

670
00:25:21.630 --> 00:25:23.789
- it's been interesting to see that ecosystem take

671
00:25:23.789 --> 00:25:26.350
- off like a hockey stick. I think Anthropic

672
00:25:26.350 --> 00:25:27.950
- did a pretty good job, like, spinning that

673
00:25:27.950 --> 00:25:29.684
- up and saying, hey. Here's the thing for

674
00:25:29.684 --> 00:25:31.765
- everybody. But it's super powerful. Right? Because somebody

675
00:25:31.765 --> 00:25:33.065
- can onboard to the platform,

676
00:25:33.365 --> 00:25:35.625
- and the tools that your MCP

677
00:25:36.244 --> 00:25:39.684
- server or servers expose, they just become consistent

678
00:25:39.684 --> 00:25:42.025
- tool sets, not only for agents, but also

679
00:25:42.244 --> 00:25:44.759
- humans who are out there using a harness,

680
00:25:44.759 --> 00:25:46.859
- if it's Claude or ChatGPT

681
00:25:47.319 --> 00:25:49.639
- or they're in a CLI and any anything

682
00:25:49.639 --> 00:25:51.880
- like that. Exactly right. And talking about agents,

683
00:25:51.880 --> 00:25:53.960
- one of the other really important things too

684
00:25:53.960 --> 00:25:55.099
- is from our perspective,

685
00:25:55.525 --> 00:25:57.704
- agents are just a more powerful

686
00:25:58.164 --> 00:26:00.404
- version of people that don't need to sleep.

687
00:26:00.404 --> 00:26:03.204
- Right? So they execute file system operations against

688
00:26:03.204 --> 00:26:05.285
- the file system. That could be creates, deletes,

689
00:26:05.285 --> 00:26:08.184
- modifications, etcetera. So the really nice thing is

690
00:26:08.325 --> 00:26:10.164
- all of that activity that an agent might

691
00:26:10.164 --> 00:26:12.539
- be driving against it gets audited just like

692
00:26:12.539 --> 00:26:15.259
- every other user client application that's touching the

693
00:26:15.259 --> 00:26:18.539
- file system. And so having that visibility, which

694
00:26:18.539 --> 00:26:19.839
- we have within the platform

695
00:26:20.299 --> 00:26:22.000
- today with something we call FileIQ,

696
00:26:22.460 --> 00:26:24.240
- we now have that audit trail.

697
00:26:24.585 --> 00:26:26.904
- So if an agent happened to go do

698
00:26:26.904 --> 00:26:30.044
- something crazy like delete 10 different directories

699
00:26:30.424 --> 00:26:30.924
- accidentally,

700
00:26:31.384 --> 00:26:32.984
- we have that full audit trail. We know

701
00:26:32.984 --> 00:26:34.984
- exactly what files were impacted, and then we

702
00:26:34.984 --> 00:26:37.065
- can restore them very fast by virtue of

703
00:26:37.065 --> 00:26:38.849
- the way that our product is marketed. Right?

704
00:26:38.929 --> 00:26:41.410
- Because we're never copying data back from backup

705
00:26:41.410 --> 00:26:43.730
- somewhere. We're just putting a shifter a pointer

706
00:26:43.730 --> 00:26:45.890
- on the file system, pointing to the previous

707
00:26:45.890 --> 00:26:47.730
- last known good version of that data, and

708
00:26:47.730 --> 00:26:49.329
- then we can restore it right away. And

709
00:26:49.329 --> 00:26:51.884
- so that's where, in a lot of ways,

710
00:26:51.964 --> 00:26:53.724
- our CEO, Sam King, has a nice phrase,

711
00:26:53.724 --> 00:26:54.845
- which is we feel in a lot of

712
00:26:54.845 --> 00:26:56.384
- ways we're made for this moment.

