1
00:00:14,750 --> 00:00:17,870
Hello and welcome to my podcast 
show Tech Talk. 

2
00:00:17,990 --> 00:00:20,990
In the April I had a chance to 
go to AWS Summit Health in 

3
00:00:20,990 --> 00:00:23,270
London. 
It was first of my kind of 

4
00:00:23,270 --> 00:00:26,070
experiences. 
I mostly work in Azure, but I 

5
00:00:26,070 --> 00:00:29,710
wanted to explore what kind of 
services they are providing and 

6
00:00:29,710 --> 00:00:32,350
what are new things they are 
demonstrating there. 

7
00:00:32,390 --> 00:00:37,120
They even start at 8:00 AM. 
But your keynote starts at 10:00

8
00:00:37,120 --> 00:00:39,800
AM. 
But before that two overs, I had

9
00:00:39,800 --> 00:00:42,800
a good chance to explore 
different kind of areas, what 

10
00:00:42,800 --> 00:00:44,840
they are offering different kind
of camps. 

11
00:00:44,840 --> 00:00:47,000
Enjoy free coffee and breakfast 
as well. 

12
00:00:47,200 --> 00:00:50,760
It was good to see those things 
that most interesting thing, 

13
00:00:50,760 --> 00:00:54,480
which I want to cover in today's
podcast is one of their demos 

14
00:00:54,480 --> 00:00:59,400
was around amplifying how they 
can transform and create a end 

15
00:00:59,400 --> 00:01:01,800
to end application using just a 
Gen. 

16
00:01:01,880 --> 00:01:04,879
AI. 
What it does like it hosting 

17
00:01:04,879 --> 00:01:09,000
your data authentication and 
even give you an instant 

18
00:01:09,000 --> 00:01:11,800
previews when you're working in 
AWS. 

19
00:01:11,800 --> 00:01:15,320
Simplify is a kind of tool which
allows you to build your end to 

20
00:01:15,320 --> 00:01:17,480
end applications from start to 
finish. 

21
00:01:17,640 --> 00:01:21,440
It tells you're mostly in React 
and then it connects your back 

22
00:01:21,440 --> 00:01:24,800
end and your middleware as well.
So it's a, it's a good way to 

23
00:01:24,800 --> 00:01:27,760
get into the market if you don't
have, if you don't want to spend

24
00:01:27,760 --> 00:01:30,640
much time in building these 
components from scratch. 

25
00:01:31,040 --> 00:01:34,600
What it does provide use ready 
to use components, which you can

26
00:01:34,600 --> 00:01:39,240
utilize it from the GUI and then
it can help you out and build 

27
00:01:39,240 --> 00:01:42,680
this end to end application. 
It's a TypeScript first approach

28
00:01:42,920 --> 00:01:45,560
where you can focus on 
application code, node 

29
00:01:45,560 --> 00:01:48,840
infrastructure, for example. 
And then you can, you can have 

30
00:01:48,840 --> 00:01:51,320
your sandbox environment set up 
hopefully. 

31
00:01:51,360 --> 00:01:55,000
And then it's easy for 
developers to work on on a 

32
00:01:55,000 --> 00:01:57,720
product in their own sandbox 
environment and have an 

33
00:01:57,720 --> 00:02:00,240
isolation environment where they
can test it. 

34
00:02:00,440 --> 00:02:03,840
It's a quick way of building an 
application where you still have

35
00:02:03,840 --> 00:02:07,080
an opportunity to work on the 
code side, but it does all by 

36
00:02:07,120 --> 00:02:09,880
itself using the chain AI. 
Then another thing which I 

37
00:02:10,199 --> 00:02:14,200
attended was like a WSQ code 
transform which helps you 

38
00:02:14,200 --> 00:02:18,240
transform your legacy code into 
the new code and what it does 

39
00:02:18,240 --> 00:02:20,960
like it analyze and creates the 
plan for you. 

