Building the Autonomous Suite from the Inside Out with Manoj Swaminathan

Welcome to the SAP Autonomous Enterprise Podcast, where we unlock the future of enterprises. Join us as we explore cutting-edge technologies, strategies and success stories, driving business towards autonomy. We talk with exciting guests from in and outside SAP, to listen how the gap between ambition and execution can be closed, exactly at the point where most organizations get stuck. We are more than happy to welcome Anos Vaminata in today's episode, discussing the evolution from the traditional enterprise software towards a gentec AI-driven business operations. Here is the General Manager and CPU, aboard the Autonomous Suite, Finance and Spend, and the Member of the Extended Board of SAP SE. Manos, welcome to our podcast today. Thank you so much, Benedick, who live having me here. That's super nice. So, Manos, we want to start really in a relaxed conversation with you, and I'm really curious about this and about the first question. And we asked this really to our every single speaker, and maybe you can give us a little bit more insights, how would you define the autonomous enterprise? And also, what problem in today's enterprise software landscapes are we really actually trying to solve with this? Absolutely, yes. I think there's certainly a top-of-mind problem for every organization right now. So, every enterprise leader now has an AI mandate, and with the changes that are happening in the industry with AI, leadership wants transformation, every organization's board wants to see the transformations with what is possible with the AI. And, you know, businesses wants to see efficiencies effectiveness, and, you know, they want to see that effectiveness and efficiencies with costs down as well. And it is the top-of-mind problem for every organization. And with all of these things, you know, you still need to run your business, you need to make sure that you're able to close the books, you know, do the movement of goods, pay the suppliers, the customers without any interruptions if you will. So, the question is, you know, is AI can deliver the value at scale without adding more complexity, more risk, and more fragmentation? And most of these companies are extremely challenged with this, and their realization with whatever that their journey has been with regards to AI is that it's not there yet. Right? So, how do we actually get this investment with AI? A real value for every customer, and how do we actually make sure that we really enable instead of somebody actually having this as a means by which as manually as is how they use to run some of these things to run their business, with an ability for them to be able to sense what's happening not just within the four walls of the organization, but also surrounding, you know, in terms of, in terms of, as is who you do business with, what's happening in the market, in the industry, around you for the likes of your business, and help the organizations to be proactive in managing every set of situations that what are possible that an organization could be running into, is really where this notion of autonomous enterprise starts. So, you have an ability for us to be able to have a set of, you know, agents that really operate across your entire set of organizations' business, really looking into our rather, you know, as is what I'd like to call a sensing, the things that need to get attended to, and it reasons those sets of things that what is of interest to the organization, with the models providing a level of intelligence and insights with the right recommendations, that the users can act, as well as based off of how the users act, the system learns, and then if it can act, and behalf of those users as well, and then, you know, you're able to actually see, you know, a continuous set of actions, gradually moving towards system taking those actions instead of humans, that's really the autonomous enterprise evolution. Our goal from SAP standpoint is truly enable every organization for them to be able to realize this notion of autonomous enterprise so that they can, you know, really realize the value as right out of the box in a way as a SAP could provide. So, that's how I would frame in a very simplistic term for us what autonomous enterprise means. Okay, that is really interesting, and to be very honest, I listened to many of your sessions at Sapphire. I had a look in the last couple of weeks about them, and you often told the audience, and you often say that the autonomous speed is the next evolution of the business speed. And now I'm asking a question that probably many customers have, what actually changed for the customers? So, absolutely, as autonomous speed is indeed based off of the SAP Business Suite as the foundation. If you recall what we announced as a spot of our business suite is, at the foundation is this data harmonization layered, and with the Business Data Cloud, we have provided the ERP as well as the line of businesses applications data harmonized in one consistent way, which becomes the data grounding in the foundation for all of the applications and AI for us to be able to provide value. The next layer is all of our applications. So, there is ERP and the line of businesses or the applications that I'm referring to. So, they are integrated as well with unified experience as out of the box capability. These two with the Business AI foundation becomes a basis for, you know, customers to be able to unlock value with AI, and that was the third layer that what we've had. With autonomous suite, what we have introduced is, we have introduced a new layer of applications as is what we call as agent experience, supporting six different set of constructs. And those are, you know, five core set of horizontal domains that we've introduced, and those domains as is also referred to as autonomous domains. And these are your finance, supply chain management, spend management, human capital management, and, you know, customer engagement. So, these five core horizontal autonomous domains gives you end to end functional breadth for any customer's business, but every customer belongs to an industry. And these industry requirements, you know, make the customer to meet certain set of compliance, regulatory