713
00:26:56.684 --> 00:26:58.545
- A lot of the work that's gone into

714
00:26:58.605 --> 00:26:59.904
- thinking just about

715
00:27:00.365 --> 00:27:00.865
- modernization

716
00:27:01.164 --> 00:27:02.625
- of file data infrastructure

717
00:27:03.325 --> 00:27:06.045
- and the data protection and recoverability that we've

718
00:27:06.045 --> 00:27:09.069
- built into the platform is really nicely matched

719
00:27:09.069 --> 00:27:11.710
- with this agentic world and some of the

720
00:27:11.710 --> 00:27:13.549
- fears that people have of what happens if

721
00:27:13.549 --> 00:27:15.230
- this stuff gets out of control. I'm a

722
00:27:15.230 --> 00:27:17.950
- little curious. So in this world where you

723
00:27:17.950 --> 00:27:18.690
- plug in

724
00:27:19.309 --> 00:27:20.910
- more and more on the top, I think

725
00:27:20.910 --> 00:27:22.674
- one of the things that I see folks

726
00:27:22.674 --> 00:27:25.315
- struggle with is just getting their environment aligned

727
00:27:25.315 --> 00:27:26.835
- in the first place. Like, you talk about

728
00:27:26.835 --> 00:27:29.634
- Copilot, not just letting up licenses, but it's

729
00:27:29.634 --> 00:27:32.035
- back to the the old SharePoint problem of,

730
00:27:32.035 --> 00:27:34.434
- like, what were the permissions on my SharePoint

731
00:27:34.434 --> 00:27:36.515
- sites and and what got overexposed and and

732
00:27:36.515 --> 00:27:38.130
- what's out there. So if

733
00:27:38.529 --> 00:27:40.710
- there's a customer, they're in the Microsoft ecosystem,

734
00:27:40.849 --> 00:27:43.009
- they're out there, they're doing SharePoint, they're doing

735
00:27:43.009 --> 00:27:44.710
- Copilot, they have file shares,

736
00:27:45.089 --> 00:27:46.769
- kind of these things. If they come and

737
00:27:46.769 --> 00:27:47.750
- they plug in

738
00:27:48.210 --> 00:27:50.724
- AI activate drive IQ, to to the degree

739
00:27:50.724 --> 00:27:52.085
- you can say because you you just got

740
00:27:52.085 --> 00:27:54.005
- it. You know, what kinda just works for

741
00:27:54.005 --> 00:27:56.325
- them there based on what they've already done?

742
00:27:56.325 --> 00:27:58.244
- Or is there kinda more of a lift

743
00:27:58.244 --> 00:28:00.005
- to come in and say, hey. There's more

744
00:28:00.005 --> 00:28:02.964
- prereqs here to really be able to take

745
00:28:02.964 --> 00:28:04.990
- advantage of the data that's out there. Let's

746
00:28:04.990 --> 00:28:06.829
- start with a customer that's already a Nasuni

747
00:28:06.829 --> 00:28:08.349
- customer and sort of how do they plug

748
00:28:08.349 --> 00:28:11.390
- into this this world of AI and different

749
00:28:11.390 --> 00:28:11.890
- solutions.

750
00:28:12.269 --> 00:28:14.450
- So, again, the the beauty of it is

751
00:28:14.670 --> 00:28:16.589
- the starting point for Nasuni, one of the

752
00:28:16.589 --> 00:28:19.234
- biggest value propositions is data unification. Right? So

753
00:28:19.234 --> 00:28:20.595
- that we have what we call a global

754
00:28:20.595 --> 00:28:22.835
- name space over all of the data. And

755
00:28:22.835 --> 00:28:24.535
- so by virtue of having that,

756
00:28:24.835 --> 00:28:27.394
- now being able to just literally plug AI

757
00:28:27.394 --> 00:28:29.394
- activate into that and be able to build

758
00:28:29.394 --> 00:28:30.994
- these indexes and the fact that they don't

759
00:28:30.994 --> 00:28:32.899
- have to deploy infrastructure to do it, that

760
00:28:32.899 --> 00:28:34.980
- makes it real easy. The harder scenario is

761
00:28:34.980 --> 00:28:37.220
- the new customer, right? Who's coming and talking

762
00:28:37.220 --> 00:28:38.740
- to us and maybe they have a bunch

763
00:28:38.740 --> 00:28:40.579
- of SharePoint today, they've got a bunch of

764
00:28:40.579 --> 00:28:43.220
- Windows file servers and their remote offices, they

765
00:28:43.220 --> 00:28:45.424
- might have some NetApp and some Isilon. One

766
00:28:45.424 --> 00:28:48.144
- of the really nice things about DriveIQ and

767
00:28:48.144 --> 00:28:49.585
- one of the very important things for me

768
00:28:49.585 --> 00:28:51.825
- in particular is we were evaluating partners and

769
00:28:51.825 --> 00:28:53.605
- then eventually targets for acquisition

770
00:28:54.065 --> 00:28:57.265
- is it's really important that solution can also

