00:00:01
[Music]
00:00:12
hello good morning and welcome everyone
00:00:15
uh I see you still have a few folks
00:00:17
coming in but maybe we we'll get get it
00:00:20
started uh good to see everyone good to
00:00:22
see a full room great to have uh be here
00:00:25
right after the kyot to follow of the
00:00:27
energy that we saw from stage uh today
00:00:31
we'll talk about how geni can help us
00:00:34
with the challenge the age uh old
00:00:37
challenge of application modernization
00:00:39
this is one of the themes that we see as
00:00:42
this interplay between gen for cloud and
00:00:45
Cloud for Gen right we really believe
00:00:47
that gen uh that that there is a lot of
00:00:50
synergy there that there is a symbiotic
00:00:52
relationship that gen can accelerate the
00:00:54
migration and modernization to cloud and
00:00:56
that no uh Enterprise grade use case of
00:00:59
General I can really exist without being
00:01:02
in public Cloud so so that relationship
00:01:04
is very strong today we'll talk about
00:01:06
how gen can can help on that migration
00:01:08
Moniz and modernization uh to Cloud uh
00:01:12
as a way of a quick introduction I'm
00:01:14
landre Santos I'm a senior partner with
00:01:16
McKenzie and Company I'm part of our
00:01:18
digital uh practice uh and I lead our
00:01:21
Cloud by Mackenzie cap capability
00:01:22
globally and here with me we have Aaron
00:01:25
hi uh Aon bam uh McKenzie partner um I
00:01:30
uh focus a lot of my time on cloud and I
00:01:33
sponsor our Global practice for the use
00:01:35
of gen to be able to power technology
00:01:38
modernization I'm a technologist by
00:01:40
trade and Leandra and I are part of what
00:01:43
we call the new McKenzie where we focus
00:01:45
on technical implementation and Delivery
00:01:47
at a keyboard level with our clients
00:01:49
which is probably different than you're
00:01:51
used to from uh hearing from Mackenzie
00:01:53
in the past and we're here because we're
00:01:55
really excited to show some of the work
00:01:57
we've been doing of how we can use gener
00:01:59
of AI to significantly accelerate and
00:02:02
reduce the developer burden for
00:02:04
modernizing technology
00:02:06
systems so uh with that the like of uh
00:02:10
Legacy Systems and Technology that is
00:02:13
very well known uh you know for age uh
00:02:16
the moment you write a piece of code it
00:02:18
starts to age right it may not show
00:02:20
rusting like physical infrastructure
00:02:22
does but we all know that it does rust
00:02:25
an age uh and it's a problem that is
00:02:28
very visible on the business side and
00:02:30
technology side so in the business side
00:02:32
it is lows down and limits the ability
00:02:35
to change and innovate for the business
00:02:37
to respond to Market pressures right for
00:02:39
you to to be able to offer new product
00:02:41
quickly and all that on the technology
00:02:43
side it just brings a ton of risks with
00:02:45
it right risks of resiliency risks of
00:02:49
continuity risks on the cyber space uh
00:02:52
and in several cases also it comes with
00:02:54
a larger operating cost a higher
00:02:56
operating cost than what you'd be able
00:02:57
to do uh uh more modernly right um now
00:03:02
if the problem is not new why why is it
00:03:05
still there and I think the reason why
00:03:06
it's there is because all of the
00:03:09
approaches all of the ways of
00:03:10
modernizing that have existed so far uh
00:03:13
come with a lot of challenge so in the
00:03:15
business side uh the business cases are
00:03:18
quite tricky right because there's been
00:03:21
hundreds of thousands or millions of
00:03:22
development hours that have gone into
00:03:24
the systems over the years and they're
00:03:26
kind of working so justifying that
00:03:28
business case is is is very very tricky
00:03:30
right really quantifying what are the
00:03:32
benefits what are the cost to do it is
00:03:34
not simple and the costs tend to be
00:03:36
fairly high right so that's one
00:03:38
Challenge on the business side and then
00:03:40
on the technology side some of the
00:03:42
systems don't have documentation they
00:03:44
sometimes don't even have the source
00:03:46
code available anymore certainly they
00:03:48
don't have enough resource uh to go
00:03:51
through the modernization the business
00:03:52
logic is not clear right and and and
00:03:55
also this Challenge on how much do I
