WEBVTT

1
00:00:00.000 --> 00:00:05.920
going. Welcome to your early morning edition. Happy October.

2
00:00:06.640 --> 00:00:10.640
Hey, Jeremy, good to see you. I know it's early for for for

3
00:00:10.640 --> 00:00:16.500
Jenny 7am. Welcome, especially West Coast and those on

4
00:00:16.500 --> 00:00:21.920
mountain time. We appreciate you making the effort. This is very

5
00:00:21.920 --> 00:00:25.460
exciting for us. We've been working at this for a while. And

6
00:00:25.460 --> 00:00:28.800
I know we've got a couple more people still coming in. And then

7
00:00:28.800 --> 00:00:32.420
we're going to just jump right in and, and give you what you

8
00:00:32.420 --> 00:00:37.740
came for, which is to talk about AI and from forge and everything

9
00:00:37.740 --> 00:00:42.960
that we're doing there. So, hey, I'm thrilled to have you, we're

10
00:00:42.960 --> 00:00:48.240
going to jump right in. You may recall at our last in person

11
00:00:48.240 --> 00:00:54.240
when we're all together, we gave you a high level, early peek

12
00:00:54.240 --> 00:00:59.220
into prompt forge as a way to really actively manage prompts

13
00:00:59.220 --> 00:01:03.160
as a first class digital asset in your business. In other

14
00:01:03.160 --> 00:01:07.220
words, so what does that mean in simple terms, it simply means

15
00:01:07.560 --> 00:01:11.220
this. And Glenn will get into the details. But this simply

16
00:01:11.720 --> 00:01:17.540
means that instead of everybody just using chat GPT, or some AI

17
00:01:17.540 --> 00:01:20.920
tool somewhere or another, and you don't capture that

18
00:01:20.920 --> 00:01:24.580
knowledge, they can't even go back and find their own ones. So

19
00:01:24.580 --> 00:01:28.880
there's no way to enhance prompts. And the reality is that

20
00:01:28.880 --> 00:01:32.100
prompts don't make themselves and not all prompts are equal.

21
00:01:32.660 --> 00:01:36.520
Good prompts equals good results and poor prompts is poor results.

22
00:01:36.960 --> 00:01:40.000
And so the question is, how do you make them good? How do you

23
00:01:40.000 --> 00:01:42.820
track them? And then if one guy cracks the code, if one person

24
00:01:42.820 --> 00:01:46.140
in the team cracks the code, how do you spread the good news? How

25
00:01:46.140 --> 00:01:49.460
do you spread that in the organization or in a team? So if

26
00:01:49.460 --> 00:01:53.500
the marketing team crack something, now does everybody

27
00:01:53.500 --> 00:01:56.460
have to go through exactly the same learning curve? And if they

28
00:01:56.460 --> 00:01:59.820
leave, they take it all with them, we just lose all that

29
00:01:59.820 --> 00:02:04.100
intellectual property. So that's what we mean by being a digital

30
00:02:04.100 --> 00:02:09.199
asset, it means that we teach people how to grow and version

31
00:02:09.199 --> 00:02:12.360
their prompts. And when they experiment and suddenly dial it

32
00:02:11.980 --> 00:02:18.380
in. And I gave the example last time, that, you know, I suddenly

33
00:02:18.380 --> 00:02:21.100
realized we're not doing this recording thing. It's my

34
00:02:21.100 --> 00:02:23.680
favorite thing. Oh, it is it is recording good automatically

35
00:02:23.760 --> 00:02:30.480
recording. So and so last time I shared with you that I was, you

36
00:02:30.480 --> 00:02:32.940
know, every time I ran a prompt on some of the stuff Graham and

37
00:02:32.940 --> 00:02:36.240
I working on, he did not like it at all. It you know, because

38
00:02:36.240 --> 00:02:39.260
it's good and brilliant. We're really particular about language

39
00:02:39.260 --> 00:02:42.380
because language is what we're developing in partnership with

40
00:02:42.380 --> 00:02:45.940
the Lord. Because we want to make him known in the way he

41
00:02:46.020 --> 00:02:49.960
wants to be made known. Right. So it's not just how we get to

42
00:02:49.960 --> 00:02:53.940
tell the world what we think it's we want to capture how he

43
00:02:53.940 --> 00:02:58.380
wants us to make him known. So that's our thing. Of course, it

44
00:02:58.380 --> 00:03:02.320
brilliant. So that also means that and it took me almost a

45
00:03:02.320 --> 00:03:08.200
year to refine that prompt. And when I got that prompt refined,

46
00:03:08.300 --> 00:03:11.440
where did it go? I couldn't share it easily. I couldn't

47
00:03:11.440 --> 00:03:15.580
click and share it with with the team. I couldn't easily do

48
00:03:15.580 --> 00:03:18.660
anything with that prompt. So it's there. And so we guess

49
00:03:18.660 --> 00:03:22.320
where did it end up in? I end up making an Apple note, copying it

50
00:03:22.320 --> 00:03:27.040
in there. Now I got a pile of Apple note, prompt things and no

51
00:03:27.040 --> 00:03:30.040
one can see that. And that's probably more disciplined than

52
00:03:30.040 --> 00:03:34.860
most people. Most people like where was that prompt I did

53
00:03:34.880 --> 00:03:37.560
scanning through the history, you know, I'm talking to you,

54
00:03:37.560 --> 00:03:44.260
right. So the thing there was, so number one, how do we capture

55
00:03:44.920 --> 00:03:48.340
quality prompts? Number two, the quality of the prompt is driven

56
00:03:48.440 --> 00:03:53.340
heavily by the context you give it. And so we shared last time

57
00:03:53.760 --> 00:03:58.040
that, that what if you had a context manager and what most

58
00:03:58.040 --> 00:04:01.360
people don't realize, you know, because we didn't know this, we

59
00:04:01.760 --> 00:04:06.700
made 176 page context, we told, we told it everything about our

60
00:04:06.700 --> 00:04:09.580
business and every product and our history and our legal

61
00:04:09.580 --> 00:04:14.080
structure and our team and what our goals were and blah, and what

62
00:04:14.080 --> 00:04:17.140
theology is. And then we found out every time we're sending

63
00:04:17.140 --> 00:04:23.200
that prompt, we're paying for 176 page prompt. Each time, we

64
00:04:23.200 --> 00:04:26.460
had no idea we were being done as a brick, because how did how

65
00:04:26.460 --> 00:04:30.340
would you know this? So it was expensive. And the reality is we

66
00:04:30.340 --> 00:04:33.460
didn't really have to have all that context. And sometimes you

67
00:04:33.460 --> 00:04:36.000
just want, hey, help me write this email, you don't need all

68
00:04:35.780 --> 00:04:41.880
that. And so, so then we found out that not all tasks are the

69
00:04:41.880 --> 00:04:44.520
same, not all contexts are the same, not all prompts are the

70
00:04:44.320 --> 00:04:50.320
same. With that in mind, you know, I'm going to hand it over

71
00:04:50.380 --> 00:04:54.160
to Glenn, to share with us where we're at and what we're doing.

72
00:04:55.180 --> 00:04:59.620
And, and just sorry, just before I do. So welcome to everybody.

73
00:04:59.800 --> 00:04:59.980
So good

74
00:05:00.000 --> 00:05:01.300
to have you on the call.

75
00:05:02.220 --> 00:05:05.900
And today's AI day, unless you didn't catch.

76
00:05:06.840 --> 00:05:08.040
So if you just jumped on,

77
00:05:08.480 --> 00:05:11.500
we're talking through Promforge, what it means.

78
00:05:12.200 --> 00:05:16.020
And so today our business goal is to hand it over to you

79
00:05:16.020 --> 00:05:19.300
for beta testing for October, play with it,

80
00:05:19.420 --> 00:05:21.540
play with your team, get people on,

81
00:05:21.720 --> 00:05:24.060
and then give us feedback for the idea

82
00:05:24.060 --> 00:05:25.240
of launching it in November.

83
00:05:26.120 --> 00:05:28.580
So we're excited about this.

84
00:05:28.580 --> 00:05:32.900
So me personally, I've been trying to build Promforge in

85
00:05:32.920 --> 00:05:34.780
as my actual working memory.

86
00:05:35.080 --> 00:05:36.180
And that's the first thing,

87
00:05:36.300 --> 00:05:38.700
that's the first resistance to change you might find is,

88
00:05:38.960 --> 00:05:41.740
well, I just figured out Claude or ChatGPT now,

89
00:05:41.760 --> 00:05:45.260
and now, and that's fine, they can still use those.

90
00:05:45.640 --> 00:05:47.440
But once they see what this can do,

91
00:05:48.060 --> 00:05:49.840
we're hoping that there's no going back.

92
00:05:49.920 --> 00:05:52.340
And so I've started to use Promforge,

93
00:05:52.520 --> 00:05:55.480
and I'll tell you upfront, my favorite ability

94
00:05:56.120 --> 00:06:00.260
is to say, here's a prompt, so I pick a previous prompt,

95
00:06:00.980 --> 00:06:03.420
I pick the context, maybe I add a few things,

96
00:06:03.600 --> 00:06:05.080
and then I say, compare this,

97
00:06:05.320 --> 00:06:07.820
and then I pick four LLMs next to each other.

98
00:06:08.240 --> 00:06:10.180
Tell me what ChatGPT would do with this,

99
00:06:10.520 --> 00:06:12.660
Claude would do with this, Gemini or whatever,

100
00:06:13.180 --> 00:06:14.900
and you run them all in simultaneous

101
00:06:14.900 --> 00:06:16.220
and you get to see the best one.

102
00:06:17.020 --> 00:06:19.980
So that means you're starting to learn

103
00:06:19.980 --> 00:06:23.240
a whole new level of subtlety, right?

104
00:06:23.920 --> 00:06:26.440
Actually, when it comes to language, Claude is better.

105
00:06:26.820 --> 00:06:29.540
When it comes to anything scientific, mathematical,

106
00:06:30.860 --> 00:06:31.660
oh, Grok's the best.

107
00:06:32.140 --> 00:06:34.660
When it comes to, and so now you're getting

108
00:06:34.660 --> 00:06:37.040
a refinement of intellectual property.

109
00:06:38.360 --> 00:06:40.360
And then maybe the last thing,

110
00:06:40.400 --> 00:06:42.500
which we sent you in that podcast,

111
00:06:42.780 --> 00:06:45.200
did anyone get a chance to listen to that podcast we sent?

112
00:06:46.100 --> 00:06:46.760
Yeah, no?

113
00:06:47.000 --> 00:06:48.500
All right, a few minutes of it.

114
00:06:48.500 --> 00:06:53.520
So the point of that was to say that

115
00:06:53.520 --> 00:06:57.240
once you're honing in your business to that level,

116
00:06:57.600 --> 00:07:01.100
what you find out is it's not just

117
00:07:01.100 --> 00:07:05.140
how do we make individual peoples go a little faster?

118
00:07:05.960 --> 00:07:09.200
What if we can have whole business processes replaced?

119
00:07:09.920 --> 00:07:11.760
But how are we gonna do that

120
00:07:11.760 --> 00:07:13.920
if we have no ability to manage that?

121
00:07:15.080 --> 00:07:16.780
And that's what we're gonna address.

122
00:07:16.780 --> 00:07:19.220
So Glenn, I think I'll tee it up for you.

123
00:07:20.300 --> 00:07:23.080
Fantastic, in fact, I think we've pretty much covered

124
00:07:23.080 --> 00:07:24.520
most of the front end.

125
00:07:25.820 --> 00:07:29.260
So that's awesome, Dion, and again,

126
00:07:29.480 --> 00:07:31.100
super excited to be back with you all

127
00:07:31.100 --> 00:07:33.720
and share the updates and share what we've got.

128
00:07:34.120 --> 00:07:35.500
So let me do this.

129
00:07:35.580 --> 00:07:39.760
I'm just gonna start by sharing some slides.

130
00:07:40.120 --> 00:07:42.220
Is that visible?

131
00:07:43.440 --> 00:07:44.720
Yeah, I can see.

132
00:07:46.180 --> 00:07:47.080
Okay, good.

133
00:07:47.440 --> 00:07:50.120
Is it in present mode or not?

134
00:07:50.940 --> 00:07:51.900
It is, yeah.

135
00:07:52.000 --> 00:07:52.940
Okay, fantastic.

136
00:07:53.220 --> 00:07:57.000
All right, so what I was planning to do today

137
00:07:57.240 --> 00:08:00.340
is to do a recap of the problem we're solving.

138
00:08:00.680 --> 00:08:02.560
Dion did a great job of covering that.

139
00:08:02.900 --> 00:08:04.620
I'm gonna run through a few extra slides

140
00:08:04.620 --> 00:08:05.400
that I put together.

141
00:08:05.620 --> 00:08:06.700
There's been some new research

142
00:08:06.700 --> 00:08:09.180
that's come out in the last month since we've last spoken.

143
00:08:09.320 --> 00:08:11.820
But I think it's really worth sort of highlighting

144
00:08:11.820 --> 00:08:13.240
and it just totally underscores

145
00:08:13.240 --> 00:08:14.540
what we're doing with Promforge.

146
00:08:15.220 --> 00:08:17.140
We're then gonna get into what is Promforge

147
00:08:18.260 --> 00:08:20.040
and then also what it's not.

148
00:08:20.200 --> 00:08:21.980
And then we'll get into the bulk of what we're gonna do,

149
00:08:22.040 --> 00:08:22.760
which is really the demo.

150
00:08:22.960 --> 00:08:26.420
And I want you to do a much more sort of thoughtful

151
00:08:26.880 --> 00:08:29.280
walkthrough of the platform versus what we did last time.

152
00:08:29.300 --> 00:08:32.159
And actually just follow a narrative all the way through

153
00:08:32.320 --> 00:08:34.520
and show the different elements of the platform.

154
00:08:34.860 --> 00:08:36.940
And then we'll talk about onboarding

155
00:08:36.940 --> 00:08:38.400
and end up with an e-Q&A.

156
00:08:38.720 --> 00:08:39.179
Sound good?

157
00:08:40.140 --> 00:08:40.760
Sounds great.

158
00:08:41.320 --> 00:08:45.220
So when we met last time,

159
00:08:45.380 --> 00:08:47.640
we referenced this McKinsey article

160
00:08:47.640 --> 00:08:49.000
that had just come out at then,

161
00:08:49.240 --> 00:08:52.920
which said 80% of companies investing in AI

162
00:08:52.920 --> 00:08:55.040
were not seeing a bottom line return.

163
00:08:55.580 --> 00:08:57.040
And that was completely shocking.

164
00:08:57.880 --> 00:09:00.280
Subsequent to that, MIT have released

165
00:09:00.280 --> 00:09:02.560
a very comprehensive study,

166
00:09:03.740 --> 00:09:07.700
300 companies, $40 billion worth of AI investment.

167
00:09:08.260 --> 00:09:11.620
Their number said 95% failed.

168
00:09:13.740 --> 00:09:16.100
It's, there's a crisis.

169
00:09:15.880 --> 00:09:21.320
There's, companies are not getting the return

170
00:09:21.320 --> 00:09:22.360
on their AI investments.

171
00:09:23.180 --> 00:09:28.360
And then OpenAI last week released a really good paper

172
00:09:28.360 --> 00:09:33.100
on an analysis of chat GPT and all the use of chat GPT.

173
00:09:34.100 --> 00:09:38.500
The shocking thing from that is only 27%

174
00:09:38.500 --> 00:09:43.200
of all chat GPT OpenAI use is businesses.

175
00:09:43.840 --> 00:09:44.600
The rest is personal.

176
00:09:45.340 --> 00:09:47.740
And that number is declining.

177
00:09:48.740 --> 00:09:49.420
It's getting worse.

178
00:09:49.660 --> 00:09:52.840
So for companies, there's a very real problem

179
00:09:53.140 --> 00:09:54.720
that they're dealing with,

180
00:09:54.760 --> 00:09:58.080
which is how they can actually start driving

181
00:09:58.080 --> 00:09:59.980
real bottom line returns from their AI.

182
00:10:00.000 --> 00:10:05.540
AI. So that's sort of the first part of the story. The one paper

183
00:10:05.540 --> 00:10:09.100
put it this way is for enterprise AI, it's essentially

184
00:10:09.100 --> 00:10:12.320
Grammarly on steroids. That's basically what they're doing. So

185
00:10:12.320 --> 00:10:18.320
even that 27% that companies are using it for the usage in

186
00:10:18.320 --> 00:10:22.400
business context is still like it's like polishing a, an email.

187
00:10:23.300 --> 00:10:27.880
That's what most companies are doing. But it's worse than that.

188
00:10:27.880 --> 00:10:32.260
Because as Dion was talking about, that it's really a

189
00:10:32.260 --> 00:10:35.140
throwaway asset. So kind of the people that are using them, it's

190
00:10:35.140 --> 00:10:39.660
just it's a Wild West out there, ungoverned, each person's

191
00:10:39.660 --> 00:10:42.880
writing their own prompts using their own preferred tool. And

192
00:10:42.880 --> 00:10:49.480
there's no, there's no compounding of value from AI

193
00:10:49.480 --> 00:10:51.980
inside of that company. I'm going to double click into that

194
00:10:51.980 --> 00:10:57.860
statement a little bit more in a moment. But essentially, it's a

195
00:10:57.860 --> 00:11:04.540
it's also risk inducing. So for enterprises, when you have these

196
00:11:05.280 --> 00:11:11.400
unmonitored prompts, unmonitored AI systems, there's a very high

197
00:11:11.400 --> 00:11:15.920
risk of IP leakage. What very few people understand is that

198
00:11:15.920 --> 00:11:20.640
most of these AI companies that you're using, actually train

199
00:11:20.720 --> 00:11:27.340
their models off of your data when you prompt. And so the

200
00:11:27.340 --> 00:11:31.300
question becomes, how do you mitigate that risk? And part of

201
00:11:31.300 --> 00:11:34.880
what PromptForge does is it provides a means by which you

202
00:11:34.880 --> 00:11:39.940
can instruct on every one of those prompts, not to train on

203
00:11:39.940 --> 00:11:43.280
that data. Now, that doesn't necessarily mean that those

204
00:11:43.280 --> 00:11:47.140
companies will abide by that. But it does give you legal

205
00:11:47.140 --> 00:11:50.780
recourse, if you find that they have, you've now actually got a

206
00:11:51.620 --> 00:11:54.740
basis, because you've instructed them specifically, but most

207
00:11:54.740 --> 00:11:58.940
companies don't even know that that is a thing. So that's the

208
00:11:59.700 --> 00:12:02.780
situation. And what was really interesting is, as I started

209
00:12:02.780 --> 00:12:05.680
reading these things, it started to come into focus for me that

210
00:12:05.680 --> 00:12:09.740
that there's really two types of, for want of a better word,

211
00:12:10.360 --> 00:12:14.360
networks, AI networks that exist in companies. And this this

212
00:12:14.360 --> 00:12:16.940
harkens back a little bit to my consulting career, I spent 20

213
00:12:16.940 --> 00:12:22.100
years in PwC. And early in my career, we actually built a very

214
00:12:22.100 --> 00:12:26.100
substantial practice around what we called shadow IT, where

215
00:12:26.100 --> 00:12:33.360
we're helping companies actually control IT usage inside of the

216
00:12:33.440 --> 00:12:36.560
organizations that was done on spreadsheets, or whatever the

217
00:12:36.560 --> 00:12:40.000
case may be, uncontrolled by the, by the business, but users

218
00:12:40.000 --> 00:12:42.900
were just using Excel or whatever, and causing all sorts

219
00:12:43.160 --> 00:12:46.760
of problems for the business. It just reminded me that we're in

220
00:12:46.760 --> 00:12:48.860
exactly that same place here right now with AI.

