Explore the Power of AI in SAP S/4HANA (with SAP Build Process Automation) [DT114]
[MUSIC PLAYING] Hi, welcome, everybody to DT114, exploring the power of artificial intelligence in SAP S/4HANA. My name is Lucas Kiesow, and I lead the Global AI Incubation Team for SAP S/4HANA. And I'm joined today by my colleague Raghu, here, from our product management team. Raghu. Thank you, Lucas.
Hi, this is Raghu Banda, and I drive the initiatives around machine learning and AI with SAP S/4HANA. Forward to you, Lucas. Thanks, Raghu.
Happy to be here with you. OK, so before we basically get into the technical details of this presentation, I briefly want to touch on the overall architecture and the strategy that we have in terms of intelligent technologies within SAP S/4HANA. And for this, let's take a look at the slide right here. And here, you can really see that basically our whole intelligent technologies portfolio is built, of course, on SAP S/4HANA as the backbone on the bottom left-hand side. Now, the further you move out, of course, we have different, different technology options available here, right? So one of them the light blue one being more embedded, meaning that they are a lot-- basically that they are embedded into SAP S/4HANA. And then the darker blue one where we really talking about side-by-side technologies based on SAP's business technology platform.
Now important to call out here that the embedded ones basically consist of situation handling, embedded machine learning use cases, as well as embedded analytics. And they feature similar extensibility options that we'll come back to in a second. Now for the side-by-side extensibility, we basically have side-by-side machine learning cases again, which will Raghu we'll come back to later. We have a side-by-side analytic solutions, as well as process automation and bots. Now really, we believe from SAP side that this whole portfolio that you can see here really offers our end users a lot of flexibility in terms of how they can and want to realize their cases in their own implementations, and Raghu will go into more details later on on what this exactly looks like. Now in terms of the extensibility options that we have, those are similarly aligned and should be fairly common by now depending on where we're looking at.
So if you're looking into the embedded use cases, you have the more or less standard extensibility options that SAP S/4HANA users have for a lot of other things as well, namely key user extensibility, as well as developer extensibility. So this is very much in line to the extensibility strategy that we have within SAP S/4HANA. Now when we're looking at the SAP BTP, here it's pretty much the same thing.
We are basically also following the extensibility options that are possible and provided by the business technology platform, namely developing custom code, as well as basically integrating SAP BTP services, which Raghu will also come back to at a later point in time. But let's also quickly take a look at what benefits our customers have and what our processes can and could look like in the future. Now, here you can see a very basic or a very simplified example of an automated order cash process. And as you move along this end-to-end process, you can really see that we've taken this more or less existing process and infused it with a multitude of different intelligent technologies throughout the process in order with the goal to improve the efficiency for our end users, for our customers. And you can see that there's a lot of different things here, whether it's actually executing things, whether it's really optimizing things, whether it's predicting things in the future. And our goal is, of course, to continue improving our existing end-to-end processes with intelligent technologies, as well as potentially even rethinking existing processes with intelligent technologies.
So changing the business process based on the technological options that we have available today. But what exactly are these enabling technologies? So let's maybe dive a little bit deeper into that. So when we're talking about the artificial intelligence technologies in SAP S/4HANA, the main ones really that we want to call out here first are around situation handling, which is really used to identify issues within certain business situations in your system, and notify the end user to really also take action based on the discoveries that were made by situation handling. Now, when we're looking more into the machine learning aspect here we, of course, come over on the right-hand side top right where we really looking at customer specific data, where we're looking into the history of what has that customer done in the system, what exceptions have there been? And also, of course, trying to derive potential future possibilities for our customers here.
And on the bottom left-hand side, we're moving more into the automation area. And here, of course, we're looking at really gaining efficiencies by automating processes that take a long time that are basically very repetitive and that the customer can more or less trust to be automated in order to improve the time being spent by your workers or by your employees. On the bottom right-hand side, to round it off, we, of course, also look from an analytics perspective. As you know, analytics is also a very big area for most of our customers.
And here, we are really trying to help our customers to make smarter and faster decisions based on the insights that they can get from their business data, and also, to certain degrees, use simple machine learning to predict and look into the future based on the analytic data that they have available. And then if we go even one level deeper when looking at not just the benefits that these technologies provide, but also really when these technologies should be used, we break it down into three, more or less, simple buckets. On the left-hand side, we can talk about the anticipation aspect.
So we use situation handling or customer can use situation handling to really recognize potentially difficult, potentially critical business situations in their system. And make sure that users or user groups are notified accordingly. So there's no time lost or no time being spent looking for this information or for the cause of this information.
