SAP: Bringing the ‘Operating System’ of a Company into the AI Era with CTO Philipp Herzig
No Priors · 23 Apr 2026
Source: https://traffic.megaphone.fm/PDP1374150825.mp3
Transcript: https://podcasts.opennash.com/shows/no-priors/sap-bringing-the-operating-system-of-a-company-into-the-ai-era-with-cto-philipp-32f8da/ (via OpenNash Podcasts, machine-generated)

Hey listeners, welcome back to No Priors. Today I'm here with Philipp Herzig, the CTO of SAP, the enterprise juggernaut. We talk about their AI strategy, why SAP has endured and thrived through several technology transitions, why entrepreneurs are underestimating the challenges of scale, why AI's a business model transition, not just a technology transition. Why he thinks that LLMs are not enough for predictive analytics, and even about the traveling salesman problem in the real world and the state of Hermes.

Welcome, Philipp. Philipp, thanks so much for being with us.

Yeah, it's a pleasure to be here, thank you.

Everybody knows the name SAP. But I do think that for, uh, lots of engineers or people who aren't close to the system in a larger enterprise, they don't really know like the breadth and function of the platform. Like, can you just describe, uh, what you guys do for customers?

Oh, absolutely. I mean, look, SAP is the market leader, right, in enterprise, uh, software applications and platforms, right? M. MySAP 4000 enterprise customers, and usually I just running their finance and HR, and you know, supply chain, manufacturing, execution, logistics, warehouse management, and then, of course, everything on the customer side, sales, services, commerce, procurement, you name it, right? Um, end to end, right? Like SAP, we always say we have the broadest portfolio in terms of, uh, end to end, running the business, end to end.

This is where SAP started with, right? Uh, uh, real giving real time insight, and usually I really describe this as it's not just software in itself, it's kind of the operating system, right? Uh, of a company, essentially, uh, in order to, um, get, you know, from from everything from order to cash and source to pay, right? End to end managed for companies around the, uh, around the entire globe.

Hm, hm. Um, I definitely wanna talk about AI, uh, LLM, some of the stuff that you guys are doing internally, and then around, um, predictive models as well. Uh, but just because the the macro backdrop is, uh, on everyone's mind, both from a technology and an economic perspective. Um, I wanna talk about, like, SAP's position in the market a little bit. SAP has, uh, has stood the test of time through multiple technology and market cycles. I, as a early stage venture capitalist, am am kind of on the other side of this, where the narrative is like, well, when you have internet and cloud and mobile and AI and social, like you have, um, uh, an opportunity for new players.

Um, what, what do you like? Uh, SAP, you know, even today is the um, I believe the largest, like market cap enterprise software vendor versus sort of the the last generation of the new guard, like the salesforces of the world. Um, w How has that happened? Like, how did you, how did you do it, and what is make? What makes it so durable?

Well, what makes it so durable right at the end of the day? I mean, if you think about this, and it's happening as a little bit the same way. Also when we talk about the Sasa stat narrative or the Saus parkellips, or I mean, anyway, we are I. I have the feeling like in this market last year, AI was in a big bubble and everybody was kind of saying, no, it's not, and not this year, Sasa stat, uh, and and so on and so forth.

Look, the the reality is now, of course, with the the the costs of building being so low, right? With specifically agentic coding and all these latest, um, powerful models. I mean, something has always prevailed over the years, because even when SAP was founded, right in 1972, right, long time ago. Um, I mean, why was it started? Because actually in the 70s, when the founders of SAP worked still at IBM, what did they do? They went to each customer, right?

And they implemented the finance system again and again and again and again. And then they said, like, Hey, this makes no sense, right? Because the economics, it doesn't scale, right? Because, of course, you can do this right, but you can only add so much value right in any given time. And by the way, we are basically programming the system very similar. Of course, there's always a little bit that is specific then to the customer. And this was the idea where standard, the the notion of a standard software was born, essentially right.

And then, of course, that stood the test of time, right? Because there is simply things and companies that need to get managed right end to end. And that also has transformed throughout the years. You've mentioned that, right? 1st from the mainframe to client server, right, then to the internet, then mobile, and now, of course, AI. So, of course, the software has changed and evolved all along with these technol. But what hasn't changed is what customers are seeking for, which is outcomes, right outcomes and return on their investment in order to get the things done right.

Um, towards, and and of course, now AI is an amazing technology that again, uh, helps to get more things done in the enterprise, right? And then that is actually what SAP is standing for, right? And uh, uh, and so what we are really doing is in giving, of course, also the breadth of the portfolio and the customers is. Of course, to help customers to achieve more by deeply embedding AI, AI agents and, of course, transforming now the user interface and so on and so forth to help them get more right done in in whichever industry that they are that they are working in.

And um, and and and we believe that still will, uh, con con continue, right? Because what? This is exactly what we're also seeing right now, with, of course, there's still, of course, there's tremendous progress. But we also see that the AI adoption in the enterprise is still not where we want to see it right. There's like, there's this Gardner curve, right? Where say, like, there's this AI innovation race, and then there's this AI outcome race, right?