771
00:28:57.265 --> 00:28:59.345
- provide huge value to data that isn't yet

772
00:28:59.345 --> 00:29:02.089
- on the Citi platform. So that same discovery

773
00:29:02.089 --> 00:29:04.250
- that you wanna do upfront to sort of

774
00:29:04.250 --> 00:29:06.490
- get your hands around what data can be

775
00:29:06.490 --> 00:29:08.169
- optimized for AI, well, that's the same thing

776
00:29:08.169 --> 00:29:09.690
- we can now go do with customers when

777
00:29:09.690 --> 00:29:11.849
- they're thinking about choosing a platform like the

778
00:29:11.849 --> 00:29:13.710
- Citi. So we can go do that discovery

779
00:29:13.769 --> 00:29:15.309
- upfront. We can help them

780
00:29:15.694 --> 00:29:18.355
- identify not just where do they have unstructured

781
00:29:18.414 --> 00:29:20.494
- data, but what's in that? What data has

782
00:29:20.494 --> 00:29:22.894
- been accessed recently? What of that is wrought,

783
00:29:22.894 --> 00:29:26.494
- right? So redundant, obsolete, trivial data. And so

784
00:29:26.494 --> 00:29:29.454
- if they have 20 petabytes of unstructured data

785
00:29:29.454 --> 00:29:32.019
- across these different places, alright, maybe 10 of

786
00:29:32.019 --> 00:29:35.140
- those are really critical operational data and we

787
00:29:35.140 --> 00:29:36.740
- can help them get their set, their arms

788
00:29:36.740 --> 00:29:38.440
- around it. And oh, by the way, DriveIQ

789
00:29:38.500 --> 00:29:40.420
- also migrates data. In fact, it migrates a

790
00:29:40.420 --> 00:29:42.420
- lot of data into SharePoint and into the

791
00:29:42.420 --> 00:29:43.480
- Microsoft ecosystem

792
00:29:43.934 --> 00:29:46.335
- today. So we have that single solution that

793
00:29:46.335 --> 00:29:47.394
- can give them visibility

794
00:29:47.934 --> 00:29:50.494
- to classify, tag, get their arms around it,

795
00:29:50.494 --> 00:29:52.434
- move the data to the Nasuni platform.

796
00:29:52.734 --> 00:29:54.654
- And then now, once we have it on

797
00:29:54.654 --> 00:29:56.815
- the platform, the the same policies that they

798
00:29:56.815 --> 00:29:59.019
- wanna define in that initial phase can all

799
00:29:59.019 --> 00:30:00.859
- be enforced once the data sits inside of

800
00:30:00.859 --> 00:30:03.180
- the Nasuni platform as well. It sounds like

801
00:30:03.180 --> 00:30:05.180
- from even where Nasuni started to what you're

802
00:30:05.180 --> 00:30:07.519
- doing today, that's one of the big advantages

803
00:30:08.059 --> 00:30:10.700
- of Nasuni is that you kinda got your

804
00:30:10.700 --> 00:30:13.154
- start in a data platform. So you have

805
00:30:13.154 --> 00:30:16.454
- a really good grasp on data storage, on

806
00:30:16.755 --> 00:30:19.554
- all things related to data. And now with

807
00:30:19.554 --> 00:30:22.694
- AI, you're just able to really build these

808
00:30:23.075 --> 00:30:23.575
- products,

809
00:30:24.515 --> 00:30:26.630
- integrate these products, the DriveIQ,

810
00:30:27.090 --> 00:30:27.590
- into

811
00:30:28.049 --> 00:30:29.109
- a really solid

812
00:30:29.490 --> 00:30:30.390
- data foundation

813
00:30:31.009 --> 00:30:33.570
- that where you got your start? Yeah. Exactly.

814
00:30:33.570 --> 00:30:35.809
- That's a great point. So, really, the start

815
00:30:35.809 --> 00:30:37.890
- of Nasuni way before I joined the company

816
00:30:37.890 --> 00:30:38.869
- was all around

817
00:30:39.274 --> 00:30:42.394
- modernization of traditional network attached storage, right? And

818
00:30:42.394 --> 00:30:44.794
- that's right in the name. Nasuni is all

819
00:30:44.794 --> 00:30:47.355
- about unified NAS, right? And it was all

820
00:30:47.355 --> 00:30:50.794
- around to our founders credit recognizing there was

821
00:30:50.794 --> 00:30:52.734
- this new thing called object storage

822
00:30:53.080 --> 00:30:56.440
- that had unbelievable characteristics from a, obviously, cost

823
00:30:56.440 --> 00:30:59.240
- of ownership perspective, from a scalability perspective, and

824
00:30:59.240 --> 00:31:00.859
- from a durability perspective.