00:03:57
just modernize without changing or do I
00:04:00
want to change the business I wonder and
00:04:01
how do I how do I do that at the end of
00:04:04
the day the opportunity cost of tackling
00:04:06
those those projects uh uh tend to be
00:04:09
too great and and other things get
00:04:11
prioritized and thus we we keep creating
00:04:13
technical data and accumulating
00:04:15
technical data as it goes now gen is a
00:04:18
game Cher on that equation right it's a
00:04:20
game Cher because uh it allows us to
00:04:23
automate a lot of the more labor
00:04:26
intensive and honestly tedious tasks
00:04:29
that developers had to do to modernize
00:04:32
uh to modernize an application and walk
00:04:33
through the application modernization
00:04:36
life cycle and how it helps along the
00:04:38
way uh but it does that the other thing
00:04:40
it does is it allows you to reverse
00:04:42
engineer the business logic that's there
00:04:45
uh which which then allows you to
00:04:48
understand what the system is doing and
00:04:50
and test it more effectively afterwards
00:04:52
right those are some of the unlocks that
00:04:54
we're seeing with Gen as it applies to
00:04:56
this particular use case uh Aon do you
00:04:58
want to maybe talk through how we have
00:05:00
been using this yeah so the big unlock
00:05:04
is being able to take a lot of that
00:05:07
manual labor for subject matter experts
00:05:09
within different areas of application
00:05:12
modernization process and be able to
00:05:15
codify what they do within workflows in
00:05:18
a multi-agent orchestrated approach so
00:05:21
if you think about that that labor that
00:05:23
was having to uh read code and be able
00:05:25
to understand it and document it in a
00:05:27
separate set of Labor to take that
00:05:29
document ation to be able to code it
00:05:31
another set of Labor to be able to
00:05:32
understand the data aspects it's being
00:05:34
able to orchestrate that through
00:05:36
cognitive agents operational agents
00:05:38
decision-based agents and critical
00:05:41
thinking type reasoning agents as well
00:05:43
and that orchestrated approach is where
00:05:45
the most significant gains for the
00:05:47
modernization process can actually come
00:05:49
and some of the kind of the the so what
00:05:52
of using this type of approach uh is on
00:05:54
the right hand side here which is at
00:05:56
minimal in increasing the the modern
00:05:59
modernization process by 40 to 50% and
00:06:02
reducing the cost by at least 30 to
00:06:06
40% so if you think about okay now how
00:06:09
does that break down even further what
00:06:11
are those agents you know actually doing
00:06:13
so if you think about a lot of the the
00:06:15
use of gen today you may take a piece of
00:06:18
code and and copy it into uh a prompt
00:06:21
for an llm to actually perform a
00:06:23
transformation and what you get may may
00:06:26
be fairly accurate you know how does
00:06:27
that all tie together what we found is
00:06:30
that there's not a magic button that you
00:06:32
can press that just says okay convert
00:06:35
this Legacy system is that being able to
00:06:38
uh tie in observability data and code
00:06:42
and knowledge from subject matter
00:06:44
experts and to be able to to reverse
00:06:46
engineer at into an intermediate State
00:06:49
an intermediate state that describes the
00:06:52
existing system from a abstract level
00:06:55
from a pseudo perspective that's not
00:06:57
actual code implementation allows for
00:07:00
developers product owners product
00:07:02
managers and Architects to look at the
00:07:04
system at a higher level to decide
00:07:06
should we continue doing it that way
00:07:08
should we modify should we add new
00:07:10
features should we remove some
00:07:11
capabilities and then once that is
00:07:13
shaped with a human that's actually
00:07:16
performing that analysis then use that
00:07:18
to be able to generate the target State
00:07:20
system where the last mile of human in
00:07:22
the loop actually
00:07:24
occurs so breaking that down a bit more
00:07:27
is that what type of data is is used to
00:07:30
be able to perform that analysis so as
00:07:32
Leandro said it's not just code code
00:07:35
observability data any documentation
00:07:37
that may or may not exist business
00:07:40
process and how the business process
00:07:42
interacts with a particular uh uh Legacy
00:07:45
