Inside Komatsu's AI-Powered ERP Transformation: Clean Core, Joule, and SAP S/4HANA Innovation
28m 22s
Yogesh Chawaini, SAP Senior Solution Architect at Komatsu, discusses the company’s Greenfield SAP ERP transformation and its concurrent AI adoption journey. The project, initiated under SAP RISE, provided AI credits that enabled the integration of Joule for Developers and Joule for Business. Komatsu’s goal is to consolidate multiple SAP and non-SAP ERPs into a single system, making it future-ready for innovation. Joule for Developers has been instrumental, acting as a catalyst that saves 30-40% of developer time by generating code, CDS views, and Fiori applications, allowing developers to focus on design and governance. Joule for Business, though not yet in production, promises conversational AI to reduce clicks and training overhead for end users.
A key focus is maintaining a "clean core" through a governance framework based on SAP’s Level A-D categorization. Komatsu infuses Joule into daily design reviews to ensure customizations align with clean core principles, achieving 52% Level A compliance. The "classic enhancement strategy" supports this by using reusable, generic code for customizations, keeping SAP standard code intact and simplifying upgrades. BTP plays a vital role, with tools like Business Application Studio, middleware integration, and SAP Analytics Cloud. Use cases include adding custom fields to product master and developing Fiori apps for production orders. With over 30 developers and 10 business users already engaged, Komatsu plans phased rollouts, leveraging these strategies to ensure seamless future expansion.
From the Americas SAP Customer Community, I'm Jim Lichtenwalter, and this is a sub-talks.
A podcast about conversations with the innovators, leaders, and change makers shaping the future
of enterprise technology and the SAP ecosystem.
This week, I am joined by Yogesh Chawaini, SAP Senior Solution Architect at Kamatsu, during
its Greenfield SAP ERP Transformation project, the manufacturing giant also kicked off its
own AI journey, leveraging SAP AI solutions, including dual-for-developers and dual-for-business.
On this episode of the podcast, we dive into how Kamatsu leveraged these solutions to save
significant developer time and the ways it managed its government strategy, along with some
specific SAP AI use cases.
All that and more this week on a sub-talks.
From customer success stories to breaking news, first five keeps you up to date on what's
what in the SAP ecosystem.
Yogesh, thank you for joining me on a sub-talks today, nice to finally meet you in person,
I'm looking forward to diving in to Kamatsu's journey so far with us for.
Sure, thank you so much for giving this opportunity to me as well as Kamatsu to represent and share
our experience.
Happy to do it.
Why don't you go ahead and give our listeners a brief overview of what this transformation
you all are doing?
Yeah, certainly, yes.
See, transformation is not a straightforward journey and it's not just about technology.
It takes lots of brainstorming about the governance, about the processes, about the decisions
that you know, dry and then it comes with the technology and if you look at the Kamatsu
and Kamatsu's business, nature of business is complex.
We have two set of giant industrial vehicles standing as a display in the Kamatsu office
and working so whenever I take a look at that, it reminds me of a movie, "Avator" and
"Pandor of Planet" and that kind of giant vehicles we have here is not a science fiction
reality and that's the manufacturing capacity of Kamatsu and so as our ERP business is also
that interesting, it's complex, we just don't have SAP as an ERP, we have non-SAP also.
So that has the complexity and after reviewing all these, we have allotted many options and
the fair decision that we got to go ahead with the green film implementation.
Okay.
So this is about our transformation.
Excellent.
First of all, Avatar, great film.
What did you think about the newest one?
News one.
Fire and ash.
Yeah.
That's so amazing.
Yeah.
So I will see.
Deep down in the sea.
All my friends are like, "Why do you still like these movies?"
I'm like, "Because they're amazing.
What are you talking about?"
I appreciate that.
It's nice to fit an Avatar and a Pandora reference and do a podcast about ERP.
So thank you for giving me the opportunity to do that.
Why don't you walk us through then Kamatsu's stated goals for this project?
What were you all hoping to get and start to do from the very beginning?
Yeah.
So as we decided to go ahead with the green film, so obviously, as we have many non-SAP ERPs
as well as SAP ERP, we wanted to go for a single system to have a system consolidation
as a part of S.O. journey.
We wanted to utilize all S.O. capabilities to transform our business operations and to
bring it in a shape so that our systems are future innovation ready.
So that's a good.
Perfect timing.
Future innovation ready.
Yeah.
As we sit here, that's a nice lead into my next question.
This has been, I know if you are aware, AI is pretty popular here at SAP Sapphire this
week.
It has been the last few sapphires.
SAP is working very hard to make it so that their customers can adopt and embrace AI
easier.