825
00:31:01.240 --> 00:31:03.420
- But there's this thing called the file system.

826
00:31:03.559 --> 00:31:05.775
- And what if we could combine them both?

827
00:31:05.855 --> 00:31:07.474
- Right? And so our

828
00:31:07.934 --> 00:31:08.994
- original architecture

829
00:31:09.375 --> 00:31:11.615
- where the entire file system lives inside of

830
00:31:11.615 --> 00:31:13.535
- the object store, all the metadata, all the

831
00:31:13.535 --> 00:31:15.615
- data was one of the number one innovations

832
00:31:15.615 --> 00:31:17.775
- that is to this day are probably our

833
00:31:17.775 --> 00:31:19.394
- biggest enduring differentiation

834
00:31:19.990 --> 00:31:21.769
- because it allows us to now

835
00:31:22.069 --> 00:31:24.329
- take advantage of again economies of scale,

836
00:31:24.789 --> 00:31:25.289
- scalability

837
00:31:25.750 --> 00:31:28.069
- of object storage, cost of ownership of it,

838
00:31:28.069 --> 00:31:31.509
- but then also to eliminate some traditional things

839
00:31:31.509 --> 00:31:33.429
- that you need along the way like backup

840
00:31:33.429 --> 00:31:36.044
- and disaster recovery built into the platform. And

841
00:31:36.044 --> 00:31:39.244
- so when we recognize, well, we're a platform

842
00:31:39.244 --> 00:31:41.804
- now, companies are entrusting petabytes of their data

843
00:31:41.804 --> 00:31:43.804
- to us, we've earned the right, and by

844
00:31:43.804 --> 00:31:45.964
- the way, they're also demanding that we provide

845
00:31:45.964 --> 00:31:47.724
- more value for the data that lives on

846
00:31:47.724 --> 00:31:50.369
- that platform. That started out with providing more

847
00:31:50.369 --> 00:31:51.589
- advanced cybersecurity

848
00:31:51.970 --> 00:31:54.869
- type capabilities like our ransomware protection service

849
00:31:55.490 --> 00:31:56.630
- that then went into

850
00:31:57.089 --> 00:32:00.130
- thinking about things like, oh, wow, AI activate.

851
00:32:00.130 --> 00:32:01.409
- I wanna be able to tap into that.

852
00:32:01.409 --> 00:32:03.674
- And then DriveIQ is the latest in that.

853
00:32:03.674 --> 00:32:06.075
- And so, yeah, we've laid a foundation and

854
00:32:06.075 --> 00:32:07.755
- the company laid a foundation in the file

855
00:32:07.755 --> 00:32:10.315
- data platform. And now we're later layering these

856
00:32:10.315 --> 00:32:12.794
- these data services, if you will, on top.

857
00:32:12.794 --> 00:32:14.274
- And then earlier in the year, we acquired

858
00:32:14.274 --> 00:32:15.615
- a company called Resilio,

859
00:32:15.980 --> 00:32:18.380
- which is some amazing technology in terms of

860
00:32:18.380 --> 00:32:19.519
- the ability to

861
00:32:19.900 --> 00:32:22.940
- move data at incredibly high speeds. And so

862
00:32:22.940 --> 00:32:24.940
- you talk about today, one of the big

863
00:32:24.940 --> 00:32:27.200
- challenges that we hear is around, again, matching

864
00:32:27.500 --> 00:32:29.019
- the data that I have to the compute

865
00:32:29.019 --> 00:32:30.619
- that I wanna run against it from an

866
00:32:30.619 --> 00:32:32.914
- from an AI perspective. Typically, we are in

867
00:32:32.914 --> 00:32:34.914
- the inference world of things, not in the

868
00:32:34.914 --> 00:32:37.154
- model training side. But again, that compute might

869
00:32:37.154 --> 00:32:39.154
- live out on the edge somewhere, and I

870
00:32:39.154 --> 00:32:40.755
- don't wanna have to create a copy of

871
00:32:40.755 --> 00:32:43.255
- it. And so the the architectural

872
00:32:43.714 --> 00:32:46.359
- fact foundation that we laid with with Nasuni

