1
00:00:00,040 --> 00:00:04,520
The percentage of all the 
completed work done autonomously

2
00:00:04,520 --> 00:00:08,640
by Blitzie, it ranged from 80 
and 95%. 

3
00:00:08,840 --> 00:00:12,560
Depending in the use case. 
We accelerated development 

4
00:00:12,560 --> 00:00:15,120
velocity of between 5:00 and 
10:00 X. 

5
00:00:15,120 --> 00:00:17,480
The human is not writing the 
code. 

6
00:00:17,680 --> 00:00:20,880
The human is directing a 
platform on how to write the 

7
00:00:20,880 --> 00:00:22,680
code. 
That's a huge change in 

8
00:00:22,680 --> 00:00:26,760
paradigm. 
That's Enrique Ibarra, CIO and 

9
00:00:26,760 --> 00:00:31,040
Head of Business Transformation 
at GNP, Mexico's largest 

10
00:00:31,040 --> 00:00:34,640
insurance company. 
Our conversation covers AI 

11
00:00:34,640 --> 00:00:39,720
adoption, change management, and
autonomous software development.

12
00:00:40,120 --> 00:00:44,320
Our main operational system is a
mainframe based map with IBM 

13
00:00:44,320 --> 00:00:46,440
components. 
System has been running for a 

14
00:00:46,480 --> 00:00:49,840
little bit over 20 years. 
The initial rationale for 

15
00:00:49,840 --> 00:00:53,160
modernizing this application was
basically cost. 

16
00:00:53,320 --> 00:00:56,360
That was one of the concerns, 
but it was not the only one. 

17
00:00:56,560 --> 00:01:01,880
In a few years from now, it's 
going to be much harder to get 

18
00:01:01,880 --> 00:01:04,080
COBOL resources. 
I mean, if you go to the 

19
00:01:04,080 --> 00:01:07,360
universities, students don't 
learn COBOL in the universities.

20
00:01:07,360 --> 00:01:10,640
There's no real interest. 
That's going to be an issue I 

21
00:01:10,640 --> 00:01:14,680
guess in the future to in terms 
of how do we give longevity to 

22
00:01:14,680 --> 00:01:17,880
this asset. 
We started a few years ago 

23
00:01:17,880 --> 00:01:22,200
incorporating the typical coding
Co pilots. 

24
00:01:22,840 --> 00:01:26,200
We, we incorporated several 
informally. 

25
00:01:26,200 --> 00:01:29,920
We started to provide all of our
developers with Co pilots 

26
00:01:30,120 --> 00:01:33,200
without formally measuring their
increasing productivity. 

27
00:01:33,200 --> 00:01:37,480
We just basically use it, find 
it useful, take advantage of it,

28
00:01:37,680 --> 00:01:42,000
and that's being expanded among 
the all of our developer base 

29
00:01:42,000 --> 00:01:45,560
and we have roughly about 1000 
developers by looking at 

30
00:01:45,560 --> 00:01:49,720
different industry solutions, we
find we basically learn about 

31
00:01:49,720 --> 00:01:54,560
blitzing the value proposition 
sounded too good to be true. 

32
00:01:55,240 --> 00:01:58,280
It was a little bit like magic 
is like you give me the 

33
00:01:58,280 --> 00:02:02,000
specifications, give me the code
and we will autonomously 

34
00:02:02,000 --> 00:02:04,560
generate everything. 
It was very, it was very 

35
00:02:04,560 --> 00:02:08,039
attractive. 
So we went and visited blitzing 

36
00:02:08,039 --> 00:02:12,240
their their offices in Boston to
learn, you know, what was their 

37
00:02:12,240 --> 00:02:15,360
vision, what was their product? 
I mean, just meet with, you 

38
00:02:15,360 --> 00:02:19,360
know, founders on the management
and we like the approach. 