Of course, we also try here to support the user with a follow-up decision or follow-up action based on the-- basically looking for a resolution based on the notification or the situation that has been discovered. And of course, here, as mentioned before, we have similar extensibility options and we are, of course, also trying to improve our automation capabilities specifically in the area of decision-making more in the future. Now if we look at the middle column, we are talking about optimization where we're looking really into the machine learning space embedded or as mentioned before side-by-side. And as I said earlier, here the focus is really on customer data where we're trying to look into historic data, where we're really trying to use what has been happening in customer systems in order to help the customer predict or at least anticipate in that sense also the future of what's going to happen. And at the same time, of course, also help the user not just identifying as we do with situation handling and solving them as we also do, but really kind of trying to prevent potential situations in the future.
And, yeah, looking at things like trending, of course, as part of this forecasting as part of this, and Raghu will go into a little bit more detail later on the use cases, but of course, there's a multitude of things happening here that customers can use to their advantage and to improve their processes. And depending on the complexity, it really differentiates whether or not we are looking into embedded machine learning cases which are more or less a little-- maybe in some cases, a little more simple or into side-by-side cases where we can really leverage business services provided and provide greater complexity. Now looking at the execution part on the right-hand side here, we're really moving into this process automation space where we're really trying to automate, extend, and adapt business processes. And that's really key because with this, we try to offer, first of all, a reduction of manual efforts for our end users. And secondly, we are, of course, trying to provide our users with more flexibility, meaning when we are defining the logic, when we are really looking into how to manage, how to schedule, how to define these use cases and these bots.
That's where we're kind of trying to provide additional value to our customers. But that's sufficient from this perspective, I think. And now, Raghu, I would like to hand it over to you to take it a little bit more into details on these individual technologies.
So Raghu, over to you. Thank you, Lucas, that was an amazing and a wonderful introduction, and also you have laid out a very good foundation for our audience about the different technologies and the aspects behind them. So now we'll walk a bit further into these different aspects of these technologies. So let's start with the situation handling framework, which can help basically with this rule-based solving of issues and anticipation of what is going to happen.
It's mainly about these alert mechanisms. These are around the user-centric behavior. The users are trying to act upon an incident that is already available or which is about to happen. Typical examples here or typical business situations here would be like in a finance scenario. You know that there are some exceeding budgets.
You know that something is going over the budget. You can already trigger an alert. Or in the case of a procurement scenario where you're ordered some item from a particular vendor or a supplier and there is a delay in that.
So you will get an automatic alert mechanism. Or if you take a sales line of business. There is some particular sales order item which has been following-- the salesperson or the sales representative has been following on this particular thing, but you know that something has happened and some alert is triggered. So this is where you could also have this kind of a scenario. And there are a lot of these scenarios that are available out of the box which SAP has provided in the standard situation handling, which is where you have about 128 use cases that are available where you could analyze these details and automate these resolutions.
There is also a possibility of extending these situations. And the customers and the partners can also provide or create their own situation objects, their own situation scenarios and their own situation templates with this extended framework. And this will help you anticipate and create scenarios or create instances in your scenario and then handle it accordingly. Now that we briefly talked about the situation handling framework or the anticipation side of the things, let us briefly touch base on the machine learning aspect of the things are optimizing aspect of the things. Like Lucas was mentioning, there are, again, different approaches of doing machine learning.
There are for the simpler use cases, we use the embedded approach where we use the machine learning algorithms that are embedded into the SAP S/4HANA business processes. And a lot of SAP homegrown machine learning algorithms are used in this embedded scenarios. Here, typically what happens in this embedded approach is that the CPU usage is low, the data, the application algorithms everything set on the same stack. Whereas now and the typical good example here would be supplier delivery prediction where a supplier is trying to procure an item and there is a delay happening.
And you could be notified ahead of time, and then you could also be notified how many days it is getting delivered, delivery delays happening on that. So these are the simpler use cases or easier use cases, which are predominantly focusing on trending or forecasting or minimal classification and clustering kind of things. Now, there are situations where you will have a lot more application data is needed or you might have to use some complicated algorithms. And here, you can build up on additional algorithms that are outside of SAP's HANA ML algorithms. You could leverage the TensorFlow library or the Python library, and you could also leverage data that is coming in from external data sources. And this is where you could build on these complex algorithms and typical use cases here would be like your image recognition or language processing or text recognition, and so on and so forth.
And here what happens is that the ML logic, the algorithms and the data are residing on the SAP BTP stack, whereas the application is residing outside of SAP BTP here. The application is on SAP S/4HANA. So the difference is that-- but it will be seamless for the end user or the customer that everything is happening seamless at the same time because the configuration is available. We'll also briefly look into the different scenarios, the different use cases that are available, whether it is embedded or side-by-side. Let us now briefly talk about the RPA or the process automation side of the things, which is where we talk about the execution of the tasks.
We know that in many scenarios, there are-- you might receive a lot of emails when there might be receiving a lot of emails or a purchaser might be receiving a lot of emails. And many of this information has to be processed manually, and you will have to upload it to the SAP S/4HANA system. So this is one such scenarios where-- this is only one of the examples I was talking about.