221
00:12:49.580 --> 00:12:53.680
So Glenn, one quick thing is, Ray brings up a good point here.

222
00:12:55.040 --> 00:12:57.300
Ray, did you want to make your own point there about AI

223
00:12:57.080 --> 00:12:58.520
investments? And

224
00:12:59.240 --> 00:13:03.860
no, I just, I just thought the podcast that you sent out was

225
00:13:04.760 --> 00:13:10.080
just really powerful and comparing just irrational

226
00:13:10.320 --> 00:13:13.920
investments without getting a return, very similar to the dot

227
00:13:13.920 --> 00:13:18.020
com boom. And so I just I'd never thought those two things

228
00:13:18.020 --> 00:13:20.580
together, but I just thought that was powerful.

229
00:13:21.000 --> 00:13:22.820
Thank you, Ray. Yeah, and appreciate you calling out down

230
00:13:22.820 --> 00:13:25.620
because I can't see the chat panel while I'm sharing it. So

231
00:13:25.620 --> 00:13:29.640
thank you for calling it out. Yeah, exactly. Right. And, and

232
00:13:29.640 --> 00:13:32.960
so this, this goes back to that time where, where it basically

233
00:13:32.960 --> 00:13:35.520
it's, it's a shadow thing companies, if they're not

234
00:13:35.520 --> 00:13:38.000
actually intentional about this, it's happening across the

235
00:13:38.500 --> 00:13:44.060
organization unseen. 73% of AI usage is in darkness, as we

236
00:13:44.100 --> 00:13:47.900
mentioned, ungoverned security issues, and essentially value

237
00:13:48.080 --> 00:13:53.060
destruction and waste. That's the one whole, very real, but

238
00:13:53.060 --> 00:13:58.020
largely hidden reality, hidden network. But on the other side,

239
00:13:58.420 --> 00:14:03.040
and this, it dawned on me that there's a, there's another

240
00:14:03.360 --> 00:14:07.800
network that also in effect, and that's also somewhat hidden, but

241
00:14:07.800 --> 00:14:10.560
it's the positive side, it's the comp, what I'm calling compound

242
00:14:10.560 --> 00:14:13.440
value network, and that podcast was called strategic value

243
00:14:13.440 --> 00:14:16.240
network. But I think compound values, maybe a little bit of a

244
00:14:16.240 --> 00:14:19.400
better term, still a little consulting, but but the point

245
00:14:19.400 --> 00:14:24.940
being is that there's, if done, if certain things are done, you

246
00:14:24.940 --> 00:14:29.080
will actually get compound value to the business. It's maybe just

247
00:14:29.080 --> 00:14:33.220
worth a comment on this. What triggered this thought was some

248
00:14:33.260 --> 00:14:39.760
research that I that I saw from the early 2000s. It was called

249
00:14:39.760 --> 00:14:48.100
value network analysis. And the hypothesis there was that the

250
00:14:48.240 --> 00:14:54.180
pathways value creation pathways in an enterprise doesn't does

251
00:14:54.180 --> 00:14:58.140
not match the organizational structure for most companies.

252
00:14:58.660 --> 00:14:59.700
And for

253
00:15:00.000 --> 00:15:02.220
they start to actually model and define

254
00:15:02.220 --> 00:15:04.420
what those value creation pathways are.

255
00:15:05.040 --> 00:15:06.760
It allows the leadership of those organizations

256
00:15:06.760 --> 00:15:08.800
to better structure their companies

257
00:15:08.820 --> 00:15:11.460
to optimize for that value creation pathway.

258
00:15:11.560 --> 00:15:14.040
When you take that thing and actually shine a light on it

259
00:15:14.040 --> 00:15:16.400
and understand how value is created in the business,

260
00:15:16.520 --> 00:15:19.060
it totally transforms how you can manage that business.

261
00:15:20.120 --> 00:15:21.500
Same thing here with AI.

262
00:15:22.380 --> 00:15:27.180
If we can shine a light on what this sort of

263
00:15:27.180 --> 00:15:29.140
compound value network is

264
00:15:29.140 --> 00:15:31.860
and allow companies to actually start to manage this

265
00:15:31.860 --> 00:15:33.060
in a disciplined way,

266
00:15:33.500 --> 00:15:36.880
we can drive significant value, I believe, for companies.

267
00:15:37.140 --> 00:15:41.500
This is really highlighted by a tweet,

268
00:15:42.140 --> 00:15:44.280
a ex-post I saw this past week.

269
00:15:44.740 --> 00:15:45.940
And I'll read it because this is,

270
00:15:47.260 --> 00:15:48.640
this absolutely encapsulates the point

271
00:15:48.640 --> 00:15:49.560
I'm trying to make here.

272
00:15:50.200 --> 00:15:51.300
This guy, JB, says,

273
00:15:51.580 --> 00:15:54.600
when I first started layering in AI into my workflows,

274
00:15:54.600 --> 00:15:56.900
I thought of it as time savings.

275
00:15:57.880 --> 00:16:00.620
One hour saved here, two hours saved there.

276
00:16:00.620 --> 00:16:03.940
But over months, I realized it's more than that.

277
00:16:04.300 --> 00:16:05.660
It's compounding.

278
00:16:06.260 --> 00:16:09.360
Shorter feedback loops equals creatives rotating faster.

279
00:16:10.200 --> 00:16:12.860
Faster reporting equals ops catching leaks earlier.

280
00:16:13.800 --> 00:16:17.080
Quicker insights equals tests run in days, not weeks.

281
00:16:17.980 --> 00:16:20.360
None of these things feel like huge needle movers

282
00:16:20.360 --> 00:16:23.120
in isolation, but stacked together,

283
00:16:23.120 --> 00:16:25.780
they change the tempo of the business.

284
00:16:26.720 --> 00:16:29.300
And in growth, tempo is everything.

285
00:16:30.600 --> 00:16:32.680
Brands that can iterate three times faster

286
00:16:32.760 --> 00:16:36.420
don't just save time, they outpace everyone else.

287
00:16:37.120 --> 00:16:39.980
I thought that was a very articulate

288
00:16:43.520 --> 00:16:45.880
way to say what we're trying to say here

289
00:16:45.880 --> 00:16:47.020
about the compound value.

290
00:16:47.280 --> 00:16:49.520
And of course, that triggered the thought

291
00:16:49.520 --> 00:16:51.240
of Elon Musk's famous statement of,

292
00:16:51.240 --> 00:16:53.240
what matters is the pace of innovation.

293
00:16:53.700 --> 00:16:55.760
That is the fundamental determinant of competitiveness.

294
00:16:56.240 --> 00:16:58.340
So all of that to say,

295
00:16:58.740 --> 00:17:02.120
compounding really beats efficiency at the end of the day.

296
00:17:02.540 --> 00:17:05.880
And that at a very sort of conceptual level

297
00:17:06.680 --> 00:17:10.339
is the sort of the backdrop of what we're talking about here.

298
00:17:10.480 --> 00:17:11.960
Are you going through the slides, Glenn?

299
00:17:12.000 --> 00:17:14.660
Because we're still just seeing the Promforge slide.

300
00:17:15.220 --> 00:17:16.920
Oh, you're not seeing any of my slides?

301
00:17:17.240 --> 00:17:17.480
No.

302
00:17:18.240 --> 00:17:18.960
Oh, terrible.

303
00:17:19.040 --> 00:17:20.740
I've been paging through all the slides.

304
00:17:20.740 --> 00:17:23.200
This is a very long introduction.

305
00:17:23.780 --> 00:17:23.780


306
00:17:26.339 --> 00:17:28.060
Okay, so I've got to go back here.

307
00:17:28.580 --> 00:17:32.620
So I'm sorry, we totally didn't see any of my amazing-

308
00:17:32.620 --> 00:17:35.160
No, it was all amazing, but we didn't see any of it.

309
00:17:36.620 --> 00:17:40.780
Just do a quick rewind and just do a quick, this is...

310
00:17:40.780 --> 00:17:41.000
Okay.

311
00:17:41.000 --> 00:17:42.320
These were such great points.

312
00:17:42.400 --> 00:17:44.100
I thought you should have made slides with these.

313
00:17:44.240 --> 00:17:44.840
They're so good.

314
00:17:44.840 --> 00:17:46.680
Oh, yeah, so I apologize.

315
00:17:46.860 --> 00:17:49.480
Well, since we're recording it, we can cut and tap.

316
00:17:50.240 --> 00:17:51.980
So I apologize for repeating myself here,

317
00:17:52.020 --> 00:17:53.240
but let me go through this.

318
00:17:54.440 --> 00:18:00.600
The key thing is 95% of enterprise AI investments

319
00:18:00.960 --> 00:18:02.100
fail completely.

320
00:18:02.580 --> 00:18:04.260
A month ago, when we first spoke,

321
00:18:04.840 --> 00:18:08.220
the McKinsey report had just come out, which said 80%.

322
00:18:08.360 --> 00:18:12.060
In the last few weeks, MIT came out with a report

323
00:18:12.360 --> 00:18:15.340
analyzing 300 large-scale AI investments

324
00:18:15.340 --> 00:18:19.400
totaling $40 billion, 95% failure rate.

325
00:18:21.260 --> 00:18:24.060
And OpenAI, in the last week,

326
00:18:24.240 --> 00:18:27.420
came out with their analysis of the use of ChatGPT,

327
00:18:27.980 --> 00:18:29.280
and that was startling.

328
00:18:30.980 --> 00:18:34.280
27%, only 27% of the use of ChatGPT

329
00:18:34.380 --> 00:18:37.640
is actually in any form of a business context.

330
00:18:37.840 --> 00:18:38.720
The rest of it is personal.

331
00:18:38.960 --> 00:18:41.560
And that 27% is declining.

332
00:18:43.180 --> 00:18:45.700
So this will just says that businesses

333
00:18:45.700 --> 00:18:49.940
have not cracked the code on how to derive value

334
00:18:49.940 --> 00:18:51.440
out of the AI investments,

335
00:18:52.060 --> 00:18:54.040
and that's what Promforge is addressing.

336
00:18:55.560 --> 00:18:58.060
I like this statement in one of the reports.

337
00:18:58.320 --> 00:19:01.300
It basically says, essentially, it's coming down to,

338
00:19:01.580 --> 00:19:04.440
for enterprise AI, it's essentially Grammarly on steroids.

339
00:19:04.600 --> 00:19:06.180
That's basically what they're doing,

340
00:19:06.520 --> 00:19:08.640
is using it to polish an email.

341
00:19:09.640 --> 00:19:12.140
That's kind of sort of what it comes down to.

342
00:19:12.440 --> 00:19:17.380
There's no understanding of the strategic value

343
00:19:17.700 --> 00:19:21.320
that AI prompts and contexts can provide to the business.

344
00:19:22.640 --> 00:19:24.080
And said differently,

345
00:19:24.200 --> 00:19:27.060
there's this sort of shadow AI network that exists.

346
00:19:28.960 --> 00:19:32.560
It harkens back to the days of,

347
00:19:33.180 --> 00:19:35.020
my consulting experience,

348
00:19:35.200 --> 00:19:38.280
we built a whole business around helping companies

349
00:19:38.280 --> 00:19:39.200
with the shadow IT.

350
00:19:39.380 --> 00:19:43.040
People were using spreadsheets to run their businesses

351
00:19:43.040 --> 00:19:45.400
and so on and ending up creating all sorts

352
00:19:45.640 --> 00:19:47.700
of massive problems for enterprise

353
00:19:47.700 --> 00:19:50.200
because it was ungoverned, uncontrolled, unmanaged.

354
00:19:50.460 --> 00:19:53.100
And we're back to exactly the same state here with AI.

355
00:19:53.860 --> 00:19:55.840
We've got a shadow network of everybody

356
00:19:55.840 --> 00:19:59.140
using their favorite tool, ungoverned, uncontrolled.

357
00:20:00.000 --> 00:20:03.000
As Dion mentioned in the opening comments,

358
00:20:03.340 --> 00:20:04.440
it's really throwaway assets.

359
00:20:05.260 --> 00:20:06.300
People kind of write a prompt,

360
00:20:07.140 --> 00:20:08.700
use it, might be good, might not be.

361
00:20:09.720 --> 00:20:11.280
And even if they do hone that over

362
00:20:11.280 --> 00:20:12.680
time and get it better and better and better,

363
00:20:12.820 --> 00:20:14.520
where does it live? How is it shared

364
00:20:14.520 --> 00:20:15.780
with the rest of the organization?

365
00:20:16.220 --> 00:20:19.220
It isn't. And if that person leaves or anything,

366
00:20:20.000 --> 00:20:22.800
that IP is essentially gone from the enterprise.

367
00:20:23.140 --> 00:20:24.440
So it's a throwaway asset.

368
00:20:24.520 --> 00:20:26.900
And then it's also risk-inducing because

369
00:20:26.900 --> 00:20:28.660
very few companies actually appreciate

370
00:20:28.660 --> 00:20:34.820
the risk that they're running with these AI models.

371
00:20:35.460 --> 00:20:37.420
Most of these companies are actually

372
00:20:37.700 --> 00:20:39.680
training off of your prompt data.

373
00:20:40.240 --> 00:20:42.060
So when you submit that prompt with

374
00:20:42.060 --> 00:20:44.340
all your financials and all of that type of stuff,

375
00:20:44.400 --> 00:20:46.260
it's now sitting inside of their servers,

376
00:20:46.520 --> 00:20:47.280
inside of their business,

377
00:20:47.320 --> 00:20:48.760
and they're training off of that data.

378
00:20:50.580 --> 00:20:53.140
Very few companies know that you can actually

379
00:20:53.840 --> 00:20:55.720
instruct those companies to

380
00:20:55.720 --> 00:20:58.040
not train off of your data when you submit the prompt.

381
00:21:00.060 --> 00:21:01.840
PromptForge takes care of that for you.

382
00:21:01.860 --> 00:21:04.080
So that's part of the governance capabilities we add.

383
00:21:05.740 --> 00:21:07.880
So that's some of the issues.

384
00:21:08.120 --> 00:21:14.320
And this brought up the realization that we really have

385
00:21:14.320 --> 00:21:17.700
these two different network

386
00:21:17.700 --> 00:21:19.740
effects taking place inside of organizations.

387
00:21:19.960 --> 00:21:23.520
There's this shadow AI network that we've just mentioned

388
00:21:23.520 --> 00:21:27.500
that is 73% of AI usage is in darkness.

389
00:21:29.340 --> 00:21:30.160
It's ungoverned.

390
00:21:30.620 --> 00:21:32.580
We have all the security and compliance risks,

391
00:21:33.020 --> 00:21:35.600
and it's essentially value destruction and waste.

392
00:21:37.260 --> 00:21:40.680
But there is another network that exists,

393
00:21:40.800 --> 00:21:44.100
and that's what I'm calling the compound value network.

394
00:21:44.440 --> 00:21:49.560
This harkens back to some of my consulting work in years past

395
00:21:49.560 --> 00:21:55.120
where there was this thing called value network analysis,

396
00:21:55.460 --> 00:21:57.860
where what they were doing is they were recognizing

397
00:21:57.860 --> 00:22:00.560
that the value pathways inside of an organization,

398
00:22:01.820 --> 00:22:04.380
how value was created in that organization

399
00:22:04.740 --> 00:22:07.060
didn't actually marry or correlate

400
00:22:07.060 --> 00:22:09.400
with the way the organization was structured.

401
00:22:10.160 --> 00:22:13.200
And so most companies go through reorgs, reorgs, reorgs,

402
00:22:13.300 --> 00:22:16.260
but it's without an actual understanding of how

403
00:22:16.820 --> 00:22:19.040
the value flow works inside of an organization.

404
00:22:19.640 --> 00:22:22.580
And so that created a whole opportunity for consulting

405
00:22:22.580 --> 00:22:24.520
to help companies shine a light

406
00:22:24.520 --> 00:22:28.440
on what was essentially a hidden network

407
00:22:28.440 --> 00:22:29.660
inside of the organization.

408
00:22:29.800 --> 00:22:31.000
And when they could do that,

409
00:22:31.100 --> 00:22:33.080
they could now actually start to manage

410
00:22:33.160 --> 00:22:34.980
that value creation pathways,

411
00:22:35.260 --> 00:22:38.040
organize and optimize the business for that value creation

412
00:22:38.040 --> 00:22:40.140
and thereby drive significant value.

413
00:22:40.560 --> 00:22:42.720
So as I was thinking back to those days,

414
00:22:43.240 --> 00:22:45.780
I realized exactly the same thing happens with AI.

415
00:22:46.840 --> 00:22:48.800
There's these two networks, the shadow network,

416
00:22:49.040 --> 00:22:51.620
but then there's also this compound value network

417
00:22:51.640 --> 00:22:54.800
if harnessed can produce significant return.

418
00:22:55.540 --> 00:22:58.060
And this was really highlighted

419
00:22:58.060 --> 00:23:01.020
and I think perfectly stated by this guy called JB.

420
00:23:01.480 --> 00:23:06.840
This was from last week where I saw this post on X

421
00:23:06.840 --> 00:23:09.960
and he said, when I first started layering AI

422
00:23:09.960 --> 00:23:13.540
into my workflows, I thought of it as time savings.

423
00:23:14.200 --> 00:23:16.720
One hour saved here, two hours saved there,

424
00:23:17.280 --> 00:23:19.940
but over months I realized it's more than that.

425
00:23:19.980 --> 00:23:21.480
It's compounding.

426
00:23:22.220 --> 00:23:25.540
Shorter feedback loops equals creators rotating faster.

427
00:23:26.640 --> 00:23:29.960
Faster reporting equals ops catching leaks earlier.

428
00:23:30.960 --> 00:23:34.100
Quicker insights equals test runs in days, not weeks.