39
00:02:19,640 --> 00:02:23,120
It was very aligned with what 
we're trying to achieve in the 

40
00:02:23,200 --> 00:02:28,520
in, you know, here in GMP. 
So we decided to execute a pilot

41
00:02:28,520 --> 00:02:32,760
that will sufficiently evaluate,
you know, the abilities of Blitz

42
00:02:32,760 --> 00:02:35,960
C in our own environment and 
with our own code. 

43
00:02:36,200 --> 00:02:41,520
Tell us about that pilot and 
what were your goals for the 

44
00:02:41,520 --> 00:02:43,960
pilot? 
The goals of the pilot were to 

45
00:02:44,000 --> 00:02:48,240
evaluate, you know, the 
capabilities of Blitz C as a as 

46
00:02:48,240 --> 00:02:53,720
a software development platform.
So what we did is we selected 

47
00:02:53,720 --> 00:02:57,840
one an existing system that that
we have real system that 

48
00:02:57,840 --> 00:03:01,560
actually had a, you know, a 
number of requirements that had 

49
00:03:01,560 --> 00:03:06,520
to be executed anyway. 
And what we decided is to test 

50
00:03:06,600 --> 00:03:10,640
different types of use cases 
using this existing system. 

51
00:03:10,640 --> 00:03:14,000
So one of them was a back end 
migration. 

52
00:03:14,000 --> 00:03:17,800
I mean the back end was written 
in an old version of Java, Java 

53
00:03:17,800 --> 00:03:23,160
8, and we needed to modernize. 
Got back into Java 21, that was 

54
00:03:23,280 --> 00:03:25,520
very clear. 
Same thing with the front end. 

55
00:03:25,520 --> 00:03:28,960
Front end was developed in a 
very old version of Angular. 

56
00:03:28,960 --> 00:03:32,640
I think it was Angular 11, I 
might be wrong and we needed to 

57
00:03:32,640 --> 00:03:34,760
upgrade it to one of the latest 
versions. 

58
00:03:34,960 --> 00:03:39,920
We also wanted to test a 
specifically use case to build a

59
00:03:39,920 --> 00:03:44,720
new feature autonomously that is
providing the right prompt to 

60
00:03:44,720 --> 00:03:49,080
the platform of you know, 
describing the new functional 

61
00:03:49,080 --> 00:03:52,240
feature in business terms that 
we wanted to build and having 

62
00:03:52,240 --> 00:03:54,680
the system built the new feature
for us. 

63
00:03:55,040 --> 00:03:59,440
And then a fourth use case where
we wanted to do the remediation 

64
00:03:59,440 --> 00:04:02,520
of a number of security 
vulnerabilities that the system 

65
00:04:02,520 --> 00:04:05,200
had. 
So we thought that, you know, 

66
00:04:05,200 --> 00:04:10,360
with that breadth of use cases, 
it will be a nice spectrum to to

67
00:04:10,520 --> 00:04:15,040
basically test the capabilities 
of the system in a real life 

68
00:04:15,040 --> 00:04:18,680
environment, which was our 
environment connected to our Git

69
00:04:18,680 --> 00:04:23,440
lab repository where all of our 
code resides and connecting it 

70
00:04:23,440 --> 00:04:26,840
to our CICD pipeline. 
Basically, that's what we wanted

71
00:04:26,840 --> 00:04:28,960
to test. 
Why didn't you just give these 

72
00:04:28,960 --> 00:04:34,840
small teams the opportunity to 
vibe code their way into this 

73
00:04:34,840 --> 00:04:37,560
modernization? 
That was perfectly possible. 

74
00:04:37,560 --> 00:04:41,240
Most of our teams, they're 
using, you know, Co pilots to 

75
00:04:41,480 --> 00:04:45,440
basically work faster and have 
the Co pilots, you know, 

76
00:04:46,000 --> 00:04:49,240
partially developer, they're 
trying to develop and test what 

77
00:04:49,240 --> 00:04:52,120
they're doing. 
But this is a different type of 

78
00:04:52,120 --> 00:04:54,840
platform. 
I mean, there's no, there's no 

79
00:04:54,840 --> 00:04:57,440
ID initially here. 
It's just a platform. 