set of expectations with the needed specialization as expected from an industry. And we also provide that as the sixth construct as part of the autonomous suite. So, an industry AI specialization on top of these core autonomous domains is what it is. So, these six constructs together forms this new layer of agent experience making up the autonomous suite. So, autonomous suite is on top of SAP's business suite, as is what we had with this app's data, with this AI value unlocked with these core autonomous domains as well as the industry AI specializations at what we are trying to provide. And an outset, what you can expect is every core autonomous domain and the industry AI will give you a set of assistance and agents that give you three fundamental constructs, one automation obviously, two intelligence and insights with right set of recommendations, three, it gives you a level of optimization unlocking value around a productivity or cost effectiveness or improved cycle times or other set of focused efficiencies that what an organization can unlock as its value. And we are doing this with just not agents with assistance and agents because we strongly believe we do need to have humans in this process as we go through this transformation. And every one of these agent experience is really enriches enhances the human set of typical day in life as is how they can actually experience their day in life in this organization. And that's collectively as is what is new with the autonomous suite on top of a SAP business suite that we've had. So, Manoj, when you heard now about autonomous suite and the different components here, do you maybe have a concrete example about what an autonomous business process might look like in practice? Absolutely, I mean, you know, as I talked about these core horizontal domains, you know, like finance supply chain spend management, human capital management as well as customer engagement, you can actually have, you know, any of these processes that we support today, we actually provide agent experience as a right out of the box. So, it could be source to pay, it could be record report could be higher to retire, could be designed to operate, we support end to end of these processes, you know, with these. agentic experience. So, let us take you know finances or autonomous finances and example. And in the prospect of record report, we actually you know have one of the assistance as an example is what we call as autonomous close assistant, right. And the construct behind this is like you know typically every organization has part of their financials close, they go through a series of activities and sometimes these activities are so intense, they have to work backwards from their public disclosure dates and in some organizations it's actually you know as bad as two to three weeks amount of time that they have to spend before they can really become ready for the public disclosure. So, working backwards from that they have to start doing this. So, how does autonomous finance close really changes this. So, now what we have is a series of agents that can actually be something that you can schedule on a daily basis now instead of doing this purely as a period and activity. So, for example, in a way this is how you can actually manage your missing accruals. In a basis is how you can actually do the clearing between your payables and receivables. In a way is a how you can actually manage your intercompany set of reconciliation that are very much needed which is very much a data intensive work and a lot of manual labor involved work today that every organization goes through as part of their period and process. Now we support this as an automated process that they can schedule on a daily basis instead of doing this as purely as a period and activity and every day as the system recognizes the things that what it needs to it provides a set of recommendations with explainability and the reasoning behind us why the system recommended those things. Wherever the system is able to go take the action on behalf of based off of you know the policies the God rails the rules that are defined the system will take the action on behalf of this. So, what used to be thousands and thousands of set of things that what needs to be processed cleared as part of the period and can be proactively managed by the system for you without needing any attention from the system and the humans. In cases where the systems not able to resolve automatically agents provide that as a request for the humans to come in and take a look at it and provide the response to that and every time the humans interact and provide a response to those questions agent continues to learn as well for the same pattern next time the agent can take the action on behalf of the user instead of prompting to the user. So, now you can actually see a close process that used to be very tedious at a things and a very time consuming process as well preventing organizations to be able to report on time. We now have an ability to be able to automate improve the cycle time of every period and process from one period into the next period and processing. Not just the cycle time aspect of it it improves the automation and involve the humans only more on the strategic aspects of things very very important without compromising the auditability traceability of the finance close process right. So, you can actually put that trust into these agents because we make sure that these agents are fully audit compliant in terms of making sure that they are ISO certified. They are sox certified and for these finance we're also doing this as sox certified as well so that you can actually do this for the reliance that you have with the humans doing these close processes you can have that with these agents doing this on behalf of you and the humans as well in the best way possible and that is why the processes are now something that you can see as it transform. Whatever I said as example as finance close you can pretty much apply to any of these processes from an end to end standpoint. Yeah, I like really your example that you explained and I think the autonomous enterprise changed the way of working you already said this but there are still many voices outside. I have the feeling