873
00:32:46.359 --> 00:32:48.680
- and UniFS, etcetera, that's why we say we

874
00:32:48.680 --> 00:32:50.279
- feel like we're made for this moment because

875
00:32:50.279 --> 00:32:52.759
- that's exactly what we're seeing our customers need

876
00:32:52.759 --> 00:32:55.880
- as they're embracing AI and they're they're trying

877
00:32:55.880 --> 00:32:57.720
- to grapple with some of these challenges. We

878
00:32:57.720 --> 00:32:59.674
- say, bring AI to your data, not your

879
00:32:59.674 --> 00:33:01.994
- data to your AI, right? And so being

880
00:33:01.994 --> 00:33:03.994
- able to present that data anywhere in the

881
00:33:03.994 --> 00:33:06.154
- world so that customers get the best of

882
00:33:06.154 --> 00:33:08.174
- both worlds. They get all of the benefits

883
00:33:08.234 --> 00:33:10.795
- of unification consolidation of their data under one

884
00:33:10.795 --> 00:33:11.295
- platform,

885
00:33:11.599 --> 00:33:13.279
- but they also get the benefit of this

886
00:33:13.279 --> 00:33:15.119
- hybrid world where we can make that data

887
00:33:15.119 --> 00:33:17.759
- available to GPUs that are running in a

888
00:33:17.759 --> 00:33:18.980
- colo in Jacksonville

889
00:33:19.519 --> 00:33:21.920
- or maybe their GPUs, but they're in another

890
00:33:21.920 --> 00:33:23.680
- Azure region. So I wanna be able to

891
00:33:23.680 --> 00:33:24.994
- take advantage of it. You don't know how

892
00:33:24.994 --> 00:33:27.394
- powerful a global namespace is until you get

893
00:33:27.394 --> 00:33:30.115
- into one and and start acting with it.

894
00:33:30.115 --> 00:33:32.115
- Yes. There's a couple folks out there who

895
00:33:32.115 --> 00:33:34.115
- do them, and, out of those, there's a

896
00:33:34.115 --> 00:33:35.875
- couple who do them really well. So I'm

897
00:33:35.875 --> 00:33:36.774
- curious if

898
00:33:37.200 --> 00:33:39.220
- based on what we've talked about today, so

899
00:33:39.359 --> 00:33:41.839
- if somebody's out there, they're listening to this,

900
00:33:41.839 --> 00:33:42.980
- they're thinking about

901
00:33:43.359 --> 00:33:46.399
- evaluating for their org, where do you see

902
00:33:46.399 --> 00:33:47.779
- kind of value in

903
00:33:48.079 --> 00:33:49.605
- proof of concept? Like, if I was gonna

904
00:33:49.605 --> 00:33:51.365
- say, like, hey. What are, like, one, two,

905
00:33:51.365 --> 00:33:53.204
- three? Because there's probably not one. Everybody's a

906
00:33:53.204 --> 00:33:55.365
- little bit different. But what's a good proof

907
00:33:55.365 --> 00:33:58.805
- of concept that you go and actually recommend?

908
00:33:58.805 --> 00:34:00.565
- Like, something that a customer could get started

909
00:34:00.565 --> 00:34:03.220
- with today, And it could be bigger, it

910
00:34:03.220 --> 00:34:05.140
- could be smaller, but something that kinda shows

911
00:34:05.140 --> 00:34:07.380
- them some value upfront. Right? You talked about

912
00:34:07.380 --> 00:34:10.099
- tokens costing money, and if they do, right,

913
00:34:10.099 --> 00:34:12.179
- the idle GPUs, they're costing you a bunch

914
00:34:12.179 --> 00:34:13.380
- of money. So if you are in the

915
00:34:13.380 --> 00:34:16.244
- inference flows, rack, things like that, these become

916
00:34:16.545 --> 00:34:19.264
- very important. Yeah. Like anything. Obviously, start with

917
00:34:19.264 --> 00:34:21.585
- the use case. Pick a business impactful use

918
00:34:21.585 --> 00:34:23.905
- case and focus on that. I don't try

919
00:34:23.905 --> 00:34:26.065
- and boil the ocean, but if you get

920
00:34:26.065 --> 00:34:28.804
- an internal champion, whether it's my legal department