system and then tying that together
00:07:48
through AI synthesis with subject matter
00:07:50
experts to complete the picture of okay
00:07:52
this is really what the system actually
00:07:55
does and that's often times a mystery
00:07:58
you know some systems are 20 or 30 years
00:08:00
old people have moved on and there's a
00:08:02
lot of Niche understanding that's just
00:08:05
very difficult to be able to figure out
00:08:07
so just being able to document is
00:08:09
incredibly critical once that
00:08:11
documentation exists that can then be
00:08:13
used to be able to rationalize different
00:08:15
systems many Enterprises have duplicate
00:08:18
systems they may have uh this instance
00:08:21
of this system and this instance of the
00:08:22
system but they're configured
00:08:23
differently they're for slightly
00:08:25
different cases but there's maybe an 80
00:08:26
to 90% overlap and some systems there
00:08:29
may
00:08:30
even nine or 10 different instances so
00:08:32
being able to rationalize what those
00:08:34
systems are doing at a service level
00:08:36
taking that to generate the lowest level
00:08:39
of technical specification which
00:08:41
includes how does this system fit into
00:08:44
the overall business what's the the
00:08:46
pseudo code what are the The Logical
00:08:49
flow what's the data that's shaping it
00:08:51
and most importantly the test cases how
00:08:54
do you know that a new modernized system
00:08:56
is actually going to work so you have to
00:08:58
have those test cases a part of that's
00:08:59
specification once you have all that
00:09:01
that is then used to generate the target
00:09:03
State system so not translating line per
00:09:07
line of code from whatever it was before
00:09:09
to the new but instead reverse
00:09:11
engineering it to an intermediate form
00:09:13
and then generating the new Target
00:09:18
state so maybe just uh one thing to add
00:09:21
is it's this is definitely more than a
00:09:24
tool and and and as much as we would
00:09:26
have liked for a magic button to exist
00:09:28
that's not the case right and honestly
00:09:31
that's not what most Enterprises need so
00:09:33
this is as much about the approach and
00:09:36
methodology that you do it and creating
00:09:38
that learning capability within the
00:09:40
Enterprise as is as it is about the
00:09:42
several tools that are going to come in
00:09:43
to support it now it is anchored on real
00:09:46
Innovation and breakthroughs that are
00:09:48
enabled through gen and those uh
00:09:51
breakthroughs are around the the
00:09:54
application modernization life cycle so
00:09:56
a lot upfront in terms of reverse
00:09:58
engineering those business domains the
00:10:01
the business Logic the data the data
00:10:02
structure the data relationships all of
00:10:04
that through that the point of reverse
00:10:07
engineering and creating the
00:10:09
documentation and the pseudo code uh and
00:10:12
and and the technical design which is
00:10:14
where the human in the loop really comes
00:10:16
in to be able to validate change it
00:10:19
improve it right and and act on it in a
00:10:22
way that there is there is an AB
00:10:23
instruction layer where you can work
00:10:24
through it and then finally on the last
00:10:27
stage to actually create the source code
00:10:29
not just the code itself in a modern
00:10:31
language but also the instantiation of
00:10:33
that code in a way that that can be
00:10:36
managed in adap sack Ops Type of Way in
00:10:38
a way that you are uh uh deploying using
00:10:41
infrastructure as code again back to the
00:10:43
point of this being a way of modernizing
00:10:45
to the cloud or even modernizing to
00:10:47
perhaps a software as a service type of
00:10:49
solution right but really not that line
00:10:51
by line trans translation but but really
00:10:54
thinking through how do I create a
00:10:56
modern application out of it and and we
00:10:58
talk about the the 20 non-trivial
00:11:00
engineering cases right or challenge
00:11:02
that that that we feel like they have
00:11:04
been breakthroughs and era maybe you can
00:11:05
highlight a couple of them and bring
00:11:07
this to life yeah some pretty exciting
00:11:10
uh kind of breakthroughs here are if you
00:11:12
think about highly integrated monolithic
00:11:15
systems where let's say you have you
00:11:18
know 10,000 functions across the system
00:11:21