And this is where you have a lot of these RPA bots. These are available that can be built into that. And also in addition to the RPA bots, we also have this additional workflow process automation, which is also embedded into this process. And this is where you have the complete the process automation aspect of it. So here, again, from SAP standpoint, there are a lot of scenarios that are available, a lot of bots that are available out of the box, which can be leveraged. Again, these bots are running on the SAP BTP, and these are configured to run with the SAP S/4HANA business processes.
Now that we talked about this, let us briefly touch base on the analytics aspect of the things. So we talked about the anticipation aspect where we are trying to-- where we understood the situation handling aspect, and then we talked about the optimizing aspect where we talked about the different machine learning scenarios and then we talked about the execution. Now, on the bottom line, analytics is a framework or analytics is available within SAP S/4HANA with our theory apps and other things. But in the same way, you also have the SAP Analytics Cloud, which is where you have a lot of these things available. There are four different buckets here.
We'll start with the conversational aspect or automated insights, and then you have your automated discovery and then you have your predictive analytics. So what we are trying to do here is that you can search to the different insights using the conversational aspects in the SAP Analytics Cloud or you can talk about the automated insights where you could provide, take the smart insights and do some forecasting, or you could also do some smart discovery that is available where you can automatically generate these different stories. Or you could also use the smart predict functionality that is available within SAP Analytics Cloud. So the concept where we are talking about in the context of SAP Analytics Cloud is that we have this additional analytics framework or analytics layer, whether it is embedded into SAP S/4HANA or it is available as a SAP Analytics Cloud, this could be leveraged.
Now that we talked about the aspects of the different technologies of AI that we have, let us briefly get into the different ML use cases that are available. Here on the right top side, you will see all the use cases around the finance line of business. Many of these use cases are either embedded or available as side-by-side.
In the context of today's session, we will not be differentiating them. But if you go into our documentation or other slides, you will see which of these are embedded, which of these are available as side-by-side, you also have detailed demos. On the bottom right, you see the use cases in the sales line of business. And on the left-hand side bottom, you see the use cases from the manufacturing line of business. And the rest of the use cases on the top left, you have from the other different lines of businesses like the procurement or the professional services.
So bottom line is that there are a lot of these use cases which are truly integrated and available with our intelligent ERP, whether it is SAP S/4HANA cloud or on-premise. We also have use cases outside of SAP S/4HANA beyond SAP S/4HANA and some of them are available here starting with the manufacturing or the digital supply chain, and so on and so forth. In the context of this session, we are providing these details at a high-level, but we have all these demos available for you to go into details and understand what is there. So now, before we wrap up the session, I also want to briefly talk about the lifecycle management that is available when we are consuming these machine learning scenarios, whether the embedded scenarios or the side-by-side scenarios. And the lifecycle management that we use is the Intelligent Scenario Lifecycle Management, or ISLM. This is a successor to the SAP predictive analytics integrator.
This is available in the SAP basis layer in SAP S/4HANA for free of charge. It is a single point of contact one central cockpit, and it has smooth features that are available. And if you go a bit further, you will see that you will-- there are two apps that come with this ISLM framework, which is embedded into the SAP S/4HANA. And these two applications or apps, one of them is called intelligent scenario lifecycle management and the ISLM, the other one is intelligent scenarios app. These are used to-- SAP has already delivered quite a lot of these scenarios that are available.
So you could out of the box use these scenarios or you can create new scenarios if you are on-premise. And the thing is that the scenarios that are already available, you could do-- you could use your own data sets and obviously, you train your data sets, you deploy them, and you can activate these for a particular business user or a particular user or a set of users. And then you can also monitor these different scenarios. So this is-- in a nutshell, this is what we have-- the different things that are available. And there are, again, demos available, which will explain you how you can go into the details about all these.
So before I conclude this session, I want to provide you thoughts about a lot of the material that is available for further reading. On the top left-hand corner, you will see details about the intelligent scenario lifecycle management that I was talking about for the cloud version, as well as the on-premise version. There is a blog series that talks about that in detail from the ISLM team. We also have quite a few recent updates, whether it is some video blogs or whether it is the other blog series, which talks about all this. We also have a trial option, which is available as a SAP CAL image, Cloud Appliance Library, which runs on HANA Enterprise Cloud.
And this can be triggered on AWS or Azure, and this has a couple of scenarios for the customers or partners to try it out. We also have a few social channels that explain about these different things in the SAP community pages. Recently, we also released a podcast that talks about these different topics, the technology topics or the different use cases that we have with different SAP guests who explain about it. And these are like 35 to 45-minute conversations, which will help you get started on your journey.
Finally, I would also want to provide you some thoughts about the other TechEd opportunities, the other TechEd sessions that are available, other TechEd opportunities that you could go ahead and look into. With this, I would like to thank you all for your time in dialing into this session. And the contact information here, you could look into that, you could reach out to Lucas or myself, Raghu, for any further information or any further things that you would need.
Thank you all. [MUSIC PLAYING]