429
00:23:35.000 --> 00:23:38.340
None of these feel like huge needle movers in isolation,

430
00:23:38.340 --> 00:23:42.980
but stacked together, they change the tempo of the business

431
00:23:43.260 --> 00:23:46.260
and in growth, tempo is everything.

432
00:23:47.980 --> 00:23:49.920
Brands that can iterate three times faster

433
00:23:49.920 --> 00:23:53.780
don't just save time, they outpace everyone else.

434
00:23:55.060 --> 00:23:58.880
So this, I think, just absolutely captured this aspect

435
00:23:58.880 --> 00:24:01.900
of really compounding beats efficiency.

436
00:24:02.340 --> 00:24:03.540
Efficiency is sort of a linear,

437
00:24:05.480 --> 00:24:08.940
this aspect of compounding takes the business exponential

438
00:24:09.280 --> 00:24:12.320
and reminds me of the quote from Elon Musk we've all heard,

439
00:24:12.760 --> 00:24:14.720
what matters is the pace of innovation.

440
00:24:15.020 --> 00:24:18.140
That is the fundamental determinant of competitiveness.

441
00:24:19.200 --> 00:24:24.400
So this is the key thing that PrimeForge is targeting.

442
00:24:24.780 --> 00:24:27.640
So it's to kind of bring it from this conceptual level

443
00:24:27.640 --> 00:24:29.560
and start bringing it down to practical things

444
00:24:29.560 --> 00:24:31.900
about what is it that we actually do.

445
00:24:32.900 --> 00:24:36.740
The key thing here is that stewardship of AI

446
00:24:38.700 --> 00:24:42.360
is really around stewardship of these prompts and contexts.

447
00:24:42.680 --> 00:24:45.400
Think of prompts and contexts as the control knobs

448
00:24:46.060 --> 00:24:49.060
of this compound value network.

449
00:24:49.840 --> 00:24:52.900
To the extent we can manage contexts and prompts,

450
00:24:53.080 --> 00:24:57.620
we can manage and influence this compound value network.

451
00:24:58.560 --> 00:24:59.980
So the stewardship in,

452
00:25:00.000 --> 00:25:04.060
drives the alignment, it drives safety and governance,

453
00:25:04.640 --> 00:25:06.860
and ultimately compounding the tempo.

454
00:25:07.380 --> 00:25:09.340
So at the end of the day, the bottom line

455
00:25:09.340 --> 00:25:11.500
is, as Dion had pointed out at the start of the call,

456
00:25:11.880 --> 00:25:15.260
prompts and contexts must become first class digital assets

457
00:25:15.260 --> 00:25:16.060
for the business.

458
00:25:17.680 --> 00:25:21.260
That's the key proposition here.

459
00:25:22.160 --> 00:25:24.300
So in summary, what is PromptForge?

460
00:25:24.600 --> 00:25:25.400
It's exactly that.

461
00:25:25.400 --> 00:25:26.920
It's a platform for organizations

462
00:25:26.920 --> 00:25:29.280
to steward their prompts and contexts

463
00:25:29.280 --> 00:25:32.120
as first class digital assets of the organization.

464
00:25:34.000 --> 00:25:36.080
And if I kind of just sort of jump forward

465
00:25:36.080 --> 00:25:37.780
to what it's not, it's worth just

466
00:25:37.780 --> 00:25:39.640
talking a little bit about this for a second.

467
00:25:40.240 --> 00:25:41.960
What we're not trying to do is to make

468
00:25:42.080 --> 00:25:45.640
this another multimodal video, audio imaging platform,

469
00:25:45.680 --> 00:25:49.620
like if you go to ChatGPT or Grok

470
00:25:49.620 --> 00:25:51.280
or any of these platforms.

471
00:25:53.160 --> 00:25:55.220
The key thing here that we're trying to do

472
00:25:55.220 --> 00:25:58.500
is we're really focusing on helping companies deal

473
00:25:58.500 --> 00:26:00.500
with prompts and contexts.

474
00:26:01.760 --> 00:26:04.700
So based on that earlier research

475
00:26:04.700 --> 00:26:08.940
that I cited from OpenAI, 93% of ChatGPT usage

476
00:26:08.940 --> 00:26:09.960
is still text-based.

477
00:26:10.920 --> 00:26:13.320
So I think this just underscores,

478
00:26:13.740 --> 00:26:19.100
let's not try and sort of compete with the multimodal.

479
00:26:19.160 --> 00:26:22.720
If you need multimodal, use ChatGPT or use Claude

480
00:26:22.720 --> 00:26:23.140
or whatever.

481
00:26:24.600 --> 00:26:29.100
But this is about managing prompts and contexts.

482
00:26:30.600 --> 00:26:32.700
And same thing with some of those rich features

483
00:26:32.700 --> 00:26:34.880
like anthropic artifacts and so on.

484
00:26:35.280 --> 00:26:37.780
By the way, and we'll talk a little bit more

485
00:26:37.780 --> 00:26:38.980
of this when I get into the demo,

486
00:26:39.580 --> 00:26:41.140
we are actually going to include

487
00:26:41.260 --> 00:26:43.040
what's called an MCP integration.

488
00:26:43.260 --> 00:26:46.980
So this is a connector into Claude in ChatGPT.

489
00:26:47.160 --> 00:26:50.420
So if you do need to use Claude or ChatGPT

490
00:26:50.420 --> 00:26:51.860
for some of these other features,

491
00:26:51.860 --> 00:26:54.620
but you still want to have the context

492
00:26:54.620 --> 00:26:57.360
and the prompts that PromptForwards is managing,

493
00:26:58.200 --> 00:26:59.800
there's a way that we can support that.

494
00:27:00.200 --> 00:27:02.880
So we'll talk a bit more about that later.

495
00:27:03.200 --> 00:27:04.940
And the other thing to say is,

496
00:27:05.000 --> 00:27:06.540
we're not trying to make this a consumer app.

497
00:27:06.640 --> 00:27:09.700
We're really focusing on enterprise business users

498
00:27:09.700 --> 00:27:11.700
and likewise, not a mobile app.

499
00:27:13.540 --> 00:27:14.800
Questions, comments on that?

500
00:27:17.320 --> 00:27:18.380
Let's open the floor.

501
00:27:22.620 --> 00:27:23.980
Thoughts, ideas?

502
00:27:26.780 --> 00:27:29.920
Any surprising revelations here for you?

503
00:27:31.940 --> 00:27:32.320
Ben?

504
00:27:34.020 --> 00:27:37.140
I just think you've done a great job

505
00:27:37.340 --> 00:27:43.160
of raising the problems and putting a fine point on it.

506
00:27:43.340 --> 00:27:45.480
And I think that PromptForwards

507
00:27:45.480 --> 00:27:47.480
kind of moves the conversation forward.

508
00:27:48.480 --> 00:27:53.680
I think for us, it's still a question of use cases

509
00:27:53.680 --> 00:27:56.880
and how it applies to our industry.

510
00:27:58.660 --> 00:28:01.080
But I think the idea of,

511
00:28:02.440 --> 00:28:05.620
I mean, understanding the value creation pathway

512
00:28:07.040 --> 00:28:10.020
and the compounding effect is interesting.

513
00:28:10.540 --> 00:28:13.820
So that made me go, hmm.

514
00:28:13.820 --> 00:28:15.420
So I'll think about that for a bit.

515
00:28:16.260 --> 00:28:17.960
Yeah, you raise a great point, Ben.

516
00:28:18.020 --> 00:28:20.860
And that is, how many of us really know,

517
00:28:20.960 --> 00:28:22.260
even in a small business,

518
00:28:23.460 --> 00:28:25.060
if you count all our folks together,

519
00:28:25.160 --> 00:28:30.180
or even if include, we've got whatever, 16 people,

520
00:28:30.660 --> 00:28:33.960
and how's value created at Brilliant, as an example?

521
00:28:34.660 --> 00:28:35.960
Do we actually know?

522
00:28:36.400 --> 00:28:39.760
And would we know how AI would help us compound that?

523
00:28:39.780 --> 00:28:41.160
And the answer is we don't.

524
00:28:41.720 --> 00:28:44.000
I think that it's helpful to know

525
00:28:44.000 --> 00:28:46.360
that almost none of us know how value is created.

526
00:28:47.320 --> 00:28:51.300
I mean, just real talk for a second, right?

527
00:28:51.300 --> 00:28:53.820
And so the bigger the organization gets,

528
00:28:53.960 --> 00:28:58.100
the more obscure that value creation mechanism is.

529
00:28:58.680 --> 00:28:59.840
Because in a small company,

530
00:28:59.920 --> 00:29:03.360
it's often the leader is the value creator, the driver.

531
00:29:03.960 --> 00:29:05.100
We're doing this now.

532
00:29:05.100 --> 00:29:06.100
Now we're running at this.

533
00:29:06.100 --> 00:29:07.080
Then we're doing this.

534
00:29:07.480 --> 00:29:11.540
But that's not a sustainable, scalable plan

535
00:29:11.540 --> 00:29:12.320
in the long term.

536
00:29:12.320 --> 00:29:14.240
So the way a company starts

537
00:29:14.240 --> 00:29:16.040
is not the way it's going to scale.

538
00:29:17.200 --> 00:29:19.020
So what happens is when we were, you know,

539
00:29:19.020 --> 00:29:20.740
five and 10 and 15 people,

540
00:29:20.820 --> 00:29:24.080
these problems all hit when you get past 15 employees,

541
00:29:24.140 --> 00:29:27.720
get to 25, it's a whole new set of problems, right?

542
00:29:28.100 --> 00:29:30.560
So you get to 50 and so on.

543
00:29:30.560 --> 00:29:34.800
So what happens is we lose sight of what creates value.

544
00:29:34.840 --> 00:29:36.720
And especially as we get, you know,

545
00:29:36.720 --> 00:29:38.880
like you start building out an executive team.

546
00:29:39.780 --> 00:29:42.060
Well, your structure doesn't reflect

547
00:29:42.060 --> 00:29:43.540
your value creation mechanism.

548
00:29:43.660 --> 00:29:44.400
So that's the first thing.

549
00:29:44.460 --> 00:29:46.940
I think it's helpful to say we don't know,

550
00:29:47.560 --> 00:29:49.620
or at least we can guess, but we're not sure.

551
00:29:50.280 --> 00:29:53.800
Two is how will the compounding work and AI contribute?

552
00:29:54.640 --> 00:29:55.940
Truthfully, no one knows.

553
00:29:56.820 --> 00:29:58.340
And it's useful to say that too.

554
00:29:58.340 --> 00:29:59.680
And that's why no one's getting a return

555
00:29:59.680 --> 00:29:59.980
because I don't know.

556
00:30:00.000 --> 00:30:05.000
I think we've applied historical management mechanisms

557
00:30:05.600 --> 00:30:07.620
to try to manage AI and it doesn't work.

558
00:30:08.300 --> 00:30:11.900
So it's helpful to stop doing that, right?

559
00:30:11.920 --> 00:30:16.220
It's like, well, we're introducing streaming video,

560
00:30:16.920 --> 00:30:20.800
but we got this blockbuster store with lots of VHS tapes.

561
00:30:20.900 --> 00:30:23.620
So maybe we should ship them faster, you know?

562
00:30:23.620 --> 00:30:26.960
So shipping faster is not gonna help a streaming business,

563
00:30:27.080 --> 00:30:29.000
you know, and it feels like that's what we're doing.

564
00:30:29.000 --> 00:30:32.020
I think we're coming up with methods,

565
00:30:32.680 --> 00:30:34.480
management methods for an old modality.

566
00:30:35.240 --> 00:30:36.960
No one's looking for the VHS tapes

567
00:30:36.960 --> 00:30:38.740
and no one's coming into the store anymore.

568
00:30:39.760 --> 00:30:41.980
And so we go, oh, we'll put up a spreadsheet

569
00:30:41.980 --> 00:30:44.440
and then they can pick their VHS tapes directly.

570
00:30:44.680 --> 00:30:46.120
You know, they have to come into the store.

571
00:30:46.260 --> 00:30:48.080
That's still, that's how it feels.

572
00:30:48.240 --> 00:30:50.660
We're in that stage, which is by the way,

573
00:30:50.660 --> 00:30:52.980
the most normal thing when a brand new

574
00:30:53.280 --> 00:30:54.500
disruptive technology enters.

575
00:30:54.720 --> 00:30:56.080
Think of when email arrived.

576
00:30:56.800 --> 00:30:59.380
We didn't know what to do with email, right?

577
00:30:59.820 --> 00:31:01.120
Well, why don't you just send a letter?

578
00:31:01.480 --> 00:31:02.480
It's more formal anyway.

579
00:31:02.820 --> 00:31:04.100
Of course, the lawyers jumped on it.

580
00:31:04.100 --> 00:31:05.080
It doesn't count if it's email.

581
00:31:05.180 --> 00:31:07.360
Today an email counts in a court of law.

582
00:31:08.080 --> 00:31:11.940
So the issue comes to, if we don't know,

583
00:31:12.040 --> 00:31:13.800
then the third question is, well,

584
00:31:13.800 --> 00:31:17.040
what pathway can we take to discover those two things?

585
00:31:17.200 --> 00:31:20.220
And that's what we were trying to build into Promforge

586
00:31:20.220 --> 00:31:23.940
so that while we don't know, we can start to learn.

587
00:31:23.940 --> 00:31:28.040
Just like that great quote you read, Glenn, from JB.

588
00:31:28.520 --> 00:31:30.860
Is that, that's not the JB, is it?

589
00:31:31.380 --> 00:31:32.060
X testing?

590
00:31:32.060 --> 00:31:32.940
I don't know.

591
00:31:33.420 --> 00:31:34.980
I went and looked, it's not the same JB,

592
00:31:35.080 --> 00:31:38.380
although he has the XAI little logo in the top.

593
00:31:38.780 --> 00:31:40.660
So I went and looked to see, hey, is he actually,

594
00:31:41.180 --> 00:31:44.520
but he's not the same JB Straubel that you're referring to,

595
00:31:44.540 --> 00:31:46.120
that I don't think it is.

596
00:31:46.660 --> 00:31:49.540
So the point of that was only that

597
00:31:51.280 --> 00:31:54.360
when the people who are supposed to know, don't know,

598
00:31:54.940 --> 00:31:56.360
when the experts don't know,

599
00:31:56.720 --> 00:31:59.520
how the heck are we supposed to know, right?

600
00:31:59.680 --> 00:32:02.020
So the good news is no one knows.

601
00:32:03.000 --> 00:32:05.940
And the data's proof, all the data Glenn showed us,

602
00:32:05.940 --> 00:32:09.400
nobody really knows, but the value is there.

603
00:32:09.600 --> 00:32:14.300
The question is, what's the massive strategic advantage

604
00:32:14.300 --> 00:32:18.800
to us as a group of kingdom enterprises here,

605
00:32:18.800 --> 00:32:20.620
no matter how big or small we are,

606
00:32:20.880 --> 00:32:23.220
to leverage this and to figure this out,

607
00:32:23.220 --> 00:32:25.720
because we have one massive advantage.

608
00:32:25.860 --> 00:32:28.440
We have the genius Holy Spirit who knows everything.

609
00:32:29.000 --> 00:32:29.680
That's exactly right.

610
00:32:29.680 --> 00:32:31.140
So what we've been trying to do is saying,

611
00:32:31.160 --> 00:32:33.980
Lord, wouldn't it be amazing for once

612
00:32:34.760 --> 00:32:37.140
that something comes from the kingdom?

613
00:32:38.480 --> 00:32:39.440
We're not always looking,

614
00:32:39.620 --> 00:32:41.920
it's not always the kingdom looking to the world,

615
00:32:41.940 --> 00:32:43.220
like what should we catch up on?

616
00:32:43.260 --> 00:32:44.160
What's the latest thing?

617
00:32:44.720 --> 00:32:47.700
By the way, JB marked himself as a vibe marketer.

618
00:32:47.700 --> 00:32:50.880
I noticed that in that thing, yeah.

619
00:32:51.180 --> 00:32:53.640
So, all right, well, any other thoughts, questions?

620
00:32:54.020 --> 00:32:56.300
Because I think you're probably keen to jump into the demo,

621
00:32:56.440 --> 00:32:58.540
but any other, I just want to take a minute

622
00:32:58.540 --> 00:33:00.300
and just breathe, it's a lot to absorb.

623
00:33:01.400 --> 00:33:04.820
Man, I'd just say, keep going, man.

624
00:33:04.820 --> 00:33:07.480
I love that you're engaged in the conversation

625
00:33:07.480 --> 00:33:10.520
and that you're in this place,

626
00:33:10.780 --> 00:33:12.580
because I think it's brilliant

627
00:33:12.580 --> 00:33:14.400
and I think the Lord loves it.

628
00:33:14.420 --> 00:33:14.760
Yeah.

629
00:33:14.780 --> 00:33:15.620
Thank you, Ben.

630
00:33:15.620 --> 00:33:17.500
I feel that so strongly.

631
00:33:17.900 --> 00:33:19.000
Dion and I keep talking about,

632
00:33:19.440 --> 00:33:20.100
we really feel,

633
00:33:21.820 --> 00:33:26.280
and we're so excited about what this is, what this can be.

634
00:33:26.480 --> 00:33:27.600
And we really think it's something

635
00:33:27.600 --> 00:33:30.320
that the Lord has given it into our hand to do.

636
00:33:30.520 --> 00:33:31.880
And so we're excited about it.

637
00:33:32.180 --> 00:33:33.840
Can't wait to see the impact.

638
00:33:33.980 --> 00:33:34.560
And I think there's,

639
00:33:35.340 --> 00:33:36.860
I think to the earlier conversation,

640
00:33:36.900 --> 00:33:37.860
I think there's a lot of learning.

641
00:33:38.060 --> 00:33:40.040
We'll talk a little bit more after the demo,

642
00:33:40.220 --> 00:33:42.800
which is around the adoption side of it.

643
00:33:42.920 --> 00:33:45.340
Ben, you were speaking about use cases, right?

644
00:33:45.340 --> 00:33:45.900
Okay.

645
00:33:46.140 --> 00:33:48.760
This is great and it's a great tool and all this,

646
00:33:48.980 --> 00:33:50.660
but at the end of the day, it has to connect.

647
00:33:50.920 --> 00:33:53.320
It has to land inside of the organization.

648
00:33:53.740 --> 00:33:56.340
So we'll talk more about that at the back end of this.

649
00:33:57.040 --> 00:33:58.880
We've no shortage of ideas,

650
00:33:59.600 --> 00:34:01.820
but we want to hold back a little bit

651
00:34:01.820 --> 00:34:04.460
and really execute based on feedback we're getting

652
00:34:04.460 --> 00:34:06.120
as opposed to, hey, we think this is a great idea.