80
00:04:57,800 --> 00:05:00,880
You provide a very detailed 
prompt of what you want to 

81
00:05:00,880 --> 00:05:05,040
achieve and then the system 
using its own internal logentic 

82
00:05:05,480 --> 00:05:10,320
architecture basically 
autonomously creates all the 

83
00:05:10,320 --> 00:05:13,080
different software changes or 
creates the new software that 

84
00:05:13,080 --> 00:05:14,960
needs to be incorporated into 
your project. 

85
00:05:15,240 --> 00:05:17,480
That's what we wanted to test, 
yes. 

86
00:05:17,480 --> 00:05:20,760
I mean, we could have given this
to a team, you know, given 

87
00:05:21,440 --> 00:05:24,800
different not only one, but you 
know, even several types of 

88
00:05:24,800 --> 00:05:26,800
copilots. 
And we're working that way 

89
00:05:26,800 --> 00:05:29,280
already. 
But we have never worked with an

90
00:05:29,280 --> 00:05:32,480
autonomous platform. 
And the idea was to explicitly 

91
00:05:32,480 --> 00:05:37,040
test a platform that has a 
different, you know, paradigm of

92
00:05:37,080 --> 00:05:40,640
usage than the rest of the byte 
coding tools. 

93
00:05:41,000 --> 00:05:46,600
So you were really looking at 
shifting the strategic value of 

94
00:05:46,600 --> 00:05:50,400
technology development? 
Right, exactly. 

95
00:05:51,000 --> 00:05:56,720
What does autonomous development
mean in practical terms at GNP 

96
00:05:56,720 --> 00:06:01,880
and and how does Blitzie fit 
along with the other tools in 

97
00:06:01,880 --> 00:06:07,360
your AI code generation stack? 
We are interested in increasing 

98
00:06:07,360 --> 00:06:10,800
agility as much as if it is 
feasible. 

99
00:06:11,000 --> 00:06:14,520
That has been our, you know, our
objective for for many years. 

100
00:06:14,520 --> 00:06:18,400
So if we're in adapting 
methodologies, incorporating 

101
00:06:18,400 --> 00:06:23,800
tools, you know, streamlining 
our CICD pipeline in order to 

102
00:06:23,800 --> 00:06:27,400
try to achieve agility, we keep 
improving, trying to make 

103
00:06:27,400 --> 00:06:31,000
changes to improve agility. 
We want to change the role of 

104
00:06:31,000 --> 00:06:33,480
humans in the software 
development process. 

105
00:06:34,200 --> 00:06:37,400
We want to do it to achieve 
agility by coding. 

106
00:06:37,640 --> 00:06:41,400
And you know, all the tools that
assist work groups are great. 

107
00:06:41,440 --> 00:06:45,320
I mean, they do provide value, 
but we wanted to test a new 

108
00:06:45,320 --> 00:06:50,280
generation of or or a new type 
of tools for generating software

109
00:06:50,680 --> 00:06:53,800
in an in an autonomous way. 
That's a value proposition of 

110
00:06:53,800 --> 00:06:55,600
bleed. 
See, it looked, sounded very 

111
00:06:55,600 --> 00:06:58,520
attractive to us and that's what
we wanted to test. 

112
00:06:58,520 --> 00:07:03,160
What's the value of that is, you
know, you have to learn a new 

113
00:07:03,160 --> 00:07:06,520
skill, which is being very good 
at prompt engineering. 

114
00:07:06,520 --> 00:07:11,760
You have to be, you have to 
create very precise and complete

115
00:07:11,880 --> 00:07:14,960
prompts for for the platform. 
But if you do, and that's the 

116
00:07:14,960 --> 00:07:18,880
skill that you know, I've seen 
that our engineers have learned 

117
00:07:18,880 --> 00:07:21,640
fast. 
Once you get the right prompt, 

118
00:07:21,920 --> 00:07:24,360
then you get the right results 
in the speed. 