You all have been on an AI journey at Khmatsu.
Why don't you walk us through that journey's been like, specifically tell us how it started
and how it's developed over the years.
Yeah.
Sure.
So this journey started last year when we signed contract with SAP RISE and as a part of
SAP RISE, RISE with SAP, that contract we got AI credits.
And that's the starting point of our AI journey.
There, I really appreciate the help and support that came from AI team from SAP's AI team.
They brought their approaches, hey, you have AI credits and these are the tools which
you can plug in your system and start utilizing as part of our credits.
See, that team just not just provided a guideline, but they also hand hold.
They had a working session with us and they ensure jewel for developer and jewel for
business, a plug in our ecosystem, our ecosystem of SAP RISE.
We perform many steps on BTP, on HANA Cloud, on Fiori.
We also establish the role security matrix and that way we plug.
And this way, we have already started utilizing jewel for developer as well as jewel for business.
Jewel for business is still under in development and quality stage is not moved to production,
but we are setting the stage for jewel for business and jewel for developer, we are
effectively utilizing it.
So SAP is very hands on with you all when you all begin embracing jewel for business and
jewel for developers.
Yes, that's good to hear.
When you think about CMATSU and its operations, again, an iconic global organization, why
were jewel and SAP business AI solutions critical to innovation in your IT infrastructure?
My perspective looking at AI or any AI app is kind of a machine, a smart machine, anything
that makes your life easy is a machine.
So jewel for developer is playing a very good role here.
So what we did is we set up a clean core governance along with the jewel utilization and
that's helping us as a time saver.
It's acting as a catalyst.
If you know the definition of chemical definition, catalyst is something which increases the rate
of chemical reaction, right?
So it's giving the efficiency 30 to 40% of time saver is generating the code, it's also
helping to develop came to an app right from CDS view, audit service and fury.
So that level of help we are getting as a oscillator throughout the jewel and certainly
yes, it's a it's a time saver, it's a good oscillator, I would say.
This is the way it's helping for the developers and jewel for business.
So that's another tool that we got.
It's very impressive when we demonstrate toward ebpuse and business, it's increasing first
side.
Okay.
Yeah, it has very good capabilities like reduced number of clicks, certainly.
That's the first capability it has.
Other capabilities conversational AI capabilities you need not to write a kind of a command or
code, even for jewel for developer, you need not to write a code or command to get a code.
It just your human interaction language, natural language, you got to talk to Jew and you
will get the answer and similarly for the business, they're getting fee of their business
cases, live business cases as an example that we can fulfill through jewel.
So jewel for business is certainly a very good tool that's going to help the users as
a, I would say, a facilitator.
So a commodity being a company that is, again, large or large global presence, I'm sure
has some organizational specific hurdles and probably also some region specific hurdles.
How has this transformation and how has jewel, how has jewel addressed some of the organizational
specific challenges that come on through faces on a day to day?
As a comma, so we have in house technology team limited candidates.
So jewel for developer that's pretty much helping us effectively.
So we can utilize, we are utilizing a jewel for developer as a time saver and we are able
to focus more productively on other stuff like design, review of the design, forming the
governance and then suddenly driving the offshore team of our partners.
So that's why jewel for developer is certainly helping us at this organizational level to
her to make this initial hurdle because you need to come out of developer shell, certainly
our developers and take a next step ahead to spend more time on design based practices
and governance.
Okay.
A critical component of leveraging AI is ensuring the organization has what SAP is called a
clean core, how did you all go about cleaning your core?
What solutions did you all use?
Yes.
So what happened is when we came across a clean core as a journey, we came across SAP's
note.
So that knows talk about level, clean core level, level A, level B, level C and level D.
So the level A is something golden standard for
it's at most quality for a clean core. It's kind of a client compliant level. Then comes
as level B. It's a it's upgrade friendly as per SAP categories level B because it's not
cloud compliant even though there are SAP provision or good standards to do your customization.
Then comes as level C that is we market as orange because it's a kind of a medium risk
and then it comes D that is high risk. But through this categorization we highly fill
the necessity of setting a funnel to filter out the screen of level A, B and C. And that
way what we did is we set up a governance and that governance we are reviewing in every
design review. So whenever your function specification comes on your way for the development
and what we do we take a look at a governance and C or this business specific customization
can be implemented through which level A level B or level A plus B and then last resource
goes to B C and D. So this governance is helping us and other than that to handle the
risk of C and D and even B we had our own classic announcement strategy. So it's very
interesting strategy I will talk about it later. But it's very interesting strategy that
helps you to keep your system future upgrade ready.