653
00:34:06.240 --> 00:34:06.880
Let's go do this.

654
00:34:07.020 --> 00:34:07.860
You know, so.

655
00:34:09.760 --> 00:34:12.280
All right, shall we dive into-

656
00:34:12.280 --> 00:34:12.280


657
00:34:12.280 --> 00:34:13.440
Let's go and see the good stuff.

658
00:34:13.440 --> 00:34:16.060
Yeah, let's go and see,

659
00:34:16.300 --> 00:34:17.659
make sure I'm in the right one here

660
00:34:17.659 --> 00:34:19.800
and hopefully you can see that.

661
00:34:20.260 --> 00:34:20.260


662
00:34:21.440 --> 00:34:24.020
Now I'm going to go through and tell me,

663
00:34:24.159 --> 00:34:25.600
make sure that my screen is updating

664
00:34:25.600 --> 00:34:26.460
because I don't want to end up

665
00:34:26.460 --> 00:34:27.820
with a situation where it's not.

666
00:34:28.159 --> 00:34:30.980
So what you can see at the moment to welcome to Promforge.

667
00:34:31.820 --> 00:34:32.080
Yeah.

668
00:34:32.080 --> 00:34:34.239
Just before you jump into that, Glenn,

669
00:34:35.120 --> 00:34:37.520
guys, just feel free to unmute yourself and talk.

670
00:34:37.620 --> 00:34:39.340
And I talk, you know, just chat.

671
00:34:39.340 --> 00:34:41.320
I mean, you're a small group here and so just talk.

672
00:34:41.320 --> 00:34:42.560
Don't, don't, just,

673
00:34:42.699 --> 00:34:44.940
what we really want is your engagement.

674
00:34:45.199 --> 00:34:46.980
So if something comes up or terms used,

675
00:34:47.000 --> 00:34:49.139
or you want to know more about something,

676
00:34:49.659 --> 00:34:52.100
just interrupt Glenn as he goes through it.

677
00:34:52.300 --> 00:34:53.060
Please do, yeah.

678
00:34:53.280 --> 00:34:55.159
And I think it goes without saying,

679
00:34:55.280 --> 00:34:58.480
but part of what Dion and I are really needing

680
00:34:58.480 --> 00:34:59.980
the feedback on is, okay, what-

681
00:35:00.000 --> 00:35:02.340
resonates, what doesn't, what's clear, what's not,

682
00:35:02.720 --> 00:35:04.400
you know, as we start to sort of dial in

683
00:35:04.400 --> 00:35:05.680
the story and the messaging.

684
00:35:06.500 --> 00:35:08.240
So really value the feedback.

685
00:35:08.800 --> 00:35:09.160
All right.

686
00:35:10.580 --> 00:35:12.640
Can see what I see?

687
00:35:13.780 --> 00:35:16.000
The thing has gone away, right?

688
00:35:16.380 --> 00:35:16.940
Yep.

689
00:35:17.200 --> 00:35:18.040
Okay, cool.

690
00:35:18.380 --> 00:35:21.400
So when you first log in, brand new,

691
00:35:21.920 --> 00:35:23.980
you've just registered, got in,

692
00:35:24.060 --> 00:35:25.580
this is the first screen you're gonna see.

693
00:35:26.040 --> 00:35:27.520
This is called the launch pad.

694
00:35:27.580 --> 00:35:29.560
This is really where all the action happens.

695
00:35:30.360 --> 00:35:32.340
We've tried to make this just the place

696
00:35:32.340 --> 00:35:33.900
that you focus on in your work.

697
00:35:34.440 --> 00:35:37.060
And to that end, we've got things like

698
00:35:37.060 --> 00:35:38.680
you can actually just hide everything.

699
00:35:39.180 --> 00:35:43.820
So it just becomes one workspace for working.

700
00:35:44.400 --> 00:35:46.360
So that's what these little things do here

701
00:35:46.360 --> 00:35:48.940
is hiding and opening up the space.

702
00:35:49.540 --> 00:35:50.940
When you first get in,

703
00:35:51.280 --> 00:35:52.920
you'll probably wanna come up over here.

704
00:35:53.020 --> 00:35:55.780
You wanna come and just set your profile up,

705
00:35:55.860 --> 00:35:57.440
you know, upload an image.

706
00:35:59.560 --> 00:36:04.140
You know, and basically do something like that,

707
00:36:04.140 --> 00:36:05.500
set up your name.

708
00:36:10.580 --> 00:36:10.780
Oops.

709
00:36:11.660 --> 00:36:12.240
Here we go.

710
00:36:13.380 --> 00:36:14.740
Can't even type straight here.

711
00:36:14.760 --> 00:36:15.080
Okay.

712
00:36:15.100 --> 00:36:16.540
And so on and so forth.

713
00:36:16.600 --> 00:36:20.300
So that's just getting you oriented, set up and so on.

714
00:36:20.420 --> 00:36:22.920
Now, there's a couple of other things worth knowing.

715
00:36:23.300 --> 00:36:24.960
Just a quick touch on them.

716
00:36:25.060 --> 00:36:27.160
You can set whether you like material,

717
00:36:27.220 --> 00:36:28.640
dark mode, light mode.

718
00:36:28.640 --> 00:36:31.140
By the way, what shows easier on Zoom?

719
00:36:31.380 --> 00:36:32.220
Is it dark or light?

720
00:36:33.160 --> 00:36:33.860
I think light.

721
00:36:34.060 --> 00:36:34.600
I think light.

722
00:36:35.300 --> 00:36:36.100
Yeah, okay.

723
00:36:36.540 --> 00:36:37.480
Thought so.

724
00:36:38.280 --> 00:36:40.520
But whatever you want, you can set that up.

725
00:36:40.960 --> 00:36:41.980
We'll talk more about this.

726
00:36:42.240 --> 00:36:43.220
There's a notification bell.

727
00:36:43.320 --> 00:36:46.500
So we have a means of being able to send messages to users.

728
00:36:48.260 --> 00:36:50.080
And at the top is your,

729
00:36:50.580 --> 00:36:54.180
what your company is and what tier is that you're on

730
00:36:55.760 --> 00:36:57.180
with the platform.

731
00:36:57.700 --> 00:37:00.120
Let me just hide this so we don't keep getting this.

732
00:37:00.260 --> 00:37:00.480
All right.

733
00:37:00.560 --> 00:37:05.740
So that's just a really quick orientation of what you see.

734
00:37:05.820 --> 00:37:08.420
The key thing is the actual chat panel over here.

735
00:37:08.480 --> 00:37:11.480
This is just like ChatGPT or Claude,

736
00:37:11.920 --> 00:37:14.540
where you're gonna enter your chats.

737
00:37:14.900 --> 00:37:18.480
However, there's some really key significant differences.

738
00:37:18.740 --> 00:37:22.080
Number one is you can see on a constant transparency

739
00:37:22.080 --> 00:37:25.360
exactly how many credits you've got at any point in time.

740
00:37:25.360 --> 00:37:27.240
We'll see this more as we go.

741
00:37:28.960 --> 00:37:29.900
Prompts are right there.

742
00:37:30.120 --> 00:37:34.720
So back to the story of being able to have managed prompts.

743
00:37:34.840 --> 00:37:36.600
We'll see how we manage these in a little bit.

744
00:37:36.720 --> 00:37:37.620
When you first start,

745
00:37:38.420 --> 00:37:42.160
there's two prompts that are preloaded into every workspace.

746
00:37:42.600 --> 00:37:44.360
One is just a welcome to PromptForge

747
00:37:44.360 --> 00:37:47.820
and another is a meta prompt,

748
00:37:48.300 --> 00:37:51.040
which is a prompt to create prompts, which is very cool.

749
00:37:51.200 --> 00:37:52.720
We'll get to play with this a little bit later.

750
00:37:52.720 --> 00:37:55.580
But if you're looking to get a high quality prompt,

751
00:37:56.300 --> 00:37:57.280
you can use this prompt

752
00:37:57.420 --> 00:37:59.380
to actually create your high quality prompt.

753
00:38:01.760 --> 00:38:03.500
But if you just wanna select one,

754
00:38:03.680 --> 00:38:05.880
you can select the prompt that you want.

755
00:38:06.880 --> 00:38:10.240
And the contexts are which contexts are turned on

756
00:38:10.440 --> 00:38:11.860
for this particular one.

757
00:38:11.900 --> 00:38:13.560
We'll play a lot more with this in a moment.

758
00:38:13.760 --> 00:38:17.100
So this is where you can see what is going to be turned on,

759
00:38:17.740 --> 00:38:20.100
which particular model you're using,

760
00:38:20.140 --> 00:38:21.600
and we'll come back to compare later.

761
00:38:22.240 --> 00:38:23.480
So now if we hit go,

762
00:38:24.480 --> 00:38:26.180
what you can do is you can see

763
00:38:26.420 --> 00:38:29.100
it's starting to stream the response.

764
00:38:30.480 --> 00:38:32.100
And Dion, you'll notice now

765
00:38:32.100 --> 00:38:33.520
you can scroll back while it's streaming.

766
00:38:33.760 --> 00:38:35.800
So that was one of the things we spoke about.

767
00:38:36.580 --> 00:38:39.020
You'll notice that it's all formatted

768
00:38:39.340 --> 00:38:42.040
according to what's called markdown formatting.

769
00:38:43.860 --> 00:38:46.920
And so it formats all the responses beautifully

770
00:38:47.100 --> 00:38:48.280
and there you go.

771
00:38:48.460 --> 00:38:49.800
What comes out at the bottom

772
00:38:49.800 --> 00:38:52.480
is it tells you exactly how many tokens it took,

773
00:38:53.200 --> 00:38:54.880
exactly what the cost of that was,

774
00:38:54.960 --> 00:38:57.020
and which model was used for it.

775
00:38:57.140 --> 00:39:00.360
So back to from a governance and a control point of view,

776
00:39:00.740 --> 00:39:04.340
you know precisely at every single step

777
00:39:04.380 --> 00:39:06.640
exactly what it's costing you.

778
00:39:07.920 --> 00:39:09.640
And also, even if you look,

779
00:39:09.880 --> 00:39:10.340
I was gonna go back.

780
00:39:10.960 --> 00:39:12.360
Glenn, if we'd had that at Brilliant,

781
00:39:12.520 --> 00:39:13.700
we would have realized that

782
00:39:13.900 --> 00:39:17.360
that massive thing we kept uploading as our base.

783
00:39:18.080 --> 00:39:20.260
Look, we have 176 page prompt.

784
00:39:21.640 --> 00:39:23.380
Why did I just pay that much money?

785
00:39:23.780 --> 00:39:27.020
We would have immediately known something was wrong.

786
00:39:27.120 --> 00:39:27.760
Exactly, exactly.

787
00:39:27.820 --> 00:39:28.340
Now look at this,

788
00:39:28.480 --> 00:39:30.360
like even on the prompt that goes out,

789
00:39:30.440 --> 00:39:31.720
if you just look and see there,

790
00:39:31.720 --> 00:39:33.700
you can just hover over the thing.

791
00:39:33.760 --> 00:39:37.980
It will tell you the user was only 117 tokens.

792
00:39:38.160 --> 00:39:40.000
That was the prompt that I gave it.

793
00:39:40.000 --> 00:39:44.020
The context that it attached to it was 1800 tokens.

794
00:39:44.420 --> 00:39:45.560
That's still very small.

795
00:39:45.560 --> 00:39:47.420
Yeah, on that one that we were doing

796
00:39:47.640 --> 00:39:50.120
with that massive 170 something pages,

797
00:39:50.320 --> 00:39:52.700
I don't know how many hundreds of thousands of tokens

798
00:39:52.700 --> 00:39:55.240
it was using every single time we submitted it.

799
00:39:55.360 --> 00:39:56.840
Yeah, but we're so smart though.

800
00:39:57.420 --> 00:39:57.820
Look it up.

801
00:39:58.440 --> 00:39:59.160
That was brilliant.

802
00:40:00.000 --> 00:40:01.840
Just very expensive.

803
00:40:03.060 --> 00:40:05.500
Part of the $87 billion learning.

804
00:40:07.000 --> 00:40:08.200
That's right.

805
00:40:08.380 --> 00:40:10.620
Yeah. Other than that, it was a great idea. Yeah.

806
00:40:10.800 --> 00:40:14.700
That's right. So anyway, so the point being is like we're giving

807
00:40:15.080 --> 00:40:19.200
max, max transparency, literally to the fractions of a penny,

808
00:40:19.560 --> 00:40:24.140
this, this, this total one was, was point eight of a cent was

809
00:40:24.140 --> 00:40:29.040
the cost of this particular transaction. So complete,

810
00:40:29.040 --> 00:40:34.500
complete visibility. So that's, that's the first the first thing

811
00:40:34.500 --> 00:40:39.040
that we will see now. I mentioned the context. So let's

812
00:40:39.040 --> 00:40:41.360
go and try something just to show you how these contexts

813
00:40:41.360 --> 00:40:45.780
work. What happens if I turn off markdown? Okay, I'll go back,

814
00:40:45.820 --> 00:40:49.020
I'll start a new, a new conversation. By the way, I was

815
00:40:49.020 --> 00:40:51.640
confused because when I ran it on the same conversation, it

816
00:40:51.640 --> 00:40:54.640
still showed markdown. I'm like, hang on, I turned off markdown

817
00:40:54.640 --> 00:40:58.100
until I realized is that the conversation maintains all the

818
00:40:58.100 --> 00:41:02.140
past history of that conversation thread. So of course, it had

819
00:41:02.140 --> 00:41:06.080
all the context of markdown as part of its history. So that's

820
00:41:07.180 --> 00:41:12.060
why it's just a little, little thing I discovered. But if we

821
00:41:12.060 --> 00:41:14.920
start a new one here, and again, we just say, let's just say,

822
00:41:15.340 --> 00:41:20.280
welcome to PromptForge. We run this one now. Now, what will

823
00:41:20.280 --> 00:41:25.940
happen is, you see, it's got, I'll actually still do some.

824
00:41:26.540 --> 00:41:31.440
Yeah, but it's pretty much no, no markdown in that one. So it

825
00:41:31.440 --> 00:41:37.200
just shows you how you can turn on and off the context. Right,

826
00:41:37.200 --> 00:41:43.100
so that's super, super simple. Let's, let's go to start playing

827
00:41:43.100 --> 00:41:48.580
with and up our game a little bit with prompts. So we saw

828
00:41:48.580 --> 00:41:51.780
that there were just these two prompts that came preloaded

829
00:41:51.780 --> 00:41:57.100
with the system when we start. So now we can go to the prompt

830
00:41:59.260 --> 00:42:04.180
template page here, and actually start to set up all prompts.

831
00:42:04.240 --> 00:42:07.420
And this is, this is the first sort of key place where you can

832
00:42:07.420 --> 00:42:10.640
start to steward and manage the prompts for your organizations

833
00:42:10.640 --> 00:42:15.560
in this place over here. Just from an orientation point of

834
00:42:15.560 --> 00:42:19.740
view, it shows you how many prompts in this particular plan,

835
00:42:19.740 --> 00:42:23.340
the pro plan, it's got 10, you can have up to 10 prompts in

836
00:42:23.340 --> 00:42:30.220
the, in the, in the pro plus, it's, it's infinite. This is 10.

837
00:42:32.020 --> 00:42:36.300
Dionne, here's your list view that you had spoke about. What

838
00:42:36.300 --> 00:42:39.320
we realized is the card view looks really cool and all when

839
00:42:39.320 --> 00:42:42.520
you've got a handful of cards, but when you start getting to a

840
00:42:42.520 --> 00:42:45.820
dozen and many, many more, it becomes very, very difficult to

841
00:42:45.820 --> 00:42:49.120
manage. So we created a list view, and you can order by most

842
00:42:49.120 --> 00:42:53.260
used last updated alphabetical date added. So you can filter

843
00:42:53.260 --> 00:42:58.720
on your, your prompts a whole lot better that way. Every one

844
00:42:58.720 --> 00:43:02.680
of these pages is a help page that will pop up and give you a

845
00:43:02.680 --> 00:43:05.500
whole lot of information about what to do on this particular

846
00:43:05.720 --> 00:43:10.520
page. So your help is there. Now I want to show the

847
00:43:10.520 --> 00:43:13.560
marketplace. So you don't have to start by creating your own

848
00:43:13.560 --> 00:43:16.600
prompts. You can, and by the way, if you are going to write

849
00:43:16.600 --> 00:43:21.040
your own prompts, I really encourage you to try this meta

850
00:43:21.040 --> 00:43:24.520
prompt. We'll play with it in a little bit, but it will help

851
00:43:24.860 --> 00:43:30.480
you craft a high quality prompt that matches the clear format,

852
00:43:30.560 --> 00:43:34.140
the standard that we put together, it quality scores the

853
00:43:34.140 --> 00:43:36.760
prompt, but make sure that the, the prompt that you're creating

854
00:43:36.760 --> 00:43:38.980
is going to be a high class prompt for your organization.

855
00:43:39.940 --> 00:43:43.940
But that said, if we go browse the marketplace, so we've

856
00:43:43.940 --> 00:43:48.680
actually, we've seeded the platform with a marketplace of

857
00:43:48.680 --> 00:43:56.080
almost 1000 prompts of pretty much, I mean, it's the gambit

858
00:43:56.080 --> 00:44:04.840
this is, I think, 72 categories of, of, of prompts in that's

859
00:44:04.840 --> 00:44:08.980
broken out. And we basically use the sort of Amazon book. You

860
00:44:08.980 --> 00:44:11.640
know, if you go to Amazon, you ordering books, the category

861
00:44:11.640 --> 00:44:14.340
categorization system that Amazon users thought would be a

862
00:44:14.340 --> 00:44:17.840
good way to do it. And so we followed the same sort of

863
00:44:17.840 --> 00:44:21.460
structure. And then we've actually loaded preloaded the

864
00:44:21.460 --> 00:44:27.000
platform with 1000 prompts that all match the quality standards

865
00:44:27.000 --> 00:44:31.400
that we've, that we've said. So that's a great place to start.

866
00:44:31.540 --> 00:44:34.840
So let's, this is instead of creating one, let's just go,

867
00:44:35.000 --> 00:44:39.060
let's say I want an NDA. So here's one here, bulletproof NDA

868
00:44:39.060 --> 00:44:41.880
creation guide. And I'm going to install that into my

869
00:44:42.500 --> 00:44:49.800
workspace. So now I've got this, this, this prompt. And now we

870
00:44:49.800 --> 00:44:52.320
can use it. Now you can launch it from in here if you want, or

871
00:44:52.320 --> 00:44:54.560
you can launch it from the launchpad. I'll just launch it

872
00:44:54.560 --> 00:44:56.940
from here. But what it's going to do is take you to the

873
00:44:57.560 --> 00:44:59.480
launchpad. And it's going to run that prompt.