921
00:34:28.864 --> 00:34:31.409
- example or whether it's a set of engineers

922
00:34:31.630 --> 00:34:34.510
- working on geospatial data. Pick a targeted use

923
00:34:34.510 --> 00:34:36.590
- case and then go through that kind of

924
00:34:36.590 --> 00:34:39.070
- checklist that I talked about. Like, think about

925
00:34:39.070 --> 00:34:41.550
- a framework of, do we understand the data

926
00:34:41.550 --> 00:34:43.250
- that we need to be able to leverage

927
00:34:43.565 --> 00:34:45.405
- to make this use case successful? Do we

928
00:34:45.405 --> 00:34:49.085
- have security and government's controls around that? Do

929
00:34:49.085 --> 00:34:51.644
- we understand what kinds of AI tooling we're

930
00:34:51.644 --> 00:34:53.565
- gonna want, whether we're gonna be building an

931
00:34:53.565 --> 00:34:56.525
- agentic workflow or whether it's just individuals leveraging

932
00:34:56.525 --> 00:34:58.969
- Copilot against that data? And then be able

933
00:34:58.969 --> 00:34:59.469
- to

934
00:34:59.769 --> 00:35:01.210
- go set up a set of metrics or

935
00:35:01.210 --> 00:35:03.849
- KPIs that represent success. Like, one of the

936
00:35:03.849 --> 00:35:06.989
- hardest conversations right now is proving ROI for

937
00:35:07.130 --> 00:35:09.210
- AI. Right? And so if you don't get

938
00:35:09.210 --> 00:35:10.109
- the other stakeholders

939
00:35:10.409 --> 00:35:11.150
- bought into,

940
00:35:11.555 --> 00:35:13.075
- hey, if we deliver on x, y, and

941
00:35:13.075 --> 00:35:15.635
- z, that's real value. And that could be

942
00:35:15.635 --> 00:35:19.235
- productivity measures, that could be comparing actually one

943
00:35:19.235 --> 00:35:21.155
- nice scenario that we can do with customers

944
00:35:21.155 --> 00:35:22.914
- is you can go say, hey, do you

945
00:35:22.914 --> 00:35:25.015
- run a comparison where you just throw Copilot

946
00:35:25.235 --> 00:35:27.949
- directly at this data store and then compare

947
00:35:27.949 --> 00:35:30.829
- that versus going in through using something like

948
00:35:30.829 --> 00:35:33.150
- AI activating our other solution. And so make

949
00:35:33.150 --> 00:35:35.230
- sure you got the metrics that define success,

950
00:35:35.230 --> 00:35:37.069
- have a real use case, and then have

951
00:35:37.069 --> 00:35:38.750
- that checklist so that you can prove to

952
00:35:38.750 --> 00:35:40.244
- the rest of the business. This is a

953
00:35:40.244 --> 00:35:42.184
- framework that we can follow

954
00:35:42.484 --> 00:35:44.005
- so that if we're able to prove that

955
00:35:44.005 --> 00:35:45.684
- out for this use case, now we can

956
00:35:45.684 --> 00:35:47.924
- wash, rinse, repeat with other use cases in

957
00:35:47.924 --> 00:35:50.324
- the organization through a repeatable pattern. I think

958
00:35:50.324 --> 00:35:52.324
- where we see things sometimes go off the

959
00:35:52.324 --> 00:35:55.000
- rails is just thinking AI is gonna just

960
00:35:55.000 --> 00:35:56.679
- do it for you. Like, hey, if I

961
00:35:56.679 --> 00:35:58.279
- make data available to AI, it's gonna be

962
00:35:58.279 --> 00:36:00.839
- great. Right? That's one. Trying to pick up

963
00:36:00.839 --> 00:36:02.859
- too big a scope is a is another.

964
00:36:03.159 --> 00:36:05.179
- And not having very clear

965
00:36:05.880 --> 00:36:09.215
- defined measures of success before you begin is

966
00:36:09.215 --> 00:36:10.974
- probably the other one that we see with.