Okay. Can you go a little bit deeper into how you specifically approached clean core
governance. Yeah. So as I discussed just about just for a short while ago we set up that
governance and that governance is a part of our design review, daily design review.
And there the best thing that we did is we infuse utilization of jewel for developer.
So J4D is playing a crucial role there. So there are many requirements that comes on
our way. We talk to jewel, we interact with jewel, we fear the approach as A or B A plus
B. There certainly it's helping jewel as that we infuse in the governance. And that's
not just our commas would be but it's also helping the developers of our partners. So combination
of this jewel for developer and classic announcement strategy that's certainly helping us to set
up the proper clean core governance.
In a world of custom code system upgrades and constant pressure to keep the core clean
SAP developers are being asked to do more faster to modernize applications to improve performance
and prepare for what's next. That's where a sub tech connect comes in join fellow developers
for deep technical sessions hands on labs and real solutions you can take back to your
organization and use on your systems. If you build extender support SAP this is the conference
or work gets done a sub tech connect registration is open now click the link in the description
to learn more.
How would you say that adopting a clean core set the stage for your eventual usage of AI
technologies like jewel can you articulate that a little bit and just talk about through
commodities experience yes so what we notice is more you go towards clean you go green green
for your air adoption let's customization that you have in your system your system will
be more adaptable for future a nids not just a clean core we also adopting APR driven architecture
so this clean core and APR driven combination is helping our system future upgraded we
can have faster upgrade we can have lower total cost of ownership for the upgrades.
This clean core governance along with the combination of jupe or developer and classes
management strategy is suddenly keeping our system to do up to that level. We have cloud
lm also and from the cloud lm we have a one tool in a system view that gives the overview
about your clean clean core achievement. From there we capture the data and we plot
the data for the development that we completed so far so there we have seen we have 52%
of a and a plus b which is great achievement and then 42% b and then rest 6% 5 to 6% as
high risk c and d okay so this way we set up our clean core governance with enablement
of jupe or developer and classic announcement strategy okay what we chatted before this
podcast planning sort of the crux the conversation you're telling me about how spptp was critical
to this transformation can you go into a little bit more detail about how come us is leveraging
the business technology platform yeah so as a part of spptp we also got btp ea credits
and btp so throughout this what we're doing is we are effectively realizing btp business
technology platform where utilizing certainly business applications studio for or custom
theory developments where utilizing a middleware integration switch that's also very effective
and very I would say helpful for seamless integration implementation okay then next tool
that we're utilizing mbc that's for bank integration then mdi master data integration
and again there we have sac spp analytics cloud that we are highly utilizing for our reports
so that's another good utilization of btp we are doing and certainly for jul btp is a starting
point because we have billburg zone there we need to set up the jules of account and security
framework and also for security we are releasing ies and ieg for took go on btp security
are you moving sapps for hana are you trying integrate systems that were never designed
to talk struggling to keep master data clean with budgets getting tighter and tighter if
any of this sounds familiar you're not alone in fact most sap customers are facing the
same challenges you are right now that's why you should attend an asub chapter meeting
they connect you with peers who have already navigated sap s for hana migrations untangled
integrations and found practical ways to move forward these are real experiences shared
openly because sap can be hard and you don't have to figure it all out alone find your local
a sub chapter by clicking the link in the description when we think about jule what would
you say some of the immediate benefits a commodity realize specific to jule for business jule
for business air as i said earlier it's conversational air capabilities wonderful okay and
whenever user login to the system the login to fury they need not to keep on searching
a specific fury ties for even though even though it is limited today for five to six business
lines but they got to they can simply start the interaction with this jule let's say show
me says order for specific sold to this is one of the use case from our koma so another is show
me purchase order pending for approvals show me geol account and all open and closing balances
right so all these kind of a use this examples are certainly helping our business to
gate the details on fingertips reduce number of clicks and suddenly it will gonna reduce the
train that we got to training overhead I would say is certainly gonna reduce okay on our user
friend how many end users do you all have using a city AI tools so from jule for developer
perspective you say there are more than 30 developers and for jule for business there are around
10 users but certainly that user count gonna grow as we move ahead to you Eddie and
production you guys we we talked about your usage of jule and i love this sort of dig a little
bit deeper and love for you the teller listeners about a use case of how you all and your end users
are using jule in kamaatsu's operations yes the best example here i can talk about
we were in the initial phase of implementing the clean core we thought about the clean core governance
we started following the clean core governance with the help of jule for developer
there was a specific business requirement that came over way we wanted to add a field product group
or on SAP standard theory app rock master and this product group along with the custom searcher
So we want to do that.
Now the beauty is
When we requested jewel for developer here, right, then it gave us step by step guide.