that SAP just adding some AI capabilities to its software. Why is adding AI to individual applications not enough anymore? I think this goes to one of the fundamental principles Benedict, you know, in terms of what we've done. You know, I call this principle as you know, we're doing a shift from AI in apps to AI on apps and the reason is why we are doing this is, you know, first and foremost, you know, we want this has this new layer of experience that what we provide that really abstracts any and all aspects of the applications that what you are indeed dealing with and the reason is why we are doing this is because we want to be able to actually go meet the customer as wherever they are and if you provide this as an in-app experience, every customer would have to go through one and implementation and an update process for them to be able to realize to get to that new layer of the new version of the application before they can actually start realizing value. So that's, you know, more technical set of reasoning if you were to but then what is more important is, you know, each one of these applications if you look at it, whether you think about it from the perspective of ERP or the other lines of businesses with the application, like, you know, when you think about HCM with success factors, when you think about spend with procurement like a rebar field glass or from a travel perspective concur or, you know, some of these things could also be a homegrown or a third party application that what a customer might have, right? Our hope with in terms of providing this new agentically is to really provide a unified harmonized experience without customer seeing these application silos in a way as a solvitus and if we provide AI to these individual applications, it can solve and provide intelligence only for that module or that area of the application. But what customers really need is end-to-end business, right? And when it comes to this end-to-end business aspects, there is a connecting tissue that is needed for each of these sets of individual modules for them to be able to interact. This new agentic layer that what we have provided as AI on apps really binds not only providing these individual set of module specific intelligence, but it also connects across modules as an end-to-end business process because when a customer thinks about their business, they think about end-to-end business process. They don't think about individual systems, individual modules and this AI on apps really unlocks that unified experience as well as this end-to-end value from a process context without compromising that that we are able to provide. And we can also go meet the customer as wherever they are instead of expecting the customer to go through a major transformation. One more construct that's very important in this is that each of these applications have different experiences as how the user is interact with. With this AI on apps, we have harmonized that experience as well with this new dual work that what we've been able to actually provide so that users interact with only one consistency of experience instead of them going through these individual applications depending on what activity that they are going to complete. And that is something that we are able to harmonize as well with this new AI on apps principle as as how we have delivered these autonomous suite with the autonomous domains and industry AI. Hopefully that helps benefit. Absolutely, yes. Yeah, thanks a lot for this explanation Manoj and what is also in my head when you explained now the question so we all came from Sapphire. We all met customers with totally different backgrounds and what I'm asking now myself many customers are still for example running ECC, some of them have hybrid environments, some of them complex landscapes. How can they really all participate in this journey and how can we bring them all on board? This is really really important and very good question as well and I think if you think about the customers, I think you're absolutely right. We have customers and wearings, you know, faces of their, you know, ERP transformation as well as the business transformation that they're going through. So some are in our oldest set of releases like ECC and the on premises that have releases. Some have already transformed either through the rise and the growth journey in terms of moving towards the private cloud and or the public cloud and they use a combination of our other lines of businesses applications as well. What is also equally important for us to recognize is, you know, customers tailor our applications out of the box, they actually go through in terms of, you know, customizing our extending our applications to meet their business requirements. You know, I always love to have customers only use SAP products but we also know that that's not the reality. They use third party solutions and sometimes they actually use the homegrown solutions as well. So it's extremely important as we think about the autonomous a suite as a layer on top of this, it is not only working on top of SAP landscape, it is something that can be tailored towards catering to customers landscape as well. And when I say customers landscape, what does it mean? It means a couple of things. One is, you know, these agents that we are providing out of the box, they have to be something that can be nicely plugged into that customer landscape without compromising. Many of their applications and the investments that what they've done. So that means if they have customized, they ought to be able to extend the agents that we provide out of the box so that they can use them based off of the customizations that they've done as well. But we also recognize they use some third party applications are sometimes homegrown applications to meet some specific business process requirements beyond SAP solution. So these agents that we're providing should be something that they can extend plug into their set of third party solutions as well. And it can work hand in hand with their investment support they've done. And hence if we limit it only to work with SAP applications, we cannot go meet these customers in a good way. So that is why these agents are designed with extensibility in mind out of the