and each function in turn calls 2,000
00:11:24
functions per so it's this massive tily
00:11:26
integrated mesh so we've actually proven
00:11:29
that that's possible using generative AI
00:11:31
to be able to reverse engineer that
00:11:33
system to able to first turn that into a
00:11:36
hierarchy of a call tree and then be
00:11:39
able to use similarity scoring to be
00:11:41
able to understand which parts of the
00:11:43
call trees are actually similar to
00:11:44
another to able to start abstracting
00:11:46
those out into reusable components to
00:11:49
start forming the basis of microservices
00:11:51
out of that monolith now we all know
00:11:53
that that's possible to do manually but
00:11:55
having the benefit of large language
00:11:57
models like Gemini and others to be able
00:11:59
to do that analysis that is very rote
00:12:02
and tedious to perform being able to
00:12:04
automate that at a very high speed
00:12:06
actually allows for that mesh of code to
00:12:09
start being pulled apart in that way uh
00:12:12
another one is a cross interface bounded
00:12:15
context synthesis what that means is
00:12:18
let's say that you have uh a call that
00:12:20
comes in from HTTP that then goes to a
00:12:23
message CU and then it goes into uh a
00:12:26
backend processing system for procedural
00:12:28
logic then it goes to a database and
00:12:30
then it goes to a database stored
00:12:32
procedure so all of that's a part of one
00:12:34
bounded context but how do you take each
00:12:36
one of those code bases each one of
00:12:38
those separate portions of the overall
00:12:41
value chain for that type of transaction
00:12:44
be able to understand it holistically
00:12:46
and then be able to document it and then
00:12:48
be able to generate that in a much more
00:12:50
simplified form potentially eliminating
00:12:53
some of the stord procedures
00:12:54
understanding the data so we've we've
00:12:56
proven that that's possible and the the
00:12:58
last is on the data side you know we we
00:13:01
find that in a lot of large Enterprises
00:13:03
that the data models that mature over 20
00:13:06
or 30 years can become incredibly
00:13:09
challenging to understand a lot of
00:13:10
duplicative data data that's misused or
00:13:13
is not marked or tagged properly
00:13:15
categorized so being able to take let's
00:13:17
say thousands or tens of thousands of
00:13:19
tables with hundreds of thousands of
00:13:20
fields being able to look at call
00:13:23
patterns being able to look at
00:13:24
observability data to figure out what
00:13:26
really should the right Enterprise data
00:13:30
model exist for the business be able to
00:13:32
suggest that and then and being able to
00:13:35
suggest that in an automated way and to
00:13:37
be able to automate the mapping of those
00:13:39
fields and then to automate the
00:13:41
Transformations that exist the existing
00:13:43
ETL that might exist within the
00:13:46
Enterprise into that model again being
00:13:48
able to use gen of AI to significantly
00:13:50
accelerate that process so these are
00:13:52
some of the come of some of the
00:13:53
individual breakthroughs that we've been
00:13:54
really excited to
00:13:57
experience so uh we're going to bring
00:13:59
this to life in a demonstration here so
00:14:02
if you think about some of the most
00:14:03
challenging Legacy systems uh a lot of
00:14:06
times they're mainframes I I think some
00:14:08
of the market estimates is that 50 to
00:14:10
60% of the planet still operates off of
00:14:12
mainframes so they still are alive and
00:14:14
well in a lot of large Enterprises and
00:14:18
so uh that was one of the first use
00:14:20
cases that we said let's go take this
00:14:22
approach and actually apply it for
00:14:24
Mainframe transformation so in this
00:14:26
demonstration we've taken some open
00:14:27
source Cobalt code and we'll go through
00:14:30
and really provide an example of what
00:14:32
this approach looks like so in an
00:14:34
integrated development environment you
00:14:35
know first we're going to see all the
00:14:36
Cobalt code and BMS maps and JCL and we
00:14:39
can actually start seeing a dashboard
00:14:40
now we're going to start taking a look
00:14:42
at what does one particular code file
00:14:44
look like well this is kind of the cobal
00:14:46
code and it's got a lot of the setup
00:14:48
with copy books but we're going to get
00:14:50