967
00:36:10.974 --> 00:36:13.454
- Alright. And if people wanna get started with

968
00:36:13.454 --> 00:36:15.535
- you, Nick, if they wanna go reach out

969
00:36:15.535 --> 00:36:16.195
- to Nasuni,

970
00:36:17.614 --> 00:36:20.619
- work with you, learn more about your current

971
00:36:20.619 --> 00:36:22.619
- products, or it sounds like maybe you have

972
00:36:22.619 --> 00:36:26.460
- some announcements here coming up shortly. Absolutely. How

973
00:36:26.460 --> 00:36:27.920
- are ways people can

974
00:36:28.380 --> 00:36:30.300
- reach out or best way for people to

975
00:36:30.300 --> 00:36:31.820
- reach out to you, get a feel for

976
00:36:31.820 --> 00:36:33.760
- how you can help them or

977
00:36:34.184 --> 00:36:36.744
- future upcoming products or integrations with some of

978
00:36:36.744 --> 00:36:38.744
- these acquisitions? Yeah. We do actually have a

979
00:36:38.744 --> 00:36:41.405
- big external announce coming up on October 6.

980
00:36:41.465 --> 00:36:43.144
- So stay tuned around that. There'll be a

981
00:36:43.144 --> 00:36:45.305
- lot of content coming out. I think we

982
00:36:45.305 --> 00:36:48.445
- can provide probably some links to blog posts

983
00:36:48.664 --> 00:36:50.869
- as well as webinars that we might be

984
00:36:50.869 --> 00:36:53.289
- doing and really start to think about

985
00:36:53.590 --> 00:36:55.130
- what are some of these use cases,

986
00:36:55.430 --> 00:36:58.329
- reach out through those those links. There's always

987
00:36:58.550 --> 00:37:00.710
- on the Nasuni website, of course. You can

988
00:37:00.710 --> 00:37:02.630
- actually go see an AI version of me

989
00:37:02.630 --> 00:37:04.469
- on the website, which, you know, can be

990
00:37:04.469 --> 00:37:05.964
- a little bit interesting. But But it's a

991
00:37:05.964 --> 00:37:07.164
- great way to start to be able to

992
00:37:07.164 --> 00:37:09.505
- interact and ask some questions and then, obviously,

993
00:37:09.644 --> 00:37:12.764
- follow-up, reach out, provide contact info, but there'll

994
00:37:12.764 --> 00:37:14.364
- gonna be a whole bunch of information that

995
00:37:14.364 --> 00:37:15.964
- go into quite a bit more depth on

996
00:37:15.964 --> 00:37:17.644
- things that I've touched on today coming up

997
00:37:17.644 --> 00:37:19.869
- as part of that October 6 events. Perfect.

998
00:37:19.869 --> 00:37:21.389
- Yeah. We'll get a bunch of links, put

999
00:37:21.389 --> 00:37:22.909
- those in the show notes for anybody that

1000
00:37:22.909 --> 00:37:24.909
- wants to go sign up for webinars, learn

1001
00:37:24.909 --> 00:37:25.650
- about those.

1002
00:37:26.429 --> 00:37:28.590
- I know we both have meetings or you

1003
00:37:28.590 --> 00:37:30.429
- and Scott at least have meetings coming up.

1004
00:37:30.429 --> 00:37:32.510
- My Friday is almost done. Yep. My Friday

1005
00:37:32.510 --> 00:37:34.905
- is not done. Else, yeah, anything else you

1006
00:37:34.905 --> 00:37:36.765
- wanna share, Nick? Anything else around

1007
00:37:37.065 --> 00:37:39.945
- Nasuni? Anything else you want people to know

1008
00:37:39.945 --> 00:37:42.344
- about before we wrap up and head off

1009
00:37:42.344 --> 00:37:43.864
- to our next meeting? No. I think the

1010
00:37:43.864 --> 00:37:45.465
- last thing I would say is it's one

1011
00:37:45.465 --> 00:37:47.065
- thing to hear me talk about it. The

1012
00:37:47.065 --> 00:37:49.050
- best thing that we always find is talk

1013
00:37:49.050 --> 00:37:50.410
- talk to one of our customers. So if

1014
00:37:50.410 --> 00:37:51.309
- you are curious

1015
00:37:51.610 --> 00:37:53.369
- and you wanna hear it from people that

1016
00:37:53.369 --> 00:37:55.130
- have actually been living through what I've been

1017
00:37:55.130 --> 00:37:56.650
- talking about, that's one of the best things

1018
00:37:56.650 --> 00:37:57.769
- we can do. We do this all the

1019
00:37:57.769 --> 00:38:00.010
- time where we have mini round tables by

1020
00:38:00.010 --> 00:38:02.385
- vertical, by industry, where we can bring an