box. And it can nicely plug and play with ability to swap to different tools, different systems out of the box. So it can work on a customer landscape instead of the SAP landscape. And that is one of the fundamental principles we applied as well. And one important point here when I can add this is also the business context, right? And why do you believe the business context is a key differentiator for enterprise AI, especially the key differentiator for SAP? You know, this is really, really good question. Every customer actually used SAP solutions to run their business and we've been in supporting customers for over the last five plus decades in terms of supporting customers business. We run some of the world's largest businesses through our solution. What it really provided us as SAP is the knowledge of, you know, how customers have been running their businesses over the last five decades. And hence, we have a good understanding of the data patterns, we have a good understanding of these processes and the industry specializations based off of different countries and, you know, regulatory requirements as well. And that is the secret sauce of SAP. And what we have been able to go do with this new layer that what we have created with the autonomous suite is we have packed that business knowledge that what we have accumulated over the past five decades as knowledge graph and the context layer that we feed to the models as well so that it can reason and provide the best tuned set of response to every customer for their business instead of a very generic AI as what most of these AI platforms can go do. The, at the beginning, I talked about, you know, customers have immersed it quite a bit with AI, but they've not realized the value with their AI investments mainly because many of these AI platforms though they are really, really good. They cannot provide the very targeted business value meeting customers expectations tuned to the customer's business, understanding the customer's data and the processes as how they have even tailored it. We can go do that because we understand the business and we understand the customers. We understand the customer's data and we have the process context as how they've done it. We use that to provide a very relevant targeted purpose driven set of experiences as our intelligence insights and recommendations through these agents. That is why it's so important and so differentiating that the customer should see us value and they would absolutely unlock some of those outcomes that I talked about when I think. Manos, what I also had in my head when I imagined being a customer from SAP and being now in this sole transformation, we often hear or explain that people direct, assist in orchestrate and agents execute. But what does that look like in a real business scenario? What does this actually mean? Let's take the same example, Julia, for the autonomous finance close that I talked about. For example, this process is very much, there are a lot of rules that really govern as how do you actually go do these? The clearing that what you need to be doing, recording accruals or doing the intercompany set of reconciliation and the management of the intercompany set of things as examples. There are a lot of regulatory expectations depending on what is the nature of business. There are a lot of rules that you need. An organization can define these policies and rules as as how they expected to be. The agents clearly understand that. But in a way, as a customer could now go define this and redefine the ways as how they operate is they can actually provide these policies, rules and stuff as input as they configure these agents in a way as how they would want to go use it. They can also provide the guidance to say that they want some of these agents to be running on a daily basis, hourly basis, etc. They can actually provide this. Assistance really take these input from the organization set of rules, policies and stuff. That set of principles around how these agents need to get execute. These assistance really orchestrate on those principles on behalf of the humans for the organization. They orchestrate these agents and these agents functions based off of those policies, applies the right data, uses the model to reason and does the set of activities that what it can actually go to automatically instead of requiring humans. And wherever it needs the help, it knows how to go and use the humans as well and continuously learn and iterate over this. So as a result, what you can expect is organizations can provide the needed set of guardrails as as how they would want. Assistance use those policies, principles, rules to orchestrate these agents and they can be one or more of these agents. And agents, you know, send space stuff of these things either based off of either time bound or event driven as an example. And it reasons the data with the models set of intelligence provide the right insights and recommendations. And the system takes the actions based off of those recommendations on behalf of those. So you can start experiencing touchless behaviors for normal processing, use the humans for exceptions processing, but also continuously learn as how humans interact with that, you know, so that you are able to get towards reducing those exceptions also to be a minimum, to less to none over a period of time, truly enabling customers to realize the autonomous enterprise. Right. And that is the value that we can actually unlock for the customer as they go through this. And that's what you should expect as part of this. Okay. Thanks a lot. That was so insightful. I really like this explanation. One of them. Yes, absolutely. And I think when talking about all of that AI agents and AI in general, we also should talk about trust, right? And it's how does SAP think about trust when agents begin taking actions on behalf of the users? So, you know, this is super important. And I think, you know, also very fundamental for SAP. Even as part of the journey that what we've been in the last five plus decades as I talked about in providing business solution, first and foremost, that SAP has been very mindful of us to not to compromise around the data usage, the security compliance, the privacy aspects of things in the way as