874
00:45:00.000 --> 00:45:07.660
what's really cool to see is that the prompt has the concept of variables. So

875
00:45:07.660 --> 00:45:12.440
when we actually create these prompts we actually put in what the

876
00:45:12.440 --> 00:45:15.580
what the variables are that this prompt needs. So in this particular case let's

877
00:45:15.580 --> 00:45:23.600
just say CTO and it's going to be a startup and it's going to be a

878
00:45:24.240 --> 00:45:29.300
industries tech just something like that right. Oh and I'll actually before I do

879
00:45:29.300 --> 00:45:32.760
that I want to just turn my context back on global markdown formatting that's

880
00:45:32.760 --> 00:45:42.060
what I want. Go back to my prompts over here and go do this one. Let's do that

881
00:45:42.160 --> 00:45:42.820
again.

882
00:45:48.200 --> 00:45:55.020
You got it covered global markdown formatting real quick again. Okay I missed

883
00:45:55.020 --> 00:46:01.060
the markdown the understanding of markdown. It's not it's basically all it

884
00:46:01.060 --> 00:46:04.420
is it's just a way to make a pretty output so I'll show you I'll show it in

885
00:46:04.460 --> 00:46:10.680
a moment. Well while Ben's typing Ben I'll just all that is is it's like you know

886
00:46:10.680 --> 00:46:15.680
when you've got a imagine a pilot text in a Word document right all this is

887
00:46:15.680 --> 00:46:20.020
just adding headings and color. Okay yeah that's all markdown is it's just saying

888
00:46:20.020 --> 00:46:24.320
hey this is heading one this is heading two this is heading three. It just makes it

889
00:46:24.340 --> 00:46:31.140
pretty and makes it more readable right and you can do things marked out that

890
00:46:31.180 --> 00:46:36.380
exposes that everything is an AI now. That's right so now we're telling the

891
00:46:36.400 --> 00:46:41.380
model when you reply you need to format your response according to these rules.

892
00:46:42.380 --> 00:46:49.760
Got it. And you can but we've created a markdown context you can clone that and

893
00:46:49.760 --> 00:46:53.360
modify it if you want and we'll see it in a moment. So you can you say you know

894
00:46:53.360 --> 00:46:57.540
what you don't really like how it's doing it when you get there that's sort

895
00:46:57.540 --> 00:47:00.860
of under the more advanced things just simply clone it and modify it to do what

896
00:47:00.860 --> 00:47:03.940
you want but the point being you're starting to see how you can start to

897
00:47:04.500 --> 00:47:08.080
control your interactions with AI and standardize them across the enterprise.

898
00:47:10.180 --> 00:47:14.700
So here's an example you can see it's nicely marked down it's got blocks of

899
00:47:14.700 --> 00:47:23.000
stuff etc etc callouts nicely marked the thing and that's a this one's using

900
00:47:23.360 --> 00:47:30.020
Claude Haiku. All right now you'll notice that it's got no signatures execution

901
00:47:30.020 --> 00:47:36.460
it's got no knowledge of who I am so we need to we need to actually go and

902
00:47:36.460 --> 00:47:41.460
create a context because and you've got the markdown formatting and you know in

903
00:47:41.720 --> 00:47:52.560
the basic platform guide here so I'm gonna just quickly copy a thing here we

904
00:47:52.560 --> 00:47:59.280
go back give me a second here there we go copy this okay. We just created a generic

905
00:47:59.280 --> 00:48:05.140
NDA for some CTO of some business owner or whatever generic yeah it didn't know

906
00:48:05.140 --> 00:48:09.520
the context so now Glenn you're creating that for us. So yeah now I'm gonna say

907
00:48:09.520 --> 00:48:14.400
okay I need a Glenn Hobbs profile okay

908
00:48:18.340 --> 00:48:31.960
so this one is all about if I can type all about Glenn. Now what I'm gonna do is

909
00:48:31.960 --> 00:48:37.660
I'm gonna paste I just copied and pasted from a document that I've got so this is

910
00:48:37.660 --> 00:48:43.000
one that we've done it brilliant so I'm just using my brilliant my brilliant one

911
00:48:43.000 --> 00:48:47.200
which what's my role what are my responsibilities I could have just put

912
00:48:47.200 --> 00:48:51.860
my contact info and it would have been fine for the NDA but I wanted to point

913
00:48:51.860 --> 00:48:56.320
out that we can start putting in a whole lot information about who I am as a

914
00:48:56.320 --> 00:49:01.700
person what I love what I don't love is context that the AI will now know who

915
00:49:01.700 --> 00:49:06.440
much more about me and so when it's replying or responding to prompts it now

916
00:49:06.440 --> 00:49:13.640
can respond with that as much as the context to optimize for me. I can also

917
00:49:14.020 --> 00:49:18.240
specify the scope what this is just to say okay by default it'll go to the

918
00:49:18.240 --> 00:49:22.740
organization so everybody in the organization will see this but there are

919
00:49:22.940 --> 00:49:29.900
many contexts that would be personal you don't you if you you want to just keep

920
00:49:29.900 --> 00:49:34.920
it for yourself and not expose that to everybody you can put that in in tag it

921
00:49:34.920 --> 00:49:41.580
personal now it will not show up in anybody else's context. Likewise you can

922
00:49:41.580 --> 00:49:44.800
scope it to the team so if you have a finance team for example there might be

923
00:49:45.720 --> 00:49:49.740
certain contexts about the finance operations that you want the finance

924
00:49:49.740 --> 00:49:54.560
team to know but you don't want to expose it to the whole organization so

925
00:49:54.560 --> 00:49:59.980
you can scope it to the team. Global is is

926
00:50:00.000 --> 00:50:03.240
what those other ones were, which will apply to the platform as a

927
00:50:03.240 --> 00:50:06.440
whole. You can't, if without the right permissions, you can't

928
00:50:06.440 --> 00:50:09.600
assign global. If you did, it would apply to everybody,

929
00:50:11.280 --> 00:50:15.460
everybody's workspaces. And so you need super admin rights.

930
00:50:16.000 --> 00:50:19.340
In other words, that will go to all Promforge users.

931
00:50:19.740 --> 00:50:23.860
That'll go to all companies, all employees everywhere. That's

932
00:50:23.860 --> 00:50:24.600
why it's called.

933
00:50:25.320 --> 00:50:27.780
Which reminds me, Dion, I have to double check and make sure

934
00:50:27.780 --> 00:50:32.540
that globally scope to super user only. But the point being

935
00:50:32.540 --> 00:50:35.560
is, that's what that is. And then proprietary is an

936
00:50:35.560 --> 00:50:40.060
interesting idea, which is this, there are, it's essentially a

937
00:50:40.060 --> 00:50:44.280
modification of global. But what it does is it makes the context

938
00:50:44.280 --> 00:50:48.200
available, but doesn't give the details of what's in it. And the

939
00:50:48.200 --> 00:50:52.240
reason for that is it's like IP protection. So if you give, let

940
00:50:52.240 --> 00:50:58.140
me, like, an example would be LeaderForge, InnerCircle, all

941
00:50:58.140 --> 00:51:03.080
the training materials that Dion has got, let's say, as part of

942
00:51:03.680 --> 00:51:08.120
InnerCircle, those need to, you want to have that as a context

943
00:51:08.120 --> 00:51:14.440
that you can use inside of your workspace. But the IP itself is

944
00:51:14.440 --> 00:51:16.840
very sensitive, you don't want people cutting and pasting all

945
00:51:16.840 --> 00:51:21.760
that IP, and then reusing. So proprietary will stop this from

946
00:51:21.760 --> 00:51:26.880
being cloned. And it would hide the content of that particular

947
00:51:26.880 --> 00:51:30.560
prompt, but still make it available. So that's a little

948
00:51:30.560 --> 00:51:33.120
more of a future thing. We haven't spent too much time on

949
00:51:33.120 --> 00:51:35.760
that one. But that's the idea of what proprietary will do.

950
00:51:36.240 --> 00:51:42.920
Yeah, I mean, to me that kind of is the heart of this whole idea

951
00:51:42.920 --> 00:51:47.640
in some regard is that, you know, man, I see this thing in

952
00:51:47.640 --> 00:51:54.540
many ways as SOPs on steroids, right? With, you know, and the

953
00:51:54.540 --> 00:52:00.320
ability to actually engage with the knowledge and context

954
00:52:00.320 --> 00:52:08.600
specific results in that proprietary context, you know,

955
00:52:08.600 --> 00:52:18.540
allows you to, it's like, you know, what would a general

956
00:52:18.540 --> 00:52:21.800
contractor think about this? Or how would they answer it? I mean,

957
00:52:21.800 --> 00:52:25.740
that opens a whole lot of possibilities. Love it.

958
00:52:26.900 --> 00:52:30.760
That's right. So yeah, it's exactly right. So okay, let's

959
00:52:30.760 --> 00:52:36.780
save this one. So now we've added a new Ben Hobbs profile.

960
00:52:37.540 --> 00:52:41.640
Let's go back to Launchpad. Now here's this last one that we ran

961
00:52:41.720 --> 00:52:47.140
that didn't have any context of who I was. Ben Hobbs profile is

962
00:52:47.140 --> 00:52:59.860
turned on. Now say, now it should be able to pick up that

963
00:52:59.860 --> 00:53:03.880
and there you go contextual approach NDA. What's it doing

964
00:53:03.880 --> 00:53:14.940
now? There you go. So it started to insert all the contact

965
00:53:15.160 --> 00:53:25.260
information now. But now I noticed that it's adding all

966
00:53:25.260 --> 00:53:27.880
sorts of things about kingdom and things into that. So that's

967
00:53:27.880 --> 00:53:33.160
a good example of why maybe you put your contact block only for

968
00:53:33.160 --> 00:53:35.840
contact information and put other contexts, you know, so

969
00:53:35.840 --> 00:53:39.680
this is a great example of, you don't want your NDA filled with

970
00:53:39.680 --> 00:53:48.700
your personal opinions about the Lord in your NDA. So what we do

971
00:53:48.700 --> 00:53:53.220
is we go back and update that one with just the contact

972
00:53:53.320 --> 00:53:59.020
information. And then redo it again, you know, I won't do it,

973
00:53:59.020 --> 00:54:02.660
you get the idea. You can start to see how it can start picking

974
00:54:02.660 --> 00:54:07.020
up the context and you dial the context in and this is why it's

975
00:54:07.020 --> 00:54:10.780
so important to be able to turn off different elements of the

976
00:54:10.180 --> 00:54:14.780
context. So for example, if we are doing a legal document, I

977
00:54:14.780 --> 00:54:20.020
might want my contact to contact details, but not my overall

978
00:54:20.900 --> 00:54:27.980
profile. So now I can start to break my context app into blocks

979
00:54:27.980 --> 00:54:32.640
of managed blocks and turn on and off which ones should be

980
00:54:33.720 --> 00:54:35.140
making sense so far.

981
00:54:36.520 --> 00:54:40.820
You can see very quickly guys that that what happens very

982
00:54:40.820 --> 00:54:43.620
quickly when you start just managing prompts and contexts

983
00:54:43.900 --> 00:54:48.620
that your your thinking changes, you know, what happens is you're

984
00:54:48.620 --> 00:54:51.480
thinking about your thinking changes, metacognition changes,

985
00:54:51.480 --> 00:54:56.040
because you're now thinking about right there, do I have to

986
00:54:56.960 --> 00:54:59.980
have a legal or more business process?

987
00:55:00.000 --> 00:55:06.300
profile versus my general profile versus my personal spiritual passions versus my own

988
00:55:06.300 --> 00:55:07.060
personal theology.

989
00:55:07.280 --> 00:55:12.620
I've suddenly come up with four contexts and all those personal, well, yeah, they personal,

990
00:55:13.080 --> 00:55:15.720
but maybe we have an SOP for our team.

991
00:55:16.120 --> 00:55:19.040
That's the context that no one has to keep typing, right?

992
00:55:19.040 --> 00:55:21.160
So we don't have to type when we send out an invoice.

993
00:55:21.340 --> 00:55:25.680
We do this, we do this, create the invoice, but here's the context.

994
00:55:25.920 --> 00:55:28.600
So now we can set it up.

995
00:55:28.600 --> 00:55:32.900
Well, Glenn's going to show us more refinements here in a second, because we're just trying

996
00:55:32.900 --> 00:55:37.200
to introduce you to the basic idea of you've got launchpad, you've got prompts, you've

997
00:55:37.200 --> 00:55:37.860
got context.

998
00:55:38.740 --> 00:55:41.580
And I've been using this, by the way, for the last month.

999
00:55:42.860 --> 00:55:47.540
And I can take a second and show some of my context and prompts and things I've done.

1000
00:55:47.620 --> 00:55:54.540
But for instance, I'm running a new elective for our school right now on church governance.

1001
00:55:55.360 --> 00:55:59.160
And, you know, for me, it's four generations of learning there.

1002
00:55:59.900 --> 00:56:06.600
So the contexts are, here is everything, you know, the 52 books my spiritual dad wrote.

1003
00:56:07.020 --> 00:56:11.900
And here is an entire previous course on each of these contexts.

1004
00:56:12.260 --> 00:56:16.720
And then I run them and I go, no, hang on, that's way too much David stuff and not enough

1005
00:56:16.720 --> 00:56:18.920
my stuff, or it's too much Graham stuff.

1006
00:56:19.140 --> 00:56:24.140
And so just like the NDA got all, you know, personal and started quoting scripture in

1007
00:56:24.140 --> 00:56:30.860
the middle of the NDA, you know, so you learn very, very quickly how to leverage this.

1008
00:56:30.880 --> 00:56:32.700
But the good news is you only do it once.

1009
00:56:33.120 --> 00:56:36.420
And you can do it for our team, or you can do it for a whole organization.

1010
00:56:37.200 --> 00:56:37.680
That's right.

1011
00:56:38.480 --> 00:56:40.760
So now I've just rerun it.

1012
00:56:40.760 --> 00:56:44.620
But you see, I didn't say use my contact info.

1013
00:56:44.820 --> 00:56:53.860
So now I'm going to say, I'm going to have to pick up the pace here a little bit, we

1014
00:56:53.860 --> 00:56:56.820
could, but anyway, you get the idea.

1015
00:56:57.580 --> 00:56:59.760
So there you go, prepared by Glenn Hobbs.

1016
00:57:00.420 --> 00:57:03.900
Now you can see prepared by now you can see all the primary contact.

1017
00:57:04.160 --> 00:57:07.520
So it's just populated everything into the document, all good to go.

1018
00:57:09.200 --> 00:57:12.520
All right, next thing I want to share and talk about is then the models.

1019
00:57:14.100 --> 00:57:21.640
So in the platform are north of 350 models that we have available.

1020
00:57:22.180 --> 00:57:26.820
Now, depending on which tier you go with, if you go with the free tier, we only include

1021
00:57:26.820 --> 00:57:35.100
three, and we careful to only select the models that are free or almost free, because we obviously

1022
00:57:35.100 --> 00:57:40.620
pay still a few fractions of a penny, you know, processing charges, and we don't want

1023
00:57:40.920 --> 00:57:44.720
to, you know, start giving away giving away the farm.

1024
00:57:44.720 --> 00:57:52.440
But when you go to pro, it's, I don't know, 15 models, whatever we've we've set.

1025
00:57:52.620 --> 00:57:56.340
And then you go to pro plus, it's all of them, all 350 models.

1026
00:57:57.180 --> 00:58:03.200
So in this particular case, you see, I selected Haiku, that's why it was constantly just using

1027
00:58:03.200 --> 00:58:04.220
that particular model.

1028
00:58:05.800 --> 00:58:09.880
Now, I've only enabled four, we'll go and see how we can turn on a whole lot more.

1029
00:58:10.820 --> 00:58:14.120
What you can do is you can start the ones that you like.

1030
00:58:14.120 --> 00:58:17.400
So if I like and say, you know what, I like that one.

1031
00:58:17.860 --> 00:58:23.520
And I like this one, you can actually start to filter because particularly when you start

1032
00:58:23.520 --> 00:58:28.480
getting to 300 models in the list, just, it's much easier just to say, I just want to see

1033
00:58:28.480 --> 00:58:29.900
my favorites, you know.

1034
00:58:30.720 --> 00:58:31.780
So that's helpful.

1035
00:58:31.880 --> 00:58:34.680
But here's the thing I want to highlight is we have an auto mode.

1036
00:58:36.580 --> 00:58:44.460
Auto mode basically will pre-select the model for you, which is going to be the optimum

1037
00:58:44.460 --> 00:58:44.820
model.

1038
00:58:45.840 --> 00:58:51.740
The way that it works, and we'll go back and we'll look at this in a second, is for every

1039
00:58:51.740 --> 00:58:57.160
one of those prompts that we put in, we can actually assign what the recommended model

1040
00:58:57.160 --> 00:58:58.640
is for that particular prompt.

1041
00:58:59.160 --> 00:59:04.660
So if you rerun that prompt, and we'll do this in a second, auto will say, oh, okay,

1042
00:59:04.660 --> 00:59:09.580
this prompt says you should use GPT 4.0 or whatever we've assigned, and it will use that

1043
00:59:09.580 --> 00:59:09.800
model.

1044
00:59:10.300 --> 00:59:15.120
So this is going back to a comment Dion made earlier on about optimizing, each model is

1045
00:59:15.120 --> 00:59:19.620
going to behave differently, depending on what you're needing it to do.

1046
00:59:19.820 --> 00:59:27.920
And so the ability to be able to attach the model with the prompt is a huge value unlock

1047
00:59:27.920 --> 00:59:28.700
for organizations.

1048
00:59:29.180 --> 00:59:30.600
So we'll show that in a second.

1049
00:59:30.720 --> 00:59:31.880
So let's just do that.

1050
00:59:32.640 --> 00:59:35.980
I'm going to turn auto mode on, but what I'm going to do is I'm going to go back to our

1051
00:59:35.980 --> 00:59:41.120
prompt templates here, go look at this Bulletproof NDA, and you'll see there's no

1052
00:59:41.220 --> 00:59:42.520
recommended model set.

1053
00:59:42.580 --> 00:59:54.320
So I'm going to edit this and go and set a, let's just see, let's go OpenAI GPT 4.0, I

1054
00:59:54.320 --> 00:59:57.980
think that does a fairly decent job, and now we're good to go.

1055
00:59:58.100 --> 00:59:59.520
Now, if I go back to this.