to some really meaty nested if
00:14:53
statements that may be difficult to be
00:14:54
able to understand and here again using
00:14:57
the large language model to be able to
00:14:58
synthesize that into documentation
00:15:00
here's a highlevel business overview
00:15:02
this is going to process credit
00:15:03
transactions in a badge process on a
00:15:05
daily basis and here's the logical flow
00:15:08
of what that does now the the product
00:15:11
manager may look at this and say well we
00:15:12
want to add some new air logging because
00:15:14
we want to be able to get some more data
00:15:16
back to be able to advise what we do so
00:15:18
we're going to actually modify the
00:15:20
overview we're to modify the
00:15:21
requirements for what the system should
00:15:23
do we're going to add that here and then
00:15:25
from there start generating human
00:15:27
readable test cases which are then turn
00:15:29
to code so these test cases actually
00:15:31
prove that that system is going to work
00:15:33
and now you can actually see from the
00:15:36
test cases from the documentation
00:15:38
generating the target state code so here
00:15:40
you have the core application logic
00:15:42
where we've converted from index flat
00:15:44
VSM files to SQL relational database
00:15:47
data manipulation you see the
00:15:49
application code where it's actually uh
00:15:52
uh um API endpoints that are actually
00:15:55
put in for the application code and the
00:15:57
the data models that are going to exist
00:16:00
across all the programs that exist
00:16:02
across the module and then being able to
00:16:03
see that manifest and be able to walk
00:16:05
through it now what you didn't see there
00:16:08
was a developer typing code that then
00:16:10
suggests something else to write what
00:16:12
you saw was a holistic view of the whole
00:16:15
system where the developer was involved
00:16:17
in every step of the way but the work
00:16:20
that the developer had to do was
00:16:21
significantly reduced to be able to
00:16:23
modernize it to a Target State and this
00:16:25
example was Java but being able to
00:16:27
Target other languages like python or
00:16:29
JavaScript simply a matter of changing
00:16:31
some of the prompts to be able to do
00:16:35
so right so while this is a su obviously
00:16:38
in early stage it's a technology that's
00:16:41
completely outside of the lab already so
00:16:43
we have been working with some of our
00:16:44
clients on some of their trickiest uh
00:16:47
Legacy challenge uh and and and and and
00:16:50
seeing Real Results in the order of the
00:16:52
improvements that Aon was talking about
00:16:53
at the beginning so Aon will walk us
00:16:56
through uh a case example with one of
00:16:58
the iminent uh US Banks and how they're
00:17:01
using this type of methodology to uh to
00:17:04
modernize one of their core systems and
00:17:06
and use that then as the the MVP to
00:17:08
scaled that across other areas within
00:17:10
the
00:17:11
Enterprise yeah so this case was uh
00:17:13
super exciting because it was such a
00:17:16
difficult system that the client had
00:17:18
been attempting to modernize it several
00:17:20
times uh over the course of several
00:17:22
years and just wasn't able to uh so when
00:17:25
we started working with them they
00:17:27
thought this was a great use case to
00:17:28
start with so this include both uh
00:17:31
online transaction processing through
00:17:33
kicks uh green screens as well as batch
00:17:36
processes for being able to compute a
00:17:38
lot of data there's a lot of realtime
00:17:40
data uh that's also involved uh to be
00:17:42
able to process and here you can
00:17:44
actually see the the detailed
00:17:45
architecture for the components as they
00:17:48
are produced in a modern uh Java form of
00:17:52
all the different components and how
00:17:53
they're working together you actually
00:17:54
see an angular user interface that was
00:17:57
uh automatically created to be able to
00:18:00
talk to API endpoints uh based off of
00:18:03
the kicks transactions being able to
00:18:04
interact with the batch processes so
00:18:06
this is kind of bringing this to life of
00:18:08
what it actually would look like for uh
00:18:11
an actual uh Mainframe application
00:18:12
transformation to a more modern language
00:18:14
like
00:18:16
Java now one of the big questions uh
00:18:18
that we get is okay you know this this