119
00:07:24,360 --> 00:07:27,040
That was very impressive. 
What about enterprise 

120
00:07:27,040 --> 00:07:32,120
requirements such as security, 
governance, architectural 

121
00:07:32,120 --> 00:07:37,080
standards? 
How does this approach support 

122
00:07:37,360 --> 00:07:41,160
the corporate technology 
requirements that are needed? 

123
00:07:41,360 --> 00:07:45,840
You definitely need to be very 
specific regarding your own 

124
00:07:46,200 --> 00:07:48,680
corporate guidelines. 
We do have our own guidelines, 

125
00:07:48,680 --> 00:07:52,040
you know, technical guidelines. 
We have our own security 

126
00:07:52,040 --> 00:07:55,800
guidelines too, regarding 
characteristics that code has to

127
00:07:55,800 --> 00:08:00,040
meet and the type of test we 
want to execute on the on the 

128
00:08:00,040 --> 00:08:04,320
components that we develop. 
But what we found is that, you 

129
00:08:04,320 --> 00:08:10,120
know, those specification, those
guard rails, those are part of 

130
00:08:10,120 --> 00:08:12,680
the input that you provide to a 
platform like Blitzing. 

131
00:08:13,600 --> 00:08:17,040
So you not only provide you know
the functional specs and the 

132
00:08:17,040 --> 00:08:20,360
intention of your project with 
all the functional you know 

133
00:08:20,360 --> 00:08:24,040
specifications, you also provide
all the technical and security 

134
00:08:24,040 --> 00:08:27,480
guidelines too. 
What results have you seen 

135
00:08:27,480 --> 00:08:31,680
across development? 
Velocity, Quality of code? 

136
00:08:31,960 --> 00:08:36,480
Cost Hours saved. 
We were very pleased with the 

137
00:08:36,480 --> 00:08:41,679
results that we got. 
We roughly were able to 

138
00:08:41,720 --> 00:08:46,480
accelerated development velocity
of between 5:00 and 10X in terms

139
00:08:46,480 --> 00:08:51,480
of engineering velocity and the 
percentage of all the completed 

140
00:08:51,480 --> 00:08:56,280
work done autonomously by 
Blitzie, it ranged from between 

141
00:08:56,280 --> 00:08:59,640
80 and 95%, depending in the use
case. 

142
00:08:59,880 --> 00:09:04,840
For example, use cases that you 
basically want to upgrade from 

143
00:09:04,840 --> 00:09:07,720
an old version of Java to a new 
version of Java. 

144
00:09:07,720 --> 00:09:10,120
And you do have to provide the 
guidelines of how to do the 

145
00:09:10,120 --> 00:09:12,000
upgrade. 
It's not, it's not as simple as 

146
00:09:12,000 --> 00:09:16,440
just mentioning just upgrade it.
Just upgrade from May to 21 and 

147
00:09:16,440 --> 00:09:18,760
I'm done requesting. 
Now you do have to provide a lot

148
00:09:18,760 --> 00:09:23,520
of information on how to do it. 
But those type of projects, we, 

149
00:09:23,520 --> 00:09:28,600
we basically saw that, you know,
the, the percentage of 

150
00:09:28,720 --> 00:09:33,480
autonomous completion was close 
to 100% front end. 

151
00:09:34,120 --> 00:09:36,640
You know, modernization of front
ends is different. 

152
00:09:36,640 --> 00:09:40,600
Those, it's a little bit more 
tricky, but we got around 80% 

153
00:09:40,800 --> 00:09:44,640
but and an 80%, you know, the, 
what basically the team told me 

154
00:09:44,640 --> 00:09:48,840
is 80% is just wonderful. 
I mean, and so I mean we still 

155
00:09:48,840 --> 00:09:53,960
have to do 20% and for the 
reminder 20%, they do by coding,

156
00:09:54,280 --> 00:09:57,240
bringing their IDs, they're 
bringing their copilots and they

157
00:09:57,440 --> 00:10:01,920
finish the reminder 20%. 
And you know, overall that 

158
00:10:02,080 --> 00:10:06,080
accelerates the process a lot. 
So they're using Blitzy to do 

159
00:10:06,080 --> 00:10:11,960
the heavy lifting and then using
lighter weight tools to do the 

160
00:10:11,960 --> 00:10:16,080
fine tune polishing. 
Exactly to basically complete 

161
00:10:16,080 --> 00:10:20,280
the project. 
Enrique, how did the developer's

162
00:10:20,400 --> 00:10:27,160
jobs, roles and daily work 
change as a result of this shift

163
00:10:27,240 --> 00:10:31,800
to autonomous development? 
The role of humans changes. 