1056
01:00:00.000 --> 01:00:05.200
we're going to do the same same thing again and say okay we're going to do

1057
01:00:06.700 --> 01:00:10.500
same bulletproof thing over here let's just say

1058
01:00:13.620 --> 01:00:16.180
do the same thing start up

1059
01:00:18.720 --> 01:00:29.640
and tick and now we go and say launch now you see it's auto so now it's going to pick

1060
01:00:29.640 --> 01:00:35.320
a a model we'll see what it picked in a moment

1061
01:00:42.240 --> 01:00:47.140
and glenn it's fair to say that it's also not just optimizing for which is the best one but

1062
01:00:47.140 --> 01:00:52.420
which is the best one at the cheapest price that is correct so what will happen is the the order

1063
01:00:52.420 --> 01:01:00.020
if you have specified a recommended model for this prompt and that model is available in the

1064
01:01:00.020 --> 01:01:05.120
plan that you add because it might not be but if it is it will use the one that was specified if

1065
01:01:05.220 --> 01:01:11.600
not it would then auto select based on what which will produce the lowest the best response with the

1066
01:01:11.600 --> 01:01:15.220
lowest cost so there's an algorithm that sits behind to do that so if you're not sure which

1067
01:01:15.220 --> 01:01:20.580
models which just set it to auto and let it be done now in this particular case it actually went

1068
01:01:21.540 --> 01:01:29.780
so that's interesting so i'll need to check on that one maybe we didn't have the open ai one

1069
01:01:29.780 --> 01:01:38.180
turned on it's quite possible but it should override the the recommended what's that

1070
01:01:38.180 --> 01:01:43.900
doesn't the auto override the recommended or does auto no the the order precedent should be it will

1071
01:01:43.900 --> 01:01:51.600
use the recommended one if it is available in the plan that you've selected so if you know because

1072
01:01:51.600 --> 01:01:56.440
you wanted to abide by it if you say use this model and that model is available it should use

1073
01:01:56.440 --> 01:02:00.840
that model so i'll go and check and see if we didn't have the one that i picked was probably

1074
01:02:00.840 --> 01:02:08.740
not turned on but let me move on because we've got less than half an hour left here i think so

1075
01:02:08.740 --> 01:02:15.860
let's uh let's let's move forward quickly um well actually that's a good segue into the next sort of

1076
01:02:16.140 --> 01:02:21.660
thing um i'll talk a little bit about the admin functionality that sits behind this the admin

1077
01:02:21.660 --> 01:02:26.560
functionality the settings and tools that you see here is only visible to the person who is the admin

1078
01:02:26.560 --> 01:02:32.260
for your organization users don't see that they only see this part over here the dashboards the

1079
01:02:32.980 --> 01:02:40.360
prompts the context that's all they see um but if you're an admin you will now be able to go and

1080
01:02:40.360 --> 01:02:48.000
see those ai models so here's in in in this particular plan there are actually 28 models

1081
01:02:48.000 --> 01:02:53.780
that i've turned on so at a super user level i can specify which models available at which plan

1082
01:02:53.780 --> 01:03:00.280
on the pro plan there's 28 so you can see all of these models that are turned on and

1083
01:03:00.280 --> 01:03:05.560
um you can see you know which we can go and turn on different ones so i probably didn't

1084
01:03:05.560 --> 01:03:08.940
have the right one turned on here for which one i picked i don't know which one i picked

1085
01:03:09.020 --> 01:03:18.080
but you get the idea so here's how as an admin you can choose which models you want to have

1086
01:03:18.080 --> 01:03:22.920
available inside of your organization and this is important because they are models from other

1087
01:03:23.100 --> 01:03:28.720
countries you might say you know what we only want to have us-based models so we're going to filter

1088
01:03:28.720 --> 01:03:36.640
filter out and only make available us-based models as an example any questions on that

1089
01:03:39.940 --> 01:03:48.440
are you implying that chinese models might steal ip because i'd be shocked well um the truth is

1090
01:03:48.440 --> 01:03:54.940
i wouldn't put it past some of the u.s companies either the difference being right the difference

1091
01:03:54.940 --> 01:04:01.880
being is if you which we will which we provide as an enterprise capability the ability to instruct

1092
01:04:01.880 --> 01:04:06.380
models not to train off of their data you have legal recourse if you find that they have because

1093
01:04:06.380 --> 01:04:13.380
you've specifically given them instructions yeah yeah so but i would i would still i didn't mean

1094
01:04:13.380 --> 01:04:17.800
sorry i didn't mean to distract you no but i would still but it's a good point john i think

1095
01:04:18.480 --> 01:04:22.180
it's really important for people to understand that's what's happening behind the scenes people

1096
01:04:22.180 --> 01:04:26.180
are hungry for data to train on so they're using everything they can get their hands on

1097
01:04:26.300 --> 01:04:31.940
a quick couple of things i'll not waste too much time on this but under users and teams here's

1098
01:04:31.940 --> 01:04:37.360
where as an admin again you can go and set up teams and create a team so if you want to create

1099
01:04:37.360 --> 01:04:42.640
a team for finance or marketing or whatever you can create your teams and then you can go and add

1100
01:04:42.640 --> 01:04:49.180
members to the teams invite members and if there's there's no teams to assign but if there were you

1101
01:04:49.180 --> 01:04:56.460
could say invite this member invite them as a member or as an org admin or team admin so you

1102
01:04:56.460 --> 01:04:59.980
can have multiple org admins if you want and

1103
01:05:00.000 --> 01:05:03.160
then assign them to a team and send the invite.

1104
01:05:03.340 --> 01:05:06.380
And then what that does is just allows them seamlessly on board,

1105
01:05:06.880 --> 01:05:09.960
no issues, and they're off to the races.

1106
01:05:10.880 --> 01:05:11.640
So that's the team.

1107
01:05:11.780 --> 01:05:14.660
And then on the billing and usage, this is where, again, as an admin,

1108
01:05:15.200 --> 01:05:20.020
you have the ability to manage payment methods, et cetera, et cetera.

1109
01:05:20.960 --> 01:05:27.240
When what we've done for the beta team folk is that as soon as you register,

1110
01:05:27.800 --> 01:05:30.840
we will preload your account with credits.

1111
01:05:32.660 --> 01:05:38.640
So we didn't want to have to worry about having to subscribe and then

1112
01:05:38.640 --> 01:05:39.540
get credits and all that.

1113
01:05:40.420 --> 01:05:43.100
So that's what this will handle.

1114
01:05:43.480 --> 01:05:47.700
Now, I wanted to call an attention to something interesting here.

1115
01:05:48.020 --> 01:05:51.800
Again, you can see literally every single transaction, that's a penny.

1116
01:05:53.420 --> 01:05:55.860
For a moment, just think about what we've been doing.

1117
01:05:55.860 --> 01:06:01.180
And some of these fairly complicated, creating an NDA,

1118
01:06:01.480 --> 01:06:05.540
this one was the most expensive, was $0.10, which is very pricey, actually.

1119
01:06:06.500 --> 01:06:10.940
But think for a second, if you eat this pricey in AI terms,

1120
01:06:11.180 --> 01:06:14.740
think for a second, yeah, think about it like this.

1121
01:06:14.820 --> 01:06:16.940
What is your chat GPT subscription?

1122
01:06:17.220 --> 01:06:17.520
$20?

1123
01:06:18.740 --> 01:06:19.220
OK.

1124
01:06:20.020 --> 01:06:24.560
How many of those things would you need to run to make up

1125
01:06:24.560 --> 01:06:26.020
your $20 subscription?

1126
01:06:27.800 --> 01:06:31.560
So it's 20 divided by 0.1.

1127
01:06:31.620 --> 01:06:38.020
It's like 200 or more things that you would have to do.

1128
01:06:38.380 --> 01:06:44.220
So what it turns out is that if you look at the cost model with chat GPT

1129
01:06:44.220 --> 01:06:48.680
or these others, you're either significantly overpaying,

1130
01:06:49.320 --> 01:06:51.600
massively overpaying for what you're getting,

1131
01:06:52.220 --> 01:06:55.860
or if you happen to have a power user that's driving the thing,

1132
01:06:56.040 --> 01:06:58.880
you're going to go over that $20 and you're going to pay a fortune

1133
01:06:58.880 --> 01:06:59.820
in overage charges.

1134
01:07:01.140 --> 01:07:04.040
So what we've done in the platform is we don't do that at all.

1135
01:07:04.040 --> 01:07:05.480
We just charge what it is.

1136
01:07:06.140 --> 01:07:11.260
And you get a number of credits and you just draw down on those credits.

1137
01:07:11.320 --> 01:07:12.840
If you don't use, you don't lose.

1138
01:07:13.440 --> 01:07:14.880
You just keep those credits.

1139
01:07:18.340 --> 01:07:19.220
On the...

1140
01:07:19.220 --> 01:07:23.460
Just to be clear, Glenn, you're saying that the credits,

1141
01:07:26.420 --> 01:07:31.040
it's not like every year, every month it resets.

1142
01:07:31.480 --> 01:07:33.480
This one, you just keep using them.

1143
01:07:33.540 --> 01:07:35.500
So if you use them over three months or three days,

1144
01:07:35.500 --> 01:07:36.580
it doesn't make any difference.

1145
01:07:36.760 --> 01:07:37.720
It makes no difference.

1146
01:07:37.840 --> 01:07:41.120
And what you would want to do is just set this auto top up over here,

1147
01:07:41.120 --> 01:07:43.760
which we'll just take care of making sure that it auto tops up

1148
01:07:43.760 --> 01:07:44.600
and doesn't run out.

1149
01:07:45.840 --> 01:07:49.220
But the crazy thing is, even in all the testing, Dion, we've been doing,

1150
01:07:50.800 --> 01:07:56.000
I don't think we've got past 20 bucks, 15 bucks off the month.

1151
01:07:56.400 --> 01:08:01.980
It's just, it's not a lot, you know, but now the point being here is that

1152
01:08:01.980 --> 01:08:06.480
we don't want that monthly thing of credits expiring.

1153
01:08:07.180 --> 01:08:07.940
They don't.

1154
01:08:08.260 --> 01:08:11.680
There's a monthly subscription fee, but the credits are...

1155
01:08:11.680 --> 01:08:15.100
When you've got credits, they live as long as you need them to be.

1156
01:08:15.380 --> 01:08:15.720
So anyway.

1157
01:08:16.979 --> 01:08:19.100
All right, let's move on.

1158
01:08:19.920 --> 01:08:22.300
Quickly, I want to get to a few more things here.

1159
01:08:22.859 --> 01:08:26.000
Dashboards, so again, back to full visibility.

1160
01:08:26.660 --> 01:08:29.680
We've got personal dashboards, so you can track how many conversations

1161
01:08:29.680 --> 01:08:32.580
you've had in the last 30 days, tokens, monthly usage,

1162
01:08:32.580 --> 01:08:34.979
and what the model is the most commonly used.

1163
01:08:35.020 --> 01:08:39.160
You can see the ones that we've just been running right now.

1164
01:08:39.160 --> 01:08:42.740
Sonnet, Haiku, Haiku, these are the ones we've been doing.

1165
01:08:44.620 --> 01:08:46.500
It's full transparency at the organization level,

1166
01:08:46.779 --> 01:08:49.640
similar type of thing where you can see it across the organization.

1167
01:08:50.840 --> 01:08:54.880
One of the things we are wanting to do here is add a little bit of gamification

1168
01:08:54.880 --> 01:08:56.380
and sort of leaderboard type things.

1169
01:08:56.439 --> 01:09:01.620
You can see who in the organization are really the power users,

1170
01:09:01.800 --> 01:09:04.580
and you can maybe do some things around that.

1171
01:09:04.680 --> 01:09:07.859
Love your feedback on what would help with adoption.

1172
01:09:08.380 --> 01:09:11.520
But the point being is that this is a way that you can actually see

1173
01:09:11.520 --> 01:09:15.260
across the organization who's doing what and how it's working.

1174
01:09:16.479 --> 01:09:18.640
Any questions, comments on that?

1175
01:09:21.600 --> 01:09:22.080
All right.

1176
01:09:22.819 --> 01:09:26.120
Let's go down to the next thing I want to highlight is this little button

1177
01:09:26.120 --> 01:09:28.939
down in the bottom right here, which is the send feedback.

1178
01:09:29.399 --> 01:09:35.140
So it's really super helpful to us to just know what your experience is

1179
01:09:35.140 --> 01:09:37.240
as you're going along here.

1180
01:09:37.240 --> 01:09:42.340
If it's a bug report or a feature request, any improvement,

1181
01:09:42.560 --> 01:09:46.600
any general feedback, anything like that, it's a very simple little form.

1182
01:09:46.700 --> 01:09:48.300
Just give us the information.

1183
01:09:48.920 --> 01:09:53.920
Automatically, it would share the contact in terms of what your browser is

1184
01:09:53.920 --> 01:09:56.400
and your information so I know exactly who that is.

1185
01:09:57.140 --> 01:09:59.060
And then what we can do is we can look at that,

1186
01:09:59.100 --> 01:09:59.980
and we can see if there's a bug or if there's a feature that we need to fix.

1187
01:10:00.000 --> 01:10:03.080
what will happen is it'll appear in my sort of notifications

1188
01:10:03.240 --> 01:10:05.460
bell at the top here that I've got a new message.

1189
01:10:06.180 --> 01:10:08.620
And I can then, once I've solved that,

1190
01:10:08.800 --> 01:10:11.100
whatever that issue is, I can send you a note back

1191
01:10:11.100 --> 01:10:12.420
and you'll see it in your bell saying,

1192
01:10:12.700 --> 01:10:15.460
there you go, sorted, or here's the feedback,

1193
01:10:15.580 --> 01:10:16.580
or here's what the issue is.

1194
01:10:16.600 --> 01:10:18.740
So we've got a means of actually engaging

1195
01:10:18.740 --> 01:10:21.200
in a bit of a conversation behind the scenes.

1196
01:10:23.340 --> 01:10:23.840
All right.

1197
01:10:24.040 --> 01:10:25.800
I've been sending Glenn all those,

1198
01:10:26.100 --> 01:10:28.280
and it's so cool to go to the notifications.

1199
01:10:28.500 --> 01:10:29.800
That's done, that's done, that's done.

1200
01:10:29.800 --> 01:10:31.800
It's like a personal request,

1201
01:10:32.180 --> 01:10:33.800
support and getting immediate feedback.

1202
01:10:33.940 --> 01:10:34.880
Oh, we nice to have this.

1203
01:10:35.440 --> 01:10:36.980
Like, wouldn't it be great if we had a list view

1204
01:10:36.980 --> 01:10:38.020
and not just card view,

1205
01:10:38.020 --> 01:10:39.780
because I got so many prompts going.

1206
01:10:40.980 --> 01:10:44.460
So anyway, that is really helpful.

1207
01:10:44.640 --> 01:10:48.240
And so we want to be super responsive to you

1208
01:10:48.240 --> 01:10:51.020
so you can see your ideas, thoughts, bugs,

1209
01:10:51.380 --> 01:10:53.080
dealt with in real time.

1210
01:10:53.720 --> 01:10:54.160
Yeah.

1211
01:10:54.160 --> 01:10:55.400
Do you want to talk to us a little, Glenn,

1212
01:10:56.040 --> 01:10:59.780
about the whole beta testing period and everything else?

1213
01:10:59.780 --> 01:11:01.920
Let me do one more quick, fast little thing,

1214
01:11:02.080 --> 01:11:03.420
and then we'll switch to the beta testing.

1215
01:11:03.640 --> 01:11:05.080
There's a few more things we can touch,

1216
01:11:05.300 --> 01:11:06.120
more advanced topics,

1217
01:11:06.400 --> 01:11:08.620
but we can talk about those at some other point.

1218
01:11:09.120 --> 01:11:11.160
I didn't want to just touch variables.

1219
01:11:11.500 --> 01:11:12.600
I think I've mentioned to that,

1220
01:11:12.660 --> 01:11:13.780
but I want to quickly just share,

1221
01:11:14.240 --> 01:11:15.500
go to the prompt over here.

1222
01:11:15.560 --> 01:11:16.700
Let's just pull this one quickly.

1223
01:11:16.820 --> 01:11:21.320
I wanted to just highlight, edit over here.

1224
01:11:22.000 --> 01:11:25.320
What you can see is, let's go down here.

1225
01:11:25.580 --> 01:11:27.440
The way that you can put a variable,

1226
01:11:27.440 --> 01:11:30.220
easiest way to do this is actually just to create

1227
01:11:30.220 --> 01:11:31.960
a quick little prompt.

1228
01:11:33.340 --> 01:11:34.760
I wanted to show you,

1229
01:11:35.380 --> 01:11:38.200
if I wanted to create my prompt,

1230
01:11:43.900 --> 01:11:47.360
and all you do is you put in open curly brackets twice,

1231
01:11:48.160 --> 01:11:49.100
prompt title,

1232
01:11:55.800 --> 01:11:56.940
and little pipe,

1233
01:11:57.120 --> 01:12:03.080
and then you could say example one, example two,

1234
01:12:03.720 --> 01:12:06.020
those are the curly brace, and that's it.

1235
01:12:06.680 --> 01:12:08.180
Now, when I run that prompt,

1236
01:12:08.260 --> 01:12:09.360
if I just save that prompt,

1237
01:12:09.480 --> 01:12:11.140
and I go and have a look at Launchpad,

1238
01:12:11.460 --> 01:12:13.480
and I just open up my prompts over here,

1239
01:12:14.060 --> 01:12:15.480
and here's my test one,

1240
01:12:15.480 --> 01:12:16.720
it'll show you the one variable,

1241
01:12:16.880 --> 01:12:17.980
there's my prompt title,

1242
01:12:18.200 --> 01:12:19.080
and example one,

1243
01:12:19.080 --> 01:12:19.080


1244
01:12:19.080 --> 01:12:19.080


1245
01:12:19.080 --> 01:12:20.160
example two.

1246
01:12:20.620 --> 01:12:22.700
So that's how simple it is to add variables

1247
01:12:22.700 --> 01:12:23.460
into your prompt.

1248
01:12:24.840 --> 01:12:26.320
So where I've used that is,

1249
01:12:27.480 --> 01:12:30.680
hey, that prompt that took me a year at Brilliant

1250
01:12:30.680 --> 01:12:34.100
to sort out on how to translate Old Covenant,

1251
01:12:34.480 --> 01:12:37.880
Old Testament scriptures into a new creation,

1252
01:12:38.080 --> 01:12:39.400
new covenant language.

1253
01:12:39.960 --> 01:12:41.340
When I got it together,

1254
01:12:41.340 --> 01:12:42.900
I needed to know the topic,

1255
01:12:43.820 --> 01:12:45.960
what translation you wanted to use,

1256
01:12:46.040 --> 01:12:46.940
NIV, whatever,

1257
01:12:46.940 --> 01:12:48.160
and it put that in there.