00:18:21
type of of you know language translation
00:18:24
has been around for a while what does
00:18:25
the code look like you know for for some
00:18:28
of those who may gone through this may
00:18:29
have heard the term uh jobal uh which is
00:18:33
Java but it just really feels and looks
00:18:35
a lot like Cobalt so how do you really
00:18:37
judge whether the code that's produced
00:18:39
is really going to be you know good
00:18:41
maintainable code so there's a a Julia
00:18:44
framework that's used to be able to to
00:18:46
measure and Benchmark the the Cod
00:18:49
production from geni but what we find in
00:18:52
addition to that is just plain old AB
00:18:55
testing with subject matter expert
00:18:57
developers to actually look at if I were
00:18:59
going to do this myself versus what
00:19:01
comes out of a large language model how
00:19:03
good is it so here you can actually see
00:19:05
scoring uh that came directly from the
00:19:08
client was this wasn't us this was
00:19:10
actually the client scoring on uh human
00:19:12
readability maintainability correctness
00:19:15
accuracy across the code uh being able
00:19:17
to use some of the more modern coding
00:19:20
standards that are actually used that uh
00:19:22
one of their actual developers would so
00:19:25
it kind of brings it to life as far as
00:19:26
what you're now uh going to actually
00:19:29
maintain through this type of
00:19:31
system and last you know this ties back
00:19:34
up to uh really the beginning is what's
00:19:37
the the the so what you know what's the
00:19:39
the value in being able to do this so uh
00:19:43
being able to reduce the cost of
00:19:45
technology modernization is really the
00:19:47
ultimate goal we all know that there's a
00:19:50
lot of of Legacy technology or technical
00:19:53
debt that may exist within our
00:19:55
corporations and our Enterprises but
00:19:57
we've never been able to make the
00:19:58
business case to actually modernize it
00:20:01
when we look at what the cost associated
00:20:03
with modernizing it the costs are
00:20:04
through the roof and they're very
00:20:06
expensive to be able to have a lot of
00:20:07
that manual labor so you can start to
00:20:09
see here that this is some of the first
00:20:12
early data around reducing the cost of
00:20:15
modernizing uh these types of systems
00:20:17
powered by
00:20:18
gen and in this case this was sufficient
00:20:21
to change that tricky business case that
00:20:23
we talked up front uh and make it more
00:20:25
clearly Roi positive right especially as
00:20:28
you take into account back to my point
00:20:30
on J for cloud and Cloud for J if you
00:20:32
take into account that once you
00:20:34
modernize this application now you're
00:20:35
also able to more easily building new
00:20:38
features that for example are using gen
00:20:40
for agents and whatnot like we've seen
00:20:42
on the keynote earlier today so once you
00:20:44
start accounting for additional upside
00:20:47
uh from from getting into a modern
00:20:49
system as well as a much uh reduced cost
00:20:53
to get there all of a sudden some of the
00:20:56
modernization case that didn't make
00:20:57
sense before start making sense now
00:21:00
right so that's one point I think the
00:21:02
other here is that uh we see the Gen
00:21:05
assisted engineering as way uh as as as
00:21:08
the next wave of significant
00:21:10
transformation in terms of software
00:21:11
development right if we think about just
00:21:13
some of the recent Innovations from how
00:21:16
we're using agil to kind of break down
00:21:18
development Cycles from a waterfall to a
00:21:20
more uh uh iterative uh in interactive
00:21:23
way with the business for example uh how
00:21:26
we've used public Cloud to automate C CD
00:21:28
pipelines and and and and uh an automate
00:21:32
provisioning of applications scaling up
00:21:34
and down and whatnot how we have uh how
00:21:37
that then the combination of those two
00:21:39
things have enabled us to uh Institute
00:21:41
that sack Ops uh uh methods within our
00:21:43
development teams and make not just
00:21:45
development but the uh operate uh the
00:21:48
operations as well as the maintenance of
00:21:50
this those applications much simpler we
00:21:52
see gen now coming on top of all of
00:21:54
those Innovations so if you look at it
00:21:57
as a methodology as as as an operating
00:21:59