1021
00:38:02.385 --> 00:38:04.065
- Asuni customer to talk to a group of

1022
00:38:04.065 --> 00:38:05.585
- people that are just kind of interested. So

1023
00:38:05.585 --> 00:38:07.025
- you have to hear it from the mouth

1024
00:38:07.025 --> 00:38:08.625
- of the people that are actually out there

1025
00:38:08.625 --> 00:38:11.505
- living and breathing this. So highly encouraged if

1026
00:38:11.505 --> 00:38:13.605
- anything that we've talked about today is interesting

1027
00:38:13.905 --> 00:38:15.744
- and you get more information coming out of

1028
00:38:15.744 --> 00:38:17.949
- that October event. We're also going to be

1029
00:38:17.949 --> 00:38:20.590
- doing an event in Chicago in November, which

1030
00:38:20.590 --> 00:38:22.430
- we call in the SUNY verse. So there

1031
00:38:22.430 --> 00:38:24.190
- could be another opportunity to come meet with

1032
00:38:24.190 --> 00:38:26.030
- a bunch of our existing customers. Think of

1033
00:38:26.030 --> 00:38:28.610
- it like we're a mobile user group event.

1034
00:38:28.750 --> 00:38:31.150
- And then also Microsoft Ignite, we will be

1035
00:38:31.150 --> 00:38:33.135
- there. I will be there. So would love

1036
00:38:33.135 --> 00:38:34.735
- to meet people if they wanna come by

1037
00:38:34.735 --> 00:38:36.655
- and find us. But, yeah, I'd say the

1038
00:38:36.655 --> 00:38:38.815
- best thing you can do is actually learn

1039
00:38:38.815 --> 00:38:40.655
- from the people in your industry that are

1040
00:38:40.655 --> 00:38:42.815
- living with the same challenges you have, who've

1041
00:38:42.815 --> 00:38:44.335
- had experience with some of what I've been

1042
00:38:44.335 --> 00:38:46.360
- talking about because they'll give you the honest

1043
00:38:46.360 --> 00:38:49.320
- opinion. Awesome. Well, thanks, Nick. Thanks for coming

1044
00:38:49.320 --> 00:38:51.880
- on and talking to us today, sharing all

1045
00:38:51.880 --> 00:38:54.039
- this information with us, and thanks as well

1046
00:38:54.039 --> 00:38:56.619
- for partnering with us, for sponsoring the podcast

1047
00:38:57.000 --> 00:38:58.760
- and partnering with us in that way. So

1048
00:38:58.760 --> 00:39:01.494
- we really appreciate that as well and hope

1049
00:39:01.494 --> 00:39:03.734
- you go and are able to have a

1050
00:39:03.734 --> 00:39:05.815
- nice weekend. Relax a little bit. Appreciate it.

1051
00:39:05.815 --> 00:39:08.454
- Really good. Guitar. Sorry? Yeah. I've got a

1052
00:39:08.454 --> 00:39:10.215
- I've got a song to record this weekend.

1053
00:39:10.215 --> 00:39:12.695
- So there you go. Alright. Awesome. Well, thanks

1054
00:39:12.695 --> 00:39:15.780
- again. Thanks again, Scott, for as always, and

1055
00:39:16.079 --> 00:39:17.679
- we'll talk to you guys later. Alright. Yep.

1056
00:39:17.679 --> 00:39:18.980
- Have a good one. Thanks, Aaron.

1057
00:39:20.880 --> 00:39:23.199
- If you enjoyed the podcast, go leave us

1058
00:39:23.199 --> 00:39:25.440
- a five star rating in iTunes. It helps

1059
00:39:25.440 --> 00:39:27.164
- to get the word out so more IT

1060
00:39:27.164 --> 00:39:29.244
- pros can learn about Office three sixty five

1061
00:39:29.244 --> 00:39:29.905
- and Azure.

1062
00:39:30.525 --> 00:39:32.125
- If you have any questions you want us

1063
00:39:32.125 --> 00:39:34.285
- to address on the show, or feedback about

1064
00:39:34.285 --> 00:39:36.605
- the show, feel free to reach out via

1065
00:39:36.605 --> 00:39:38.864
- our website, Twitter, or Facebook.

1066
00:39:39.164 --> 00:39:41.005
- Thanks again for listening, and have a great

1067
00:39:41.005 --> 00:39:41.505
- day.