1258
01:12:48.160 --> 01:12:50.500
And so you just create those,

1259
01:12:50.640 --> 01:12:52.380
and so anyone who's using it now just writes,

1260
01:12:52.540 --> 01:12:54.200
I wanna know about topic,

1261
01:12:55.280 --> 01:12:57.720
and then how many verses do you want?

1262
01:12:57.780 --> 01:12:59.100
So it asks for how many verses,

1263
01:12:59.300 --> 01:12:59.960
what translation,

1264
01:13:00.100 --> 01:13:01.180
and what's the topic got?

1265
01:13:01.580 --> 01:13:04.420
And we're gonna find 10 Old Covenant things,

1266
01:13:04.920 --> 01:13:05.960
do everything in NIV,

1267
01:13:06.200 --> 01:13:07.840
or whatever your preferred thing is,

1268
01:13:08.100 --> 01:13:09.460
come back, list it out,

1269
01:13:09.480 --> 01:13:11.440
and translate it using Brilliant's language.

1270
01:13:11.560 --> 01:13:12.880
That was a short prompt,

1271
01:13:12.920 --> 01:13:14.360
it was like 10 lines long.

1272
01:13:14.660 --> 01:13:16.420
But man, adding those variables,

1273
01:13:16.420 --> 01:13:18.940
I mean, everybody can use it for everything.

1274
01:13:19.880 --> 01:13:21.480
Yeah, yeah, super powerful.

1275
01:13:21.980 --> 01:13:23.360
Quick thing, and then I'll switch back

1276
01:13:23.420 --> 01:13:25.160
and talk about next steps.

1277
01:13:25.220 --> 01:13:27.320
But I wanted to also just highlight something new

1278
01:13:27.320 --> 01:13:28.260
that we're just putting in,

1279
01:13:28.300 --> 01:13:29.220
which is around the affiliate,

1280
01:13:29.620 --> 01:13:31.460
building mechanism for affiliate program.

1281
01:13:31.980 --> 01:13:33.880
So if you go up to your contact,

1282
01:13:34.000 --> 01:13:36.060
there'll be this thing called affiliate program.

1283
01:13:36.520 --> 01:13:37.500
And you can say,

1284
01:13:37.660 --> 01:13:40.060
I want my code to be Glenn Hobbs.

1285
01:13:42.420 --> 01:13:44.660
And you could say, okay, create my affiliate.

1286
01:13:44.660 --> 01:13:47.980
And now I've got a affiliate dashboard

1287
01:13:47.980 --> 01:13:51.640
set up for my affiliate.

1288
01:13:51.780 --> 01:13:55.260
And I can start sharing and tracking my affiliates.

1289
01:13:56.540 --> 01:13:59.040
So we can talk more about this later.

1290
01:13:59.800 --> 01:14:02.060
But this is, you know, Vanessa,

1291
01:14:02.360 --> 01:14:03.860
this is large part inspired

1292
01:14:03.860 --> 01:14:05.840
by some of the challenges we've had with flight

1293
01:14:05.840 --> 01:14:07.200
and all of this type of stuff.

1294
01:14:07.520 --> 01:14:08.900
And so I've been thinking about

1295
01:14:08.900 --> 01:14:10.000
what might it look like

1296
01:14:10.000 --> 01:14:11.720
if we could build something ourselves?

1297
01:14:11.960 --> 01:14:13.800
So I thought, well, let me try from Comfort,

1298
01:14:13.800 --> 01:14:15.540
see how it might work out.

1299
01:14:16.300 --> 01:14:18.720
And I started building, you know, this,

1300
01:14:19.240 --> 01:14:21.800
and it actually is working quite well.

1301
01:14:22.140 --> 01:14:23.140
It's working on the referrals.

1302
01:14:23.360 --> 01:14:24.940
All of that is working right now.

1303
01:14:25.240 --> 01:14:27.420
We still need to cook up the payouts,

1304
01:14:27.520 --> 01:14:29.100
which I found out actually Stripe

1305
01:14:29.100 --> 01:14:30.760
allows you to handle the payouts.

1306
01:14:30.980 --> 01:14:33.220
So we can handle all of that.

1307
01:14:34.100 --> 01:14:36.240
And you've got now a whole bit dashboard

1308
01:14:36.240 --> 01:14:38.300
to check in your recent commissions.

1309
01:14:39.660 --> 01:14:42.020
Also, what I've got is resources in here.

1310
01:14:42.020 --> 01:14:43.820
So some pretty cool things,

1311
01:14:44.080 --> 01:14:45.080
like if you want to start a kit

1312
01:14:45.080 --> 01:14:47.020
for some marketing materials,

1313
01:14:47.280 --> 01:14:48.320
like if you want to,

1314
01:14:48.460 --> 01:14:49.660
let's just pick one here,

1315
01:14:49.940 --> 01:14:53.260
like this AI is just family on steroids one.

1316
01:14:53.500 --> 01:14:55.360
So here's a series of tweets,

1317
01:14:56.720 --> 01:14:58.480
you know, that you could use

1318
01:14:58.540 --> 01:14:59.980
and tailor to your liking.

1319
01:15:00.000 --> 01:15:06.140
King. But he has a very cool, you know, think about it, best

1320
01:15:06.140 --> 01:15:08.420
prompts to supply after use and organizational learning,

1321
01:15:08.580 --> 01:15:11.160
interaction starts from zero, no competitive advantage

1322
01:15:11.280 --> 01:15:15.620
accumulation. So it's a nice little tweet thing. And so

1323
01:15:15.620 --> 01:15:19.220
there's a whole series of these, which could help a lot with

1324
01:15:19.220 --> 01:15:26.020
just sort of starting to share the story. So there we go, I'll

1325
01:15:26.020 --> 01:15:28.380
stop sharing on this one, any questions on that, and then we

1326
01:15:28.380 --> 01:15:31.500
can switch back to just wrapping up on next steps.

1327
01:15:33.300 --> 01:15:35.680
Yeah, do you have?

1328
01:15:37.040 --> 01:15:41.720
One thing I think that would be really helpful is I'm just

1329
01:15:41.720 --> 01:15:44.340
thinking with adoption and people adopting this, and

1330
01:15:44.460 --> 01:15:49.780
obviously, it's, it's going to be evolving, you know. And

1331
01:15:49.780 --> 01:15:52.640
that's, that's kind of the part of the challenge and the

1332
01:15:52.640 --> 01:15:55.220
struggle it feels like is everything is so fluid, and

1333
01:15:55.220 --> 01:16:00.800
it's moving so fast. And, you know, but with, I'm just some

1334
01:16:00.800 --> 01:16:04.540
of the platforms that we use the softwares for different things,

1335
01:16:04.540 --> 01:16:06.840
they have like a little knowledge base, you know, you

1336
01:16:06.840 --> 01:16:11.100
got some basic, yeah, little tutorial things, those would be

1337
01:16:11.100 --> 01:16:12.520
super to actually,

1338
01:16:13.360 --> 01:16:15.740
actually, let me share that I didn't go there. But the answer

1339
01:16:15.740 --> 01:16:20.500
is yes. Let me quickly just share so you can see. Let's go

1340
01:16:20.500 --> 01:16:24.740
back here. I think it's this one. Can you see the page is the

1341
01:16:24.740 --> 01:16:28.760
help center over here with quick start guide, enterprise

1342
01:16:28.900 --> 01:16:33.380
government, governance, and then a number of resources and

1343
01:16:33.380 --> 01:16:36.900
learning. And so and so the idea is to start seeding this, this

1344
01:16:36.900 --> 01:16:40.920
is this is still very light, it still needs a lot more work to

1345
01:16:40.920 --> 01:16:46.900
see that. But it's exactly that Jeremy is to really create a set

1346
01:16:46.900 --> 01:16:51.140
of artifacts that would help you in the adoption of helping to

1347
01:16:51.140 --> 01:16:53.700
communicate the story. So, you know, it's one thing when you

1348
01:16:53.700 --> 01:16:57.060
hear me talk about it, but how do you have the message in an

1349
01:16:57.060 --> 01:16:59.520
encapsulated format, you can take it very easily to someone

1350
01:16:59.280 --> 01:17:02.700
else. And, and by the way, on that list, that's actually a

1351
01:17:03.000 --> 01:17:06.800
great pivot to the next thing that we wanted to talk about,

1352
01:17:06.920 --> 01:17:12.620
which is over the next 30 days on the beta during this beta

1353
01:17:12.620 --> 01:17:16.600
test period, let me let me switch back to the slides of a

1354
01:17:16.600 --> 01:17:21.120
slide or two on that. Let's go over here. And I have to put

1355
01:17:21.120 --> 01:17:26.560
this into play mode. Otherwise, it doesn't work. Okay. So just

1356
01:17:26.560 --> 01:17:33.620
from an onboarding, what, you know, go to mypromptforge.com.

1357
01:17:34.580 --> 01:17:37.180
If you go there right now, you'll see a landing page, if

1358
01:17:37.180 --> 01:17:39.000
you haven't already seen it, it's got a lot of that

1359
01:17:39.120 --> 01:17:42.540
messaging that I've shared at the top end of the call. It's

1360
01:17:42.540 --> 01:17:46.340
got a link to that podcast, a lot of the slides, it's all

1361
01:17:46.340 --> 01:17:52.660
there. Pick a plan. And start, start, start playing with it,

1362
01:17:52.660 --> 01:17:55.880
start using it started getting, you know, getting your hands on

1363
01:17:55.880 --> 01:17:59.560
it and and providing the feedback. I mentioned free

1364
01:17:59.560 --> 01:18:01.660
credits, we'll load it up with free credits, you don't have to

1365
01:18:01.660 --> 01:18:05.020
worry about that. And then we'll also run a weekly discovery

1366
01:18:05.020 --> 01:18:09.840
call. I actually think I put it on the next. Yeah. Next page.

1367
01:18:09.840 --> 01:18:15.800
Yeah. Basically, I was gonna say as your role as an admin,

1368
01:18:16.260 --> 01:18:20.120
what what would you need to do one is obviously register and

1369
01:18:20.120 --> 01:18:23.300
select a plan, but then manage your AI models go as I showed,

1370
01:18:23.580 --> 01:18:25.540
just go and decide which ones you want to have available to

1371
01:18:25.540 --> 01:18:29.800
your team, from the list that you have access to invite your

1372
01:18:29.800 --> 01:18:34.800
users set up the teams, manage the payment subscriptions, and

1373
01:18:34.800 --> 01:18:37.760
then the you know, drive adoption and organization. I

1374
01:18:37.760 --> 01:18:41.320
think that's really the conversation is the next sort

1375
01:18:41.320 --> 01:18:45.580
of step. And what you and I were thinking about with that is, I

1376
01:18:45.580 --> 01:18:49.140
think that's where we can help each other in that respect.

1377
01:18:49.140 --> 01:18:52.500
Because I think the next sort of frontier, if you will, it's not

1378
01:18:52.500 --> 01:18:55.320
the platform itself, it's now going to be the adoption of the

1379
01:18:55.320 --> 01:18:59.520
platform you use within an organization. And so starting to

1380
01:18:59.520 --> 01:19:02.880
understand what the points of friction are, what are the what

1381
01:19:02.880 --> 01:19:07.980
you need in order for this to, to really help you be able to

1382
01:19:07.980 --> 01:19:10.120
connect this into your organization and make it work.

1383
01:19:10.420 --> 01:19:15.440
And Deon and I have a long list of possible things. For example,

1384
01:19:16.040 --> 01:19:20.620
maybe it could be well, in a journey, you asked for some sort

1385
01:19:20.620 --> 01:19:23.720
of messaging, some sort of package story that you can share

1386
01:19:23.720 --> 01:19:27.620
in a very digestible fashion. And we'll certainly make sure

1387
01:19:27.780 --> 01:19:31.200
that that we provide that that's vital. We'll have that there.

1388
01:19:31.200 --> 01:19:34.460
That's super important. But maybe there are other things

1389
01:19:34.460 --> 01:19:38.820
that could be helpful. For example, you know, certainly my

1390
01:19:38.820 --> 01:19:41.420
experience from consulting, one of the things we always did for

1391
01:19:41.420 --> 01:19:44.560
clients is we do some sort of baseline assessment that helped

1392
01:19:44.560 --> 01:19:47.760
create visibility to the reality in the organization, here's,

1393
01:19:47.760 --> 01:19:50.200
here's what's happening, here's who's doing what, here's what's

1394
01:19:50.200 --> 01:19:53.380
going on, so that we could bring bring up shine a light, if you

1395
01:19:53.380 --> 01:19:56.140
will, onto the reality of what's going on. And that also can

1396
01:19:56.140 --> 01:19:59.840
baseline to help understand, okay, from

1397
01:20:00.000 --> 01:20:03.400
From here, how have we moved organization forwards?

1398
01:20:04.420 --> 01:20:06.740
And so I think that can be something,

1399
01:20:07.160 --> 01:20:09.500
it's in our thinking to work to kind of create

1400
01:20:09.500 --> 01:20:10.200
something like that,

1401
01:20:10.200 --> 01:20:11.680
if that's going to be meaningful or helpful.

1402
01:20:12.260 --> 01:20:13.900
But the key point is,

1403
01:20:14.140 --> 01:20:16.780
is that we're going to have a beta support call

1404
01:20:16.780 --> 01:20:18.760
every Wednesday from one to 1.30.

1405
01:20:20.000 --> 01:20:21.540
And the nature of that call

1406
01:20:21.540 --> 01:20:23.080
would be to help answer these questions,

1407
01:20:23.140 --> 01:20:24.460
because once you start getting in,

1408
01:20:25.240 --> 01:20:26.700
I mean, I've covered so much ground,

1409
01:20:26.820 --> 01:20:28.840
it's going to be hard to remember everything.

1410
01:20:28.840 --> 01:20:30.220
You know, you'll get a recording,

1411
01:20:30.400 --> 01:20:32.140
but it might be easier just to get on a call and say,

1412
01:20:32.560 --> 01:20:34.880
hey, I'm struggling with this, or how do I do that?

1413
01:20:34.960 --> 01:20:36.100
Or whatever the case may be.

1414
01:20:36.560 --> 01:20:38.420
But it can also be that in,

1415
01:20:38.440 --> 01:20:39.860
we'd love to use that as a time

1416
01:20:39.860 --> 01:20:41.420
to kind of get that feedback to say,

1417
01:20:41.480 --> 01:20:43.120
okay, here's the thing that we need

1418
01:20:43.220 --> 01:20:45.100
in order to help drive better adoption

1419
01:20:45.100 --> 01:20:46.280
inside of the organization

1420
01:20:46.280 --> 01:20:47.880
and start that conversation going.

1421
01:20:48.000 --> 01:20:50.160
So we know what we need to work on

1422
01:20:50.800 --> 01:20:54.460
to really, you know, get it to connect well for you.

1423
01:20:55.580 --> 01:20:56.860
So that's the story.

1424
01:21:01.800 --> 01:21:04.840
So you definitely hit the questions I was thinking.

1425
01:21:06.140 --> 01:21:07.340
I mean, great job.

1426
01:21:07.600 --> 01:21:09.620
I, you know, adoption, how you onboard it,

1427
01:21:09.700 --> 01:21:10.940
how you introduce it.

1428
01:21:11.140 --> 01:21:14.840
How do we continue to be a construction company

1429
01:21:14.840 --> 01:21:17.060
and not an AI development company

1430
01:21:17.140 --> 01:21:18.740
inside of a construction company

1431
01:21:19.360 --> 01:21:23.380
as, you know, employees get these fun tools,

1432
01:21:23.380 --> 01:21:24.440
if that makes sense.

1433
01:21:24.780 --> 01:21:27.440
So I think that'll be a key thing for me,

1434
01:21:27.440 --> 01:21:28.820
how to encourage adoption

1435
01:21:29.460 --> 01:21:31.480
yet not go down the rabbit hole

1436
01:21:31.480 --> 01:21:34.780
and spend five hours a day, you know,

1437
01:21:34.780 --> 01:21:36.220
creating context, if that makes sense.

1438
01:21:37.100 --> 01:21:37.440
100%.

1439
01:21:37.440 --> 01:21:41.040
And, you know, use the phrase right at the top of the call,

1440
01:21:41.480 --> 01:21:42.420
Ben, about use cases,

1441
01:21:42.600 --> 01:21:44.060
I think was the phrase you used.

1442
01:21:45.020 --> 01:21:46.640
I think it's a great way to think about it.

1443
01:21:46.780 --> 01:21:48.780
It's like, okay, what are those use cases

1444
01:21:48.780 --> 01:21:49.640
for your organization?

1445
01:21:49.720 --> 01:21:51.780
And that'd be something we'd be very happy

1446
01:21:51.780 --> 01:21:53.000
to partner with you to work on

1447
01:21:53.000 --> 01:21:53.960
and have that conversation,

1448
01:21:54.120 --> 01:21:55.780
say, you know, as you have those questions,

1449
01:21:55.780 --> 01:21:57.260
because you did correctly.

1450
01:21:57.340 --> 01:21:59.260
So your business is construction.

1451
01:21:59.420 --> 01:22:02.880
You don't want to have people sort of sidetracked off now,

1452
01:22:02.960 --> 01:22:05.420
going, chasing this next new shiny object.

1453
01:22:05.480 --> 01:22:08.520
You want this thing to empower your business

1454
01:22:08.520 --> 01:22:10.060
in a meaningful way without distracting.

1455
01:22:11.060 --> 01:22:15.560
And so Dion and I would love to have that conversation

1456
01:22:15.560 --> 01:22:17.920
with you to kick that around a little bit,

1457
01:22:18.000 --> 01:22:20.740
figure out, okay, how do we help you do this

1458
01:22:20.740 --> 01:22:24.600
in a way that's, you know, not depletive,

1459
01:22:24.600 --> 01:22:26.360
but accretive to your business?

1460
01:22:27.120 --> 01:22:29.640
And I would argue that they're probably doing it now.

1461
01:22:30.160 --> 01:22:32.380
They are some, I mean, we're sharing.

1462
01:22:32.760 --> 01:22:34.820
The shadow version.

1463
01:22:35.460 --> 01:22:38.580
But I think once we have transparency and it's managed,

1464
01:22:38.840 --> 01:22:40.540
so we've got step one is we've got to move

1465
01:22:40.540 --> 01:22:42.360
out of the shadows into the managed.

1466
01:22:43.200 --> 01:22:45.140
Step two then is we got to show,

1467
01:22:45.580 --> 01:22:47.660
let them discover the power in steps.

1468
01:22:48.060 --> 01:22:48.360
Yeah.

1469
01:22:48.880 --> 01:22:52.100
We just blew out everything for you to see everything,

1470
01:22:52.140 --> 01:22:53.580
but they don't need to stop there.