How to go with the clean coding implementation of this business requirement?
We follow the step.
So very first step is about we got to go to custom field logic app.
There we got to find the business context and business context for this product master.
Then add your custom field all there.
When you define a type for this custom field, it has to be code list based on CDS view.
And then it also provided the CDS view for your search app, your custom search that you're
looking for.
Then through this configuration, when we hit OK, it generated the O data services, it generated
the classes, it generated the binding and then we got this field on our product master.
So this is the example of key user, so this is the example of developer extensibility.
Now key user extensibility, we got to go to the product master field app and just do
add of UI.
OK, the moment we do add of UI, this will appear on that SAP standard field app and that
to be the search is so quick, so seamless.
If you imagine you go to do this customization through classical way, it takes lot of time,
you go to lots of complexity, but with this very simplified step, it's helping us.
Other app we developed for our production order.
So if you notice production order, there is a Feury app on Feury app library, but that's
GUI app.
We wanted to implement a Feury UI app, we asked you, "Jule provided us, CDS view, is
provided us API, O data service, and it also provided the model view controller code
for UI.
And that help us to build is app very quickly.
So these are the use cases from the Jule for developer, that's exciting.
And one example I would like to tell you, how to deal with the Jule for developer.
Who is?
Yeah.
So it's a kind of your colleague, extended helping hand, you can suddenly get from him,
but he's kind of your be specific friend, be specific colleague.
So last year in Sapphire, I was here and one of my friends was standing in queue, I called
him, he said he's standing in a queue to get a cup of coffee from Barista Bar.
Okay, asking, "Oh, okay, please get me for, get me also one cup of coffee."
So kind of him, "Hey, get me a cup of coffee, we had coffee cheers, and then I enjoyed
a first sip of sip of coffee."
Then I realized, "Oh, it's full of sugar."
I asked my friend, "Hey, you know, I don't need sugar, and you get me a coffee with
a sugar."
He replied, "But I didn't know at this moment you need a coffee with a sugar or without sugar."
You got to be be specific.
Exactly the same way, whenever you interact with the Jule for developer, you got to be
specific.
The specific, "Hey, act as a web developer, act as a solution architect or do not hallucinate."
And then put your requirement, and also you can put your requirement from standard specific,
do not give me a serious view or association with a join on serious views, not any sip of
a table.
Then he gives you accurate results.
So this is one hint and technique to deal with the Jule.
Thank you for sharing that.
It's very funny.
And good context.
As it pertains to the conversation we've had today, Yogesh, what's the future hold for
Kamatsu?
What are the next steps that you are taking in this project?
So as a part of Greenfield, we decided to go ahead with the phase wise manner.
Okay.
So pilot phase is certainly is to shape up our strategies, our reusable strategies I would
say, so we have clean core governance along with the infuse AI and classic announcement
strategy, and that we are going to utilize effectively to for our future roll outs.
So we have strategy in place.
We just talk about the processes, we took her to talk about the decisions and the scope
for our business requirements, and that comes from our business.
This runbook and strategy, clean core governance and strategy, we're going to utilize to have,
I would say, well plan and seamless roll outs and infusion.
Last question for you.
My asses of everyone I interview, I want you to think that you're speaking to a potential
future Jule for business, Jule for developers, customer, someone that's considering using
it.
At least you have for those listeners that are about to leverage that solution and maybe
even embark on a project similar to this one.
Yeah, I would say, see, you need a solid foundation from your infrastructure first.
So that's SAP rise, you can certainly consider.
And once you have the solid foundation on system build, then achieving a clean core is not
straightforward.
You got to think through, you got to set up a governance and you've got to utilize
enablers for this clean core governance.
Okay.
So enabler is Jule for developer that I explain a lot.
Now let me take this approach to explain the classic announcement strategy that we are
utilizing for category C and D to make a clean core.
So implicit and explicit announcement is a little bit hard, hard technical here, but the
core we are putting in this announcement is a generic and reusable code.
The actual business specific customization, we are having in a separate class configured
in a table and then that code we are invoking here into this customer exit, user exit and
implicit announcements through field symbol programming, reusable, a generic core.
So this way we are keeping the code of SAP as close to standard possible.
Okay.
One is something goes wrong.
You're doing your upgrade SPA, you underscore ENH process, SPDD SPA, you just got to replace
that reusable code because it is same all over these exits.
Only the differentiator is the constant.
So this is another, I would recommend to best take a base utilization of it.
So done in a nutshell, there is a clean core governance with the combination of Jule for
developer and classic announcement strategy can part of your system to make it future
ready.
Okay.
Yogeesh, thank you so much for walking us through Komatsu's journey with Jule.