164
00:10:32,560 --> 00:10:36,320
You still need developers. 
If you're going to develop a 

165
00:10:36,320 --> 00:10:40,160
system, you need to provide 
technical guidelines, you need 

166
00:10:40,160 --> 00:10:44,640
to provide guidelines related to
the platforms where the software

167
00:10:44,640 --> 00:10:47,720
is going to run. 
And an end user cannot do that. 

168
00:10:47,840 --> 00:10:51,280
I mean, in in our case, we, we 
deploy our systems mainly in 

169
00:10:51,280 --> 00:10:53,720
Google Cloud. 
So you need to know about the 

170
00:10:54,280 --> 00:10:57,680
technical features and then and,
and the technical settings that 

171
00:10:57,680 --> 00:11:01,320
you need to enable in the in the
cloud platforms in order for the

172
00:11:01,320 --> 00:11:04,240
system to run. 
An end user is not going to do 

173
00:11:04,240 --> 00:11:06,480
that. 
So you still need a program, but

174
00:11:06,480 --> 00:11:10,320
the role changes completely. 
I mean, it's the human is not 

175
00:11:10,320 --> 00:11:14,280
writing the code. 
I mean, the human is directing a

176
00:11:14,280 --> 00:11:16,400
platform on how to write the 
code. 

177
00:11:16,760 --> 00:11:20,240
So that's, I mean, that's a 
huge, you know, changing 

178
00:11:20,240 --> 00:11:22,520
paradigm. 
How did the developers react? 

179
00:11:22,880 --> 00:11:27,680
You have to be careful with the 
change management and the human 

180
00:11:27,680 --> 00:11:30,520
resistance to change is 
something that you cannot 

181
00:11:30,520 --> 00:11:34,160
overlook in general. 
Some of them were skeptic at the

182
00:11:34,160 --> 00:11:38,440
beginning, You know, it was like
this sounds too good to be true.

183
00:11:39,000 --> 00:11:42,480
As we started work with the 
platform and getting results, 

184
00:11:43,000 --> 00:11:48,600
their attitude changed very 
rapidly and they started to get 

185
00:11:48,600 --> 00:11:51,320
interested and they started to 
see value. 

186
00:11:51,760 --> 00:11:55,000
They were intellectually 
challenged by, you know, this 

187
00:11:55,000 --> 00:11:58,080
new type of work. 
So they not only accepted it, 

188
00:11:58,080 --> 00:12:02,720
but they were very motivated in,
in testing the platform in, in 

189
00:12:03,520 --> 00:12:07,680
testing the different prompt, 
you know, techniques to try to 

190
00:12:07,680 --> 00:12:10,680
achieve better results from, 
from the platform. 

191
00:12:10,760 --> 00:12:14,960
And I think that overall the 
group that so far has worked 

192
00:12:14,960 --> 00:12:18,760
with Blitzi at GMP, they're 
they're excited about it and 

193
00:12:18,760 --> 00:12:23,040
they're very enthusiastic about,
you know, basically expanding 

194
00:12:23,040 --> 00:12:26,040
the use of Blitzi to the rest of
of our software development. 

195
00:12:26,320 --> 00:12:31,600
What advice would you give to 
other CIOs on transforming their

196
00:12:31,600 --> 00:12:35,720
engineering organization to be 
AI native in this same way? 

197
00:12:36,200 --> 00:12:39,600
You don't just flip a switch to 
full autonomy. 

198
00:12:40,040 --> 00:12:41,760
That doesn't doesn't work that 
way. 