1471
01:22:53.580 --> 01:22:55.660
They can just start, here's a launch,

1472
01:22:56.060 --> 01:22:57.120
write your things, done.

1473
01:22:57.240 --> 01:22:59.860
Oh, do you know that you can make these prompts permanent?

1474
01:23:00.260 --> 01:23:01.960
Do you know that there's a thousand prompts

1475
01:23:01.960 --> 01:23:02.620
you can choose from?

1476
01:23:02.680 --> 01:23:03.360
You don't have to write them.

1477
01:23:03.500 --> 01:23:05.280
Okay, so maybe that's where we start.

1478
01:23:05.640 --> 01:23:07.480
We can add context over time,

1479
01:23:08.940 --> 01:23:12.420
but when we're demonstrating to you as business owners,

1480
01:23:12.720 --> 01:23:14.100
the full context of everything,

1481
01:23:14.220 --> 01:23:16.740
we're trying to show you how you're going to manage it.

1482
01:23:17.280 --> 01:23:19.220
What would be useful every week?

1483
01:23:19.320 --> 01:23:20.820
And we only set up for half an hour,

1484
01:23:20.900 --> 01:23:22.700
but if it needs to go longer, we'll make it longer.

1485
01:23:22.700 --> 01:23:23.400
That's not the problem.

1486
01:23:23.400 --> 01:23:25.580
And you can bring whoever you like to those calls.

1487
01:23:25.600 --> 01:23:27.420
It's not any so-called people.

1488
01:23:27.480 --> 01:23:30.720
It's just bring your team, bring your whole team,

1489
01:23:30.780 --> 01:23:31.260
if you like.

1490
01:23:31.420 --> 01:23:36.860
The issue would be how do we get them mobile and productive,

1491
01:23:36.900 --> 01:23:39.100
but in a way that you can manage.

1492
01:23:39.480 --> 01:23:42.000
And I think this is the first time

1493
01:23:42.000 --> 01:23:43.680
we're going to bring it out of the shadows.

1494
01:23:44.400 --> 01:23:46.800
Second thing, we want to discover where the value is.

1495
01:23:46.800 --> 01:23:49.860
And that's where you'll start having conversations,

1496
01:23:49.860 --> 01:23:51.880
I think, like we've been having at Brilliant.

1497
01:23:54.560 --> 01:23:58.080
Hey, why are we all generating the same context?

1498
01:23:58.600 --> 01:24:01.080
We don't all have to describe what Brilliant does.

1499
01:24:01.280 --> 01:24:03.540
We don't all have to describe Brilliant's products.

1500
01:24:03.660 --> 01:24:04.400
We don't all have,

1501
01:24:04.460 --> 01:24:06.900
and we don't need to have all that information all the time.

1502
01:24:07.520 --> 01:24:11.120
So we need to break those up into smaller contexts.

1503
01:24:11.780 --> 01:24:14.600
And then what's actually useful?

1504
01:24:15.060 --> 01:24:16.500
Because you might create a context

1505
01:24:16.500 --> 01:24:19.140
and find it actually makes no difference to the output.

1506
01:24:19.140 --> 01:24:20.160
It's just worthless.

1507
01:24:20.280 --> 01:24:21.900
So we can delete that.

1508
01:24:22.080 --> 01:24:26.320
So I think what we're going to do is over a period of time,

1509
01:24:26.320 --> 01:24:28.400
hone in on where the value is.

1510
01:24:28.660 --> 01:24:31.220
And I think that's where we had to start with honesty

1511
01:24:31.220 --> 01:24:33.900
at the front end of, we don't know where the value is,

1512
01:24:34.060 --> 01:24:35.620
but at least let's get it in the light.

1513
01:24:35.680 --> 01:24:36.840
Let's start the journey.

1514
01:24:37.260 --> 01:24:39.300
But I think what you'll see immediately,

1515
01:24:39.380 --> 01:24:41.840
and we try to set the subscription separate

1516
01:24:42.080 --> 01:24:44.060
from the tokens thing,

1517
01:24:44.220 --> 01:24:48.760
so that there was an immediate 40 to 60% savings

1518
01:24:48.760 --> 01:24:51.260
off the bat, because you don't need a high-end tool.

1519
01:24:52.000 --> 01:24:53.640
You know, I'm a super grok user.

1520
01:24:54.080 --> 01:24:56.360
I have the highest on all of them,

1521
01:24:56.440 --> 01:24:58.700
and I have four of them, and I use them.

1522
01:25:00.000 --> 01:25:02.600
probably spend hundreds of dollars,

1523
01:25:02.880 --> 01:25:04.500
maybe several hundred dollars,

1524
01:25:04.640 --> 01:25:07.580
if I could just control all that

1525
01:25:07.580 --> 01:25:09.600
and have control over the costs.

1526
01:25:09.840 --> 01:25:11.840
So I think we also want to go into cost management.

1527
01:25:12.300 --> 01:25:16.620
And now it's a predictable amount of money each month.

1528
01:25:17.140 --> 01:25:20.240
And you can see who's using it before you couldn't.

1529
01:25:21.820 --> 01:25:26.000
So I think knowing what will be great is,

1530
01:25:26.000 --> 01:25:30.420
I think some making heroes of the story,

1531
01:25:31.080 --> 01:25:33.700
using our edge training, leader forge training,

1532
01:25:34.500 --> 01:25:36.580
like, hey man, this person did this,

1533
01:25:36.780 --> 01:25:38.980
cracked the code, did that, saved us.

1534
01:25:39.880 --> 01:25:41.540
Whoa, you know, whatever.

1535
01:25:42.840 --> 01:25:44.440
So I think we want to board,

1536
01:25:45.000 --> 01:25:46.900
I'm excited that we get to give this

1537
01:25:46.900 --> 01:25:48.280
to kingdom people first.

1538
01:25:48.840 --> 01:25:52.400
And I'm excited that, in a circle.

1539
01:25:52.840 --> 01:25:55.260
So we're gonna figure it out with you.

1540
01:25:55.260 --> 01:25:56.060
Short answer.

1541
01:25:56.240 --> 01:25:57.180
So we don't know entirely,

1542
01:25:57.320 --> 01:25:59.240
but we're starting the journey.

1543
01:25:59.420 --> 01:26:01.220
So if you want to come with us,

1544
01:26:01.660 --> 01:26:03.900
Glen, what do we have to do to get people connected?

1545
01:26:05.180 --> 01:26:07.620
Simply go to mypromforge.com

1546
01:26:09.380 --> 01:26:12.540
and you can just register there.

1547
01:26:14.540 --> 01:26:17.220
No issues, very straightforward.

1548
01:26:18.720 --> 01:26:21.120
Just get on board and start using it.

1549
01:26:22.020 --> 01:26:23.840
And the bottom point there,

1550
01:26:23.840 --> 01:26:26.560
Hypercare will be right there with you all the way along.

1551
01:26:27.200 --> 01:26:28.740
Feel free, you have my cell phone number,

1552
01:26:29.040 --> 01:26:30.780
feel free to text me if you have any issue,

1553
01:26:30.840 --> 01:26:33.180
more than happy to help and get going.

1554
01:26:33.980 --> 01:26:36.340
So that of course will only be this month

1555
01:26:36.340 --> 01:26:37.280
with us beta group.

1556
01:26:37.340 --> 01:26:38.820
I'm not gonna share that with the world,

1557
01:26:38.980 --> 01:26:42.480
but for us, I just want to, you know,

1558
01:26:42.480 --> 01:26:45.280
just really help get you going.

1559
01:26:46.480 --> 01:26:49.380
So nothing more than go to mypromforge.com,

1560
01:26:49.620 --> 01:26:50.600
pick the plan.

1561
01:26:53.960 --> 01:26:57.320
And start using it and please share the feedback.

1562
01:26:57.420 --> 01:26:59.040
And then we've got the Wednesday calls.

1563
01:26:59.140 --> 01:27:00.600
We'll have that invite sent out

1564
01:27:01.220 --> 01:27:05.540
and we'll be looking forward to the conversation with you

1565
01:27:05.540 --> 01:27:06.560
and learning how it goes.

1566
01:27:06.640 --> 01:27:09.520
At the end, you know, obviously for Dionne and I,

1567
01:27:09.700 --> 01:27:12.120
it's really vital that we start getting stories

1568
01:27:12.120 --> 01:27:16.860
of the success and be able to start sort of capturing that

1569
01:27:16.860 --> 01:27:18.240
and working with you to achieve that.

1570
01:27:19.360 --> 01:27:22.060
Your stories are very valuable to us too.

1571
01:27:22.200 --> 01:27:23.300
Let's do a quick round robin,

1572
01:27:23.420 --> 01:27:25.840
give everyone a chance just to say feedback.

1573
01:27:26.480 --> 01:27:27.600
Jim, do you want to kick it off?

1574
01:27:28.260 --> 01:27:30.520
You're a big user of AI.

1575
01:27:31.180 --> 01:27:32.780
Yeah, yeah, my head's just kind of spinning,

1576
01:27:33.900 --> 01:27:35.080
you know, looking at it.

1577
01:27:35.080 --> 01:27:36.660
I think what was really helpful was to see an example.

1578
01:27:37.040 --> 01:27:39.780
I wasn't aware that there was like 1,000 prompts

1579
01:27:39.780 --> 01:27:40.720
that could be used.

1580
01:27:41.100 --> 01:27:42.800
I think it was really helpful to know that,

1581
01:27:43.220 --> 01:27:46.260
you know, the last I heard was that 275 page prompt.

1582
01:27:46.740 --> 01:27:47.980
And so I was feeling a little overwhelmed,

1583
01:27:47.980 --> 01:27:51.100
I'm like, oh my gosh, it's gonna take us forever.

1584
01:27:51.560 --> 01:27:52.900
So yeah, yeah, this is helpful.

1585
01:27:52.960 --> 01:27:54.080
It's helpful to see it play out.

1586
01:27:54.220 --> 01:27:55.840
I mean, I can see how it'd be helpful.

1587
01:27:56.080 --> 01:27:58.580
I've just seen firsthand how the better prompt,

1588
01:27:58.760 --> 01:27:59.740
the better answer you get.

1589
01:28:00.780 --> 01:28:02.640
And yeah, and I've had that where I'm going back

1590
01:28:02.640 --> 01:28:04.900
and searching for that prompt, where did I have that?

1591
01:28:05.680 --> 01:28:08.200
And, you know, I have no idea where it is.

1592
01:28:08.820 --> 01:28:10.380
So yeah, it's super cool.

1593
01:28:10.500 --> 01:28:11.540
Thanks, thanks for this, this is good.

1594
01:28:12.140 --> 01:28:15.380
I'm thrilled you're on and it's great, Jeremy.

1595
01:28:16.900 --> 01:28:22.180
Yep, so this is amazing, super, super ecstatic and excited.

1596
01:28:22.620 --> 01:28:27.080
Thank you, everybody who's been working on super timely.

1597
01:28:27.640 --> 01:28:32.900
We're about to do an event of this group we're in

1598
01:28:32.900 --> 01:28:35.700
called the Roofing and Solar Reform Alliance.

1599
01:28:35.760 --> 01:28:37.260
They're having their annual gathering

1600
01:28:37.260 --> 01:28:42.220
and the whole topic is AI and, you know,

1601
01:28:42.380 --> 01:28:44.380
everybody coming together and sharing

1602
01:28:44.380 --> 01:28:46.540
how they're using it in their company.

1603
01:28:46.880 --> 01:28:49.060
And they've, you know, a lot of people

1604
01:28:49.060 --> 01:28:50.360
are starting to do it.

1605
01:28:50.640 --> 01:28:53.940
And so this is, it's a little bit of the wild, wild West

1606
01:28:53.940 --> 01:28:54.880
though, you know what I mean?

1607
01:28:54.960 --> 01:28:56.980
And so this is just amazing.

1608
01:28:57.240 --> 01:28:59.620
So I'm really, really excited

1609
01:28:59.620 --> 01:29:02.700
to start getting familiar with this.

1610
01:29:03.240 --> 01:29:05.520
Appreciate it, Vanessa.

1611
01:29:09.480 --> 01:29:12.540
I love it, I've signed up already

1612
01:29:12.540 --> 01:29:14.960
and I can't wait to get started.

1613
01:29:16.160 --> 01:29:17.180
Was I the first, Glenn?

1614
01:29:17.680 --> 01:29:21.940
I think you are, yeah, we'll have to find some,

1615
01:29:21.940 --> 01:29:23.980
some surprise, you know, some prize for first.

1616
01:29:24.580 --> 01:29:24.580


1617
01:29:25.280 --> 01:29:26.100
That's so cool.

1618
01:29:26.500 --> 01:29:28.180
No, the prize is I get to use it.

1619
01:29:28.180 --> 01:29:29.860
That's good enough, thank you.

1620
01:29:30.080 --> 01:29:31.820
Awesome, awesome.

1621
01:29:33.320 --> 01:29:35.240
Jenny, then Ben, we'll wrap it.

1622
01:29:35.600 --> 01:29:39.260
I love that you hit the main pain points of, you know,

1623
01:29:39.260 --> 01:29:42.460
not capturing what you, previous prompts

1624
01:29:42.460 --> 01:29:44.380
and as well, the team projects.

1625
01:29:44.620 --> 01:29:46.540
I love the collaboration piece of it

1626
01:29:46.540 --> 01:29:50.720
because you have some folks who are stronger in this

1627
01:29:50.720 --> 01:29:52.060
and that's gonna help the others

1628
01:29:52.060 --> 01:29:54.140
who are not as strong benefit

1629
01:29:54.180 --> 01:29:56.740
from that experience and that learning.

1630
01:29:57.780 --> 01:29:59.920
I'm really excited to dive

1631
01:30:00.000 --> 01:30:03.220
and to get the team on boarded with it.

1632
01:30:03.580 --> 01:30:05.760
Awesome, Ben, give you the last word.

1633
01:30:06.360 --> 01:30:09.540
Yeah, I think you guys have been very busy.

1634
01:30:09.780 --> 01:30:11.880
Man, it's come a long way since we were together.

1635
01:30:12.060 --> 01:30:12.500
Great job.

1636
01:30:12.680 --> 01:30:14.340
It really looks great.

1637
01:30:14.380 --> 01:30:18.040
Looks like you've thought, it's well thought out.

1638
01:30:18.400 --> 01:30:21.240
And I'm excited to jump in and learn it.

1639
01:30:21.600 --> 01:30:24.440
And ultimately I think it is an answer

1640
01:30:24.440 --> 01:30:26.600
to a problem we've had, right?

1641
01:30:26.600 --> 01:30:31.580
Which is how do we across the team leverage this tool

1642
01:30:31.580 --> 01:30:34.500
in a way that seems not chaotic.

1643
01:30:34.760 --> 01:30:35.860
It pulls it out of the shadows.

1644
01:30:36.280 --> 01:30:38.780
So I love the visibility and transparency

1645
01:30:38.780 --> 01:30:40.340
in the pricing as well.

1646
01:30:41.200 --> 01:30:44.580
I'm interested to learn kind of the different tiers

1647
01:30:44.580 --> 01:30:45.740
and what makes sense.

1648
01:30:46.640 --> 01:30:51.760
Right now, I signed up and didn't get the magic link back.

1649
01:30:51.780 --> 01:30:53.700
So I don't know, I'm trying to figure that out, Ben Reid.

1650
01:30:53.980 --> 01:30:55.500
Okay, I'll help you with that.

1651
01:30:55.500 --> 01:30:58.380
If, you know, I'll sync with you offline

1652
01:30:58.380 --> 01:31:00.480
if you had any issue there.

1653
01:31:00.700 --> 01:31:02.600
Just check the spam box if, you know,

1654
01:31:02.620 --> 01:31:03.880
that could be an issue.

1655
01:31:04.560 --> 01:31:04.920
Yeah, I see.

1656
01:31:05.260 --> 01:31:06.700
But thank you, it looks great.

1657
01:31:06.960 --> 01:31:08.040
I'm excited to learn.

1658
01:31:09.360 --> 01:31:09.360


1659
01:31:09.840 --> 01:31:10.340
Thanks, everyone.

1660
01:31:10.480 --> 01:31:11.660
I'd like to try to keep as close

1661
01:31:11.660 --> 01:31:13.040
to the 90 minutes as possible.

1662
01:31:13.160 --> 01:31:14.000
We know you're really busy.

1663
01:31:14.400 --> 01:31:15.660
I know we certainly are.

1664
01:31:15.800 --> 01:31:17.520
We're about to launch a big event for the year

1665
01:31:17.640 --> 01:31:20.600
in a few hours, so yeah.

1666
01:31:20.600 --> 01:31:24.360
So we're excited and thanks again.

1667
01:31:24.360 --> 01:31:26.280
And, you know, I think a lot of this

1668
01:31:26.280 --> 01:31:28.400
was highly informed by our own experience

1669
01:31:28.400 --> 01:31:29.700
as a business at Brilliant

1670
01:31:30.400 --> 01:31:32.760
and, you know, learning and trying to manage this.

1671
01:31:32.860 --> 01:31:34.980
And we've been very active in this AI space

1672
01:31:34.980 --> 01:31:37.840
for well over two years with active training

1673
01:31:37.840 --> 01:31:39.400
and made a lot of mistakes.

1674
01:31:39.620 --> 01:31:41.880
So hopefully we can shortcut all of that for you,

1675
01:31:42.200 --> 01:31:43.560
take you straight to the bottom line

1676
01:31:43.560 --> 01:31:44.820
of what actually works.

1677
01:31:45.100 --> 01:31:47.760
And hopefully this tool will take us to a whole new level.

1678
01:31:48.240 --> 01:31:50.960
So, hey, if you would, please pray for us this

1679
01:31:50.960 --> 01:31:54.180
because we would love to see something big

1680
01:31:54.180 --> 01:31:56.360
come out of the kingdom, you know, stuff.

1681
01:31:56.680 --> 01:32:00.260
And I'd love to have a whole opportunity

1682
01:32:00.600 --> 01:32:04.720
to give this as an advantage to you.

1683
01:32:05.200 --> 01:32:06.220
So thanks so much.

1684
01:32:06.480 --> 01:32:06.960
Thanks for joining.

1685
01:32:07.180 --> 01:32:08.800
I know there was a lot of just dialogue,

1686
01:32:09.040 --> 01:32:11.320
you know, from us to you, not much interaction,

1687
01:32:11.560 --> 01:32:13.020
but there was a lot to share.

1688
01:32:13.700 --> 01:32:16.940
So we'll go back to normal format next month

1689
01:32:16.940 --> 01:32:18.040
where we actually talk.

1690
01:32:19.040 --> 01:32:19.580
Enjoy.

1691
01:32:20.960 --> 01:32:21.340
Thanks all.