40
00:02:20,960 --> 00:02:23,920
So for example, you have an 
application in.net or Java which

41
00:02:23,920 --> 00:02:27,520
is outdated and you want to 
transform into that upgraded 

42
00:02:27,520 --> 00:02:30,080
framework. 
So it does it by solve by 

43
00:02:30,240 --> 00:02:32,960
creating a plan for you. 
The good thing is like whatever 

44
00:02:32,960 --> 00:02:36,840
the changes, it's such as you, 
it actually gives you a manual 

45
00:02:36,920 --> 00:02:40,120
human intervention. 
So for example, if the plan 

46
00:02:40,120 --> 00:02:43,480
doesn't work or if you have some
kind of limitation, it does let 

47
00:02:43,480 --> 00:02:47,000
you manually intervene those 
kind of process to actually make

48
00:02:47,000 --> 00:02:50,040
like informed decisions. 
But it's a good way of like 

49
00:02:50,040 --> 00:02:52,960
doing the manual rather than 
doing the manual work, it just 

50
00:02:52,960 --> 00:02:56,680
automate it and just deploys the
full application into a new 

51
00:02:56,680 --> 00:02:59,120
Facebook and then it it 
generates the. 

52
00:02:59,520 --> 00:03:03,240
The good thing is like if you 
have like multiple depositories,

53
00:03:03,240 --> 00:03:07,960
let's say report 1234 sitting in
GitHub, you can actually do the 

54
00:03:07,960 --> 00:03:11,600
interaction with it give you a 
kind of understanding where you 

55
00:03:11,600 --> 00:03:15,120
can transform several 
repositories at once. 

56
00:03:15,400 --> 00:03:18,920
And it can create a plan for 
each repositories according to 

57
00:03:18,960 --> 00:03:22,520
whatever the code is present in 
those repositories, either Java 

58
00:03:22,520 --> 00:03:26,600
or in a similar kind of 
approach, Microsoft got like a 

59
00:03:26,600 --> 00:03:29,880
GitHub Copilot app 
modernization, which is actually

60
00:03:29,880 --> 00:03:33,000
in the early stage of testing. 
But Microsoft does have Azure 

61
00:03:33,000 --> 00:03:36,720
micro application and code 
assessment for upgrading technic

62
00:03:36,720 --> 00:03:38,760
application. 
What I found it is like this 

63
00:03:38,760 --> 00:03:42,160
tool is quite impressive. 
It can handle the things very 

64
00:03:42,160 --> 00:03:44,800
well. 
And in, in terms of maturity, 

65
00:03:44,800 --> 00:03:48,960
this tool, Amazon Q Core 
transform is, is better than 

66
00:03:48,960 --> 00:03:52,680
what I found it with as compared
to GitHub Copilot, which is 

67
00:03:52,680 --> 00:03:56,720
again in the early stage of 
modernization and the existing 

68
00:03:56,720 --> 00:03:59,240
tool, which is Microsoft micro 
application. 

69
00:03:59,240 --> 00:04:03,600
I did tried Amazon Q together 
per visual core to see like how 

70
00:04:03,600 --> 00:04:06,600
does it work? 
And I'm very impressed by how it

71
00:04:06,600 --> 00:04:09,360
actually works. 
In reality, it's not about the 

72
00:04:09,360 --> 00:04:11,920
demo, which I've seen it there. 
But in reality, if you want to 

73
00:04:11,920 --> 00:04:16,040
give it a go up with Amazon Q 
code transform to test like any 

74
00:04:16,079 --> 00:04:18,680
legacy code to migrate to the 
new framework. 

75
00:04:18,800 --> 00:04:21,000
Or you can use the intelligent 
Gen. 

76
00:04:21,000 --> 00:04:25,080
EI agent to actually give you 
suggestions on your, how you 

77
00:04:25,720 --> 00:04:31,920
actually works in there. 
Another thing which I've really 

78
00:04:32,240 --> 00:04:35,760
looked at is like how they're 
using AI to, to, to automate 

79
00:04:35,760 --> 00:04:38,840
some of the task. 
So one of the thing which I had 

80
00:04:38,840 --> 00:04:42,760
a demo about like how you can 
choose AI to do kind of like 

81
00:04:42,840 --> 00:04:46,680
changes in your complaint system
in financial institution. 

82
00:04:46,680 --> 00:04:51,480
So they showed us like a bit 
good kind of like an agent 

83
00:04:51,480 --> 00:04:55,640
application which actually works
on by automating your complaint 

84
00:04:55,640 --> 00:04:57,760
system. 
So for example, they had five 

85
00:04:57,760 --> 00:04:59,920
agents. 
So one is customer advocate 

86
00:04:59,920 --> 00:05:03,120
agent, which actually speaks on 
the behalf of the customer. 