199
00:12:42,640 --> 00:12:46,440
You have to build trust through 
a faced human in the loop 

200
00:12:46,440 --> 00:12:49,280
approach. 
You need to target the friction.

201
00:12:50,680 --> 00:12:56,560
You know, we need to deploy 
agents to 1st to handle high F4,

202
00:12:56,680 --> 00:13:01,360
low risk friction points first. 
Such as you know, this 

203
00:13:01,360 --> 00:13:04,360
modernization projects of 
basically upgrading the 

204
00:13:04,360 --> 00:13:09,360
programming languages or writing
system documentation for 

205
00:13:09,360 --> 00:13:13,040
example, or generating test 
suites. 

206
00:13:13,040 --> 00:13:15,960
This type of tasks that 
generally are low friction. 

207
00:13:16,760 --> 00:13:19,160
Then you need to shift the 
engineering mindset. 

208
00:13:19,280 --> 00:13:22,120
We have to train our engineers 
to transition from being 

209
00:13:22,120 --> 00:13:25,280
creators to editors and 
orchestrators. 

210
00:13:25,320 --> 00:13:29,120
It's a different role, and the 
human leader's job now becomes 

211
00:13:29,120 --> 00:13:34,120
defining the prompt, reviewing 
the architecture, and validating

212
00:13:34,120 --> 00:13:38,160
the AI's execution. 
You are a technologist, You're 

213
00:13:38,160 --> 00:13:43,280
also a business person. 
Can you describe the strategic 

214
00:13:43,480 --> 00:13:47,920
benefits that this approach and 
the faster development speed 

215
00:13:48,400 --> 00:13:53,880
unlocks for GNP? 
Development speed is isn't just 

216
00:13:53,880 --> 00:13:57,800
about writing code faster. 
I think it on lots different 

217
00:13:57,800 --> 00:14:03,520
strategic advantages and 1 is 
like first to market innovation.

218
00:14:03,800 --> 00:14:09,000
I mean, you know, we can design,
deploy and iterate on like new 

219
00:14:09,000 --> 00:14:12,720
insurance products in weeks 
rather than in months or 

220
00:14:12,720 --> 00:14:16,600
allowing us to capture market 
demand before our competitors 

221
00:14:16,600 --> 00:14:19,400
even react. 
That's one of the goals that we 

222
00:14:19,400 --> 00:14:21,360
have. 
The other benefit and the other 

223
00:14:21,720 --> 00:14:24,760
objective that we have as a 
company is improving customer 

224
00:14:24,760 --> 00:14:27,440
experience. 
So this speed and this agility 

225
00:14:27,440 --> 00:14:32,760
empowers us to continuously ship
digital improvements, so like 

226
00:14:33,320 --> 00:14:38,120
instant claim processing or, or 
seamless onboarding and 

227
00:14:38,200 --> 00:14:41,840
basically try to meet the, the 
high expectations of the modern 

228
00:14:41,840 --> 00:14:45,800
consumer who, who is, you know, 
basically they're more demanding

229
00:14:45,960 --> 00:14:48,200
all the time. 
It shifts our technology 

230
00:14:48,200 --> 00:14:53,160
organization from simply 
maintaining the business to 

231
00:14:53,400 --> 00:14:57,040
actively dictating the pace of 
the Mexican insurance market. 

232
00:14:57,160 --> 00:15:02,560
That's that's our goal. 
As you plan to roll this type of

233
00:15:02,560 --> 00:15:07,520
agentic product development out 
across the company, what are 

234
00:15:07,520 --> 00:15:09,400
your thoughts? 
What are your concerns? 

235
00:15:09,480 --> 00:15:11,320
What's the approach that you're 
taking? 