model transformation more so than a tool
00:22:02
that's going to assist an individual
00:22:04
developer to write an individual file uh
00:22:07
this becomes truly transformative right
00:22:09
and that's where that that is what we're
00:22:11
seeing uh for example on the on the case
00:22:13
that Aon just walked you through is the
00:22:15
power of doing all of that in a combined
00:22:17
way right so uh if we go to the next
00:22:21
page or do you want to add anything here
00:22:22
yeah I think uh when I recall one of my
00:22:25
favorite conversations with with a CIO
00:22:27
to to Leandra point about this being
00:22:29
really more than just tooling uh that
00:22:31
tooling is going to you know save the
00:22:33
software development process you we used
00:22:34
to develop software without agile we
00:22:37
discovered agile and now we can't really
00:22:39
Envision developing without agile and
00:22:40
then we used to develop without Dev
00:22:42
secops and then we discovered Dev seov
00:22:44
so I think that that the use of gen as a
00:22:47
part of the software development process
00:22:49
is going to become quite prominent and
00:22:51
as one CIO said it that he said I don't
00:22:54
want one tool to be able to solve one
00:22:56
problem I have a hundred problems I need
00:22:58
to be the ability to actually solve this
00:23:00
across my entire software engineering
00:23:03
process and all my software development
00:23:05
capabilities and uh the developers that
00:23:08
exist so I think that's the trend of
00:23:10
where we're going
00:23:11
here now like I said this is already
00:23:14
outside of the the lab and the timing is
00:23:16
now uh so our recommendation is that
00:23:19
obviously the the the the use case the
00:23:22
problem the specific challenges are
00:23:24
going to be unique to each Enterprise
00:23:26
right so so the first thing that we
00:23:28
encourage you all to do and we've been
00:23:30
encouraging our clients to do is to go
00:23:32
back to those modernization use case
00:23:34
those thorny issues that you know you
00:23:36
have and you've had for a while that
00:23:38
just didn't quite make the cut in the
00:23:40
past from a cost benefit standpoint and
00:23:43
let's take a fresher look at that and
00:23:46
now assuming again additional upside and
00:23:48
assuming a more streamlined way to go
00:23:50
about it like do some of those at least
00:23:52
to start making more sense once you have
00:23:55
that and you really understand that
00:23:56
there is value right and and we start
00:23:58
certain about that that there is also
00:23:59
value on the business side not just on
00:24:01
the technology side because this will
00:24:02
ultimately be a change management
00:24:04
challenge right once you validate that
00:24:07
then the next step is to create a PC or
00:24:08
even better an MVP so pick one of those
00:24:11
or a subset of them uh and and and and
00:24:14
test it and test it not just to test the
00:24:16
tools or to validate that it can work
00:24:18
from your source code to to a more
00:24:20
modern source code but tested so that
00:24:23
you can actually develop your own
00:24:24
methodology your own approach on how
00:24:26
you're going to do that within the
00:24:27
Enterprise and ultimately test and learn
00:24:30
and stand up the factory that's going to
00:24:32
go through this process going forward
00:24:33
right that's really what that uh that PC
00:24:36
or the MVP sta uh should achieve and
00:24:39
then from there it's about scaling right
00:24:41
scaling to other domains uh scaling to
00:24:43
other technolog that you have uh and
00:24:46
then just sort of accelerating ad option
00:24:47
we've seen this model where you know the
00:24:50
initial team that works through it can
00:24:51
then be split and helping in other areas
00:24:53
and all of a sudden uh as you improve
00:24:56
your your own efficiency on doing this
00:24:58
obviously as the technology continues to
00:25:00
M mature as we've been seeing right so
00:25:02
rapidly more and more of those business
00:25:04
case that didn't make sense before will
00:25:06
start making sense now right so if you
00:25:08
get into this journey it becomes a a a
00:25:11
bit of a Snowball Effect right and and
00:25:14
as you get better as the technology get
00:25:15
gets better more and more can be done
00:25:17
with it so again I think the time to
00:25:19
start is
00:25:21
[Music]
00:25:25
now