87
00:05:03,120 --> 00:05:06,200
So if you receive a complaint 
from the customer into your 

88
00:05:06,360 --> 00:05:08,920
complaints management. 
So this customer advocate 

89
00:05:08,960 --> 00:05:11,480
teacher actually assess the 
complaint from the customer 

90
00:05:11,480 --> 00:05:13,440
point of view. 
Then you have a business agent 

91
00:05:13,440 --> 00:05:17,360
which actually looks at what 
your current business is and 

92
00:05:17,360 --> 00:05:20,920
gathers all the business 
policies around it that you have

93
00:05:20,920 --> 00:05:23,840
Internet policy agent, which 
actually looks like what's your 

94
00:05:23,840 --> 00:05:26,240
policy in terms of like 
resolving one of these 

95
00:05:26,240 --> 00:05:28,320
complaints and what are the 
action could be taken. 

96
00:05:28,360 --> 00:05:31,320
And then there was like FCA 
compliance agent which actually 

97
00:05:31,320 --> 00:05:35,240
looks at the FCA documentation 
and gives you suggestions from 

98
00:05:35,360 --> 00:05:38,240
from that point of view. 
And lastly, we have a judge 

99
00:05:38,320 --> 00:05:42,680
agent which actually get us all 
the results from all these four 

100
00:05:42,680 --> 00:05:45,200
regions and then compile them 
and give you the suggestion. 

101
00:05:45,200 --> 00:05:49,080
So they outwork automatically. 
So each agent collects their job

102
00:05:49,080 --> 00:05:52,440
data in on their own behalf and 
compile the record. 

103
00:05:52,520 --> 00:05:55,200
And it takes like a few minutes 
to actually get to this. 

104
00:05:55,440 --> 00:05:59,160
And at the end of the day, the 
judge agent actually used those 

105
00:05:59,160 --> 00:06:02,800
records to create a case and 
give you what should be done for

106
00:06:02,800 --> 00:06:05,240
that specific complete again the
customer. 

107
00:06:05,280 --> 00:06:08,120
And before we sent this thing to
the customer, that human 

108
00:06:08,120 --> 00:06:11,840
intervention can be done to 
actually feed more data or maybe

109
00:06:11,960 --> 00:06:15,200
change the responses before we 
actually sent to the customer. 

110
00:06:15,280 --> 00:06:18,960
So that's was a good way of 
demonstrating how AI can 

111
00:06:19,080 --> 00:06:22,520
automate your task, especially 
in the complaints handling 

112
00:06:22,760 --> 00:06:27,040
management and how it can speed 
up the process of like resolving

113
00:06:27,040 --> 00:06:32,200
customer queries and basically 
using the agents model in and, 

114
00:06:32,320 --> 00:06:37,280
and, and helping you out with 
quick resolving those customers 

115
00:06:37,280 --> 00:06:41,400
to most of the conference was 
around AI, how the AI can be 

116
00:06:41,400 --> 00:06:46,720
utilized to speed up the work, 
automate some of the manual work

117
00:06:47,040 --> 00:06:50,640
and how you can build these 
agents to actually help you out 

118
00:06:50,960 --> 00:06:54,360
to be more productive. 
So I think it was a good flavor 

119
00:06:54,640 --> 00:06:57,360
of AWS. 
If I compare it with Azure, they

120
00:06:57,360 --> 00:06:59,520
both have like identical 
services. 

121
00:06:59,520 --> 00:07:02,920
They both are doing good in 
their cloud platforms and they 

122
00:07:02,920 --> 00:07:06,720
have both are providing like 
hundreds of services which can 

123
00:07:06,720 --> 00:07:09,960
help us transform your business,
help you build ancient 

124
00:07:10,160 --> 00:07:16,800
capability in your applications.
And they also in WS they are 

125
00:07:16,800 --> 00:07:19,320
building everything is similar 
to what you are building in 

126
00:07:19,400 --> 00:07:23,320
Azure as well. 
And no code probably in terms of

127
00:07:23,320 --> 00:07:25,840
like putting those agents, you 
just need to to train those 

128
00:07:25,840 --> 00:07:29,240
agents your models and then you 
are good to go and use them into

129
00:07:29,240 --> 00:07:31,720
your application. 
Overall, it was a good 

130
00:07:31,720 --> 00:07:34,400
experience at AWS London 
conference. 

131
00:07:34,640 --> 00:07:37,120
Thank you for listening me to 
Nick in today. 

132
00:07:37,120 --> 00:07:40,720
I will come back soon with 
another actually. 

133
00:07:40,720 --> 00:07:42,440
So tell them see you.