236
00:15:11,680 --> 00:15:16,560
We are now a process of 
expanding the use of the Blitzy 

237
00:15:16,560 --> 00:15:19,240
platform. 
We have incorporated seven 

238
00:15:19,240 --> 00:15:22,520
additional teams now that 
they're going to be trained and 

239
00:15:22,520 --> 00:15:24,920
they're going to be Twitter, 
they're going to be some hand 

240
00:15:24,920 --> 00:15:29,400
holding of working with them 
between Blitzy engineers and the

241
00:15:29,400 --> 00:15:32,560
our engineers that have already 
gone through the different 

242
00:15:32,560 --> 00:15:36,280
projects the last four months. 
So we're going to gradually 

243
00:15:36,280 --> 00:15:40,200
start expanding the use and what
we are going to measure and make

244
00:15:40,200 --> 00:15:45,520
sure that it gradually happens 
and becomes a reality is that 

245
00:15:45,560 --> 00:15:50,560
our speed of execution basically
improves at a very noticeable 

246
00:15:50,560 --> 00:15:53,600
and exponential rate. 
And gradually we will continue 

247
00:15:53,600 --> 00:15:55,400
to expand. 
I mean initially we have seven 

248
00:15:55,400 --> 00:15:57,240
groups. 
Once this groups are 

249
00:15:57,240 --> 00:16:00,520
sufficiently mature, they will 
continue working in this fashion

250
00:16:00,640 --> 00:16:02,680
and they will continue 
incorporating groups. 

251
00:16:03,000 --> 00:16:06,280
We think that in the two year 
time frame we will be able to 

252
00:16:06,280 --> 00:16:09,120
change the paradigm in in the 
company. 

253
00:16:09,520 --> 00:16:10,920
And how many developers do you 
have? 

254
00:16:11,240 --> 00:16:14,680
Currently we use in between our 
own developers and external 

255
00:16:14,680 --> 00:16:16,400
developers from software 
factories. 

256
00:16:16,400 --> 00:16:20,200
We have around 1000. 
Ideally we should not rely on 

257
00:16:20,200 --> 00:16:24,840
external developers in the short
to medium term and we will only 

258
00:16:24,840 --> 00:16:28,080
keep internal employees 
basically working with 

259
00:16:28,080 --> 00:16:31,000
platforms. 
So speed was the driver, but 

260
00:16:31,040 --> 00:16:34,840
ultimately you will also be 
reducing cost. 

261
00:16:35,280 --> 00:16:39,520
Yes, definitely. 
Any final advice on how 

262
00:16:39,920 --> 00:16:43,520
engineering leaders can get 
started or should get started 

263
00:16:43,520 --> 00:16:46,760
with autonomous development? 
You can easily pick within your 

264
00:16:46,760 --> 00:16:51,080
organization a use case. 
I mean, either a very old system

265
00:16:51,080 --> 00:16:54,760
that need to be modernized or a 
system that is giving you a lot 

266
00:16:54,760 --> 00:16:58,480
of problems, you know, during 
regular operation because this 

267
00:16:58,480 --> 00:17:02,000
has a lot of bugs or has a lot 
of security vulnerabilities. 

268
00:17:02,000 --> 00:17:06,800
And try these type of tools. 
I mean, you know, big bigger use

269
00:17:06,800 --> 00:17:10,920
case. 
Pick an existing system and just

270
00:17:10,920 --> 00:17:14,520
try it and you know, sandbox it,
do it carefully, you know, 

271
00:17:14,520 --> 00:17:18,160
select the right initial set of 
engineers to work in this 

272
00:17:18,160 --> 00:17:22,160
project to make sure that they 
will be able to to do the 

273
00:17:22,160 --> 00:17:24,560
transition and to understand 
this new paradigm. 

274
00:17:25,280 --> 00:17:29,160
Test it, see how it goes. 
And you basically elaborate from

275
00:17:29,160 --> 00:17:31,720
there. 
You just presented a very 

276
00:17:31,760 --> 00:17:36,120
practical and wise textbook on 
AI adoption. 

277
00:17:36,600 --> 00:17:38,680
I think it's just common sense, 
but thank you. 

278
00:17:39,760 --> 00:17:43,200
Enrique Ibarra, thank you so 
much for taking time to speak 

279
00:17:43,200 --> 00:17:45,520
with us today. 
No thanks to you, Michael. 

280
00:17:45,520 --> 00:17:46,600
It was an honor. 
Thank you.