If that's for, really appreciate your insights here today.
I mean, thank you for joining the podcast.
Thank you so much for giving us a promotion day again.
Thanks, Yogeesh, let's take a look what's happening around the ASA community this week.
First, learn about SAP's recent commitments around on-premise maintenance and support,
including more flexibility on unused licenses, greater ability for customers to tailor support
arrangements to their operational commercial priorities, and streamline processes for returning
customers to resume SAP maintenance and support without administration fees.
Click the link in the description to read the full article and what it means for you and
your organization.
Then dive into another ASA editorial article around Puget Sound Energy's SAP sales and service
cloud platform implementation.
Read how the utility managed a seamless cutover and achieved its main goal of having the
least amount of impact on its business as possible.
You can read the full article by clicking the link in the description.
For ASAug, I'm Jim Lichtenwater, thanks for listening.
Podcast Summary
Key Points:
Komatsu is undergoing a Greenfield SAP ERP transformation, aiming for system consolidation and future innovation readiness.
The AI journey began with SAP RISE, providing AI credits that enabled the adoption of Joule for Developers and Joule for Business.
Joule for Developers has saved 30-40% developer time by generating code, CDS views, OData services, and Fiori apps, acting as a catalyst in development.
Joule for Business, still in development/quality stages, offers conversational AI capabilities to reduce clicks and simplify user interactions for business users.
A clean core governance strategy, using SAP’s Level A-D categorization, is central, with Joule infused into design reviews to filter customizations and maintain upgrade readiness.
The "classic enhancement strategy" uses reusable, generic code for customizations, keeping SAP standard code close to baseline and easing upgrades.
SAP Business Technology Platform (BTP) is critical, utilizing Business Application Studio, middleware integration, MDC for bank integration, MDI for master data, and SAC for reporting.
Specific use cases include adding custom fields to product master and building Fiori apps for production orders, all guided by Joule.
Komatsu has over 30 developers using Joule for Developers and around 10 business users for Joule for Business, with plans to expand.
1
Future steps involve phased rollouts, leveraging established strategies for seamless implementation across the organization.
Summary:
Yogesh Chawaini, SAP Senior Solution Architect at Komatsu, discusses the company’s Greenfield SAP ERP transformation and its concurrent AI adoption journey. The project, initiated under SAP RISE, provided AI credits that enabled the integration of Joule for Developers and Joule for Business. Komatsu’s goal is to consolidate multiple SAP and non-SAP ERPs into a single system, making it future-ready for innovation. Joule for Developers has been instrumental, acting as a catalyst that saves 30-40% of developer time by generating code, CDS views, and Fiori applications, allowing developers to focus on design and governance. Joule for Business, though not yet in production, promises conversational AI to reduce clicks and training overhead for end users.
A key focus is maintaining a "clean core" through a governance framework based on SAP’s Level A-D categorization. Komatsu infuses Joule into daily design reviews to ensure customizations align with clean core principles, achieving 52% Level A compliance. The "classic enhancement strategy" supports this by using reusable, generic code for customizations, keeping SAP standard code intact and simplifying upgrades. BTP plays a vital role, with tools like Business Application Studio, middleware integration, and SAP Analytics Cloud. Use cases include adding custom fields to product master and developing Fiori apps for production orders. With over 30 developers and 10 business users already engaged, Komatsu plans phased rollouts, leveraging these strategies to ensure seamless future expansion.
FAQs
Komatsu's AI journey began last year when it signed a contract with SAP RISE, which included AI credits. These credits allowed them to plug in SAP AI tools like Joule for Developers and Joule for Business.
Joule for Developer acts as a catalyst, saving 30-40% of developer time by generating code and helping build apps from CDS views to OData services and Fiori. It allows developers to focus more on design, review, and governance.
Komatsu uses SAP's clean core levels (A, B, C, D) to filter customization risks. They review governance in daily design reviews, infuse Joule for Developer, and use a classic enhancement strategy to keep the core upgrade-friendly.
For a business requirement to add a custom field to SAP's standard product master app, Joule provided step-by-step guidance, generated OData services, classes, and bindings, enabling a quick and seamless implementation.
Joule for Business offers conversational AI, reducing clicks and training overhead. Users can interact naturally, such as asking to show sales orders for a specific customer or pending purchase order approvals.
Komatsu uses BTP for custom developments via Business Application Studio, middleware integration, bank integration with MBC, master data integration with MDI, and reporting with SAP Analytics Cloud.
Chat with AI
Loading...
Pro features
Go deeper with this episode
Unlock creator-grade tools that turn any transcript into show notes and subtitle files.