a service. As we moved along with the industry shifts around, you know, our ECC to on prem to private cloud to public cloud in the true, you know, SAP managed sort of experiences. That is the trust we've been able to establish with our customers and our customers want the same going forward with these new agent experience as well. And this is also one of the fears that what every customer has and as they go into more regulated industries, it becomes much more of an elevated requirement. And it's an absolute requirement without which they won't even encourage adopting AI or these agents if you were right. So, having these security fundamentals, privacy fundamentals, data management fundamentals, the authentication, authorization aspects of things without compromising any of those or a fundamental requirement for our agents as well for them to be able to and we have factored them in a way as as how we have designed our agents right from the get go. And we also make this visible to our customers. for them to be able to go manage. In a way, as is how we have provided this is, we have extended what we have offered as the AI agent hub as part of the Lean Iax. Instead of that being component of Lean Iax, we have now made that as a standard component that every customer would get with autonomous suite. So in the agent hub, you can actually go manage the identity management for these agents who can execute on behalf of the users and a full fledged ability for you to be able to audit and manage every instance of agent's execution. And what path these agents took in terms of as they went through end to end. And you can without any compromise see the complete trace of these agents execution as well. So you have full visibility towards the agents execution, you have the full traceability of this. There is no compromise from the audit requirements standpoint as what the agent did, what was the state of that data and why the agent took a certain decision based off of that, you will be able to get that complete visibility as well without compromising any of those underlying security that what customer has already established whether they have established those authorization policies in the ERP or in any of these lines of business applications, we take those and honor those policies without compromise. And that is one of the fundamental principles as how we have offered. And something that the customers really has appreciated and value as well for the trust that they put with SAP, we have not messed up with agents as well. Panatic, hopefully that makes sense. This is so important, this is really important. And yes Manosh, as we are coming to the end, I have to tell you that the little birdie told us you are a dedicated marathon runner too. So if becoming an autonomous enterprise where marathon rather than a sprint, what is the one thing leaders need to get right in the first few miles? That's a very, very interesting question. So, yes, indeed, I use run to actually keep me somewhat, you know, honest to as well as you know, it teaches me perseverance as is what it is. And that's what I would say for this for any leaders embracing autonomous enterprises part of their transformation. One of the fundamental things that we tell as a spot of running marathon is, you know, start with the intent that you know, you don't run at a full speed at the beginning and gradually improve your speed so that you can actually have your energy throughout the full run if you were to. That's the same principle I would say just because we say this as autonomous enterprise, please don't expect this to be an overnight transformation, right? We do see this as a journey, as as how you would go through in terms of realizing the autonomous enterprise. As I described, you would be able to have your policies, your God rails, your rules that you would be able to establish, take the help of the assistance to be able to have it orchestrate and these agents to be able to go do the work. Initially, there would be a lot more human engagement in terms of these agents requiring human input for it to be able to but trust that these agents will learn through these humans interaction and truly enable a touchless processing for both normal as well as exception processing in helping you to truly realize autonomous enterprise. So don't run fast, you know, take it gradually and you would be able to realize autonomous enterprise with SAP guiding you throughout. That's what I would recommend. Thank you so much. This is really good answer. Thank you so much Manoj and as Benedict mentioned, we are coming to an end and what we'll listen to should really take from this series. I would say really this one sentence, the autonomous suite is the next evolution of business suite and I think what we can also add here is that the boundary is no longer really the application is the boundary is the customer's landscape so I would say and I want to ask you Manoj, do you have any final thoughts or something you want to give our listeners with so really your closing sentence or something like this? In a way, this is how we've actually provided the autonomous enterprise that with the autonomous suite as well as the SAP Business AI platform. We have applied the knowledge that what we have gained over the last five plus decades in terms of working with customers so that we take into account, you know, really enabling every customer to realize value with this new agent experience is that what we have provided and we have done it in such a way so that they can start realizing this value wherever they are instead of going through a major transformation to start realizing value and we would love to actually work with every customer as a part of this so look forward to working with you and have customer taking the opportunity to be able to you know go through the transformation and realizing the true autonomous enterprise for their own business so thank you again Julien Benedict for the opportunity. Thank you so much, thank you again for listening and if this episode resonates to you, share it with someone who needs to listen to it. Follow us on your favorite podcast platform for upcoming deep dives, industry episodes and specials and we'll see you next time on the autonomous enterprise podcast.