And the gap almost increases, right versus getting getting narrow and and that is what what we are really focused on, right? That customers get the outcome, uh, from from AI to to achieve more, given, of course, uh, the foundation we have. But simultaneously, of course, the system we are, we are kind of re engineering the entire system, right, with the help of AI, uh, in in a, in a totally new way. Yeah,

hm, hm. Um, you, uh, now, as CTO as of SAP, have, like, a very broad purview. Um, it does so, includes the AI strategy piece of it, um, internal and for your customers. Like, uh, what do you think of as your your own top priorities for the organization? And where is SAP on this, um, re engineering or reimagination journey?

no, look, I mean, we, we are. We are in the meantime all in on AI, right? Uh, so I mean, everybody in the company is using agentic coding, right? Like, because that's, of course, an amazing uh, um, uh uh, productivity boost, right? That our developers have, no matter in which programming language they are building, uh, the the software for the customers. But of course, it's also really, um, again, focusing on customer outcomes, right? And we've seen this, for example, early in the early days now with consulting, for example, right?

We built this thing called Ju for Consulting, which is phenomenal because, uh, it's one of our fastest growing AI products. Because what this actually helps is to build the the um, uh, to to help the consultants, right in the in an SAP project, or in a complex landscape, if they're, I mean, a again, right. We are serving some of the largest customers. They have a lot of heritage, they have a lot of complex landscape to help them actually to move into the cloud, to adopt the latest AI capabilities and so on.

And with Ju for consultant, they can reduce like 30 of their efforts, right, to get to the outcome faster, which, of course, then directly reduces the cost, not just the time, but also the costs that are necessary in order to get to, to, to get to the to, to the latest software. And we've seen this, of course, uh, with Concur, for example, right, where now our travel booking agents or expense agents are are alive. And so so there's many of these outcomes that we are designing.

But when you look from a CTO perspective, really, it it's in my mind, it's three things that are really getting not not disrupted, but are massively changing. Yeah, to me, the metaphor is a little bit like when we move from on prem to the cloud, right? Originally, everybody thought, Hey, we just take the on prem software we already built, we put it on the internet, call it cloud, right? But then only over over a certain period of time, people realized, Oh, what does actually CICD mean, right?

You can deploy daily or multiple times per day. And then you realize, oh, hm, we always had multi tenancy in the on prem software already. But then, of course, in the cloud, you have to learn how to scale it up and down, right? And like, all of a sudden, the software kind of got re engineered, right, to really, you know, build real SaaS. Software that that, uh, that, uh, uh, with all the concepts right in cloud computing, and with AI, the same is happening, right?

And then it's happening on three levels. It happens, of course, on the UI side, right? The time is clearly over. Where you design software, where the dumb software, where the that requires the intelligence to sit in front of the computer, right? So, I mean, if you look at classical software, what did you do? You designed a user interface trying to. Hopefully, you did some user research. Try to figure out in the easiest way or the most intuitive way to figure to.

To teach a human how they get their tasks done by clicking through the UI, essentially, this is over, right? It's now we call this generative UI, right? So the UIs get dynamically generated, right? If you have analytical questions, for example, or if you wanna do your deep research, not just the deep research you find on Perplexity or the the usual chatbots, but deeply rooted. Let's say tariffs are being introduced or new taxes, or the state of formals.

What does this mean for my supply chain? And then you can analyze this in conjunction with your SAP data. So there's a lot of exciting opportunities and new things you can build that were almost impossible that that we only dreamed of in the last, I don't know, 20 years, at least since I'm a developer. Where now, you know, the system becomes much more multimodal, much more proactive, right? Because it can run overnight. And then only if you wake up in the morning, tell you, Hey, Sarah, have you looked at this, right?

Here's a problem on the sales side, maybe the order entry is going down. You should do something here. And here are some recommendations already for this. Or here, there's a problem in the supply chain, right? Uh, because now, if you're an oil and gas customer, obviously you wanna know what are my options? You need to replan, right? Like, so all these things become super important for customers, and that changes the UI. Then the 2nd one is, of course, the business processes, like in order to cash.

In the past, of course, it has variances and so on and so forth, but it was a rather rigid process, but not like the standard operating procedure of a company. But now, of course, with these agents, right, we can blend the structured and the unstructured world. Yeah, more seemingly so, to get actually more work done. So this whole move from software as a service to some call it service as a software, outcome as a service. That is, of course, what these agents are building for us.

And then, of course, below that, you have the whole data layer, right, the whole data layer of bringing. Of course, SAP has a lot of super valuable data for a company, right? Like all your general ledger and your invoices, and and your warehouse and inventory information, et cetera. But of course, you now want to combine this right with the plethora of other data, right, in order to kind of build this one harmonized semantical view. Because only we always say AI is only as powerful as the data is right.

So, and and that is exactly what we are then also doing and transforming on the data side to help our customers to benefit from a from a, from a globally harmonized data model to fuel the AI.

What is the, um, biggest engineering or technical challenge for for for you guys? When you look at, you know, these three bodies of work or anything else that you're doing at SAP?

Well, the the biggest challenge, quite frankly, is, of course, when you, when you look at this, is how do you? It's actually not the AI so much, right? But it's actually teaching the AI to do the right thing at scale, right? Because, I mean, you can look, you can build two years ago, right? Everybody build a rack service, right? And you can. You could easily with a POC blew off. Everybody's, you know, the CEO's socks and like, look how easy it is to build a chatbot on ten documents, right?

But that, of course, but but SAP and and all these large customers, right, they are always have a problem of scale. Okay, what do you now with 100 documents? Well, it becomes a little harder, a thousand documents becomes a deep engineering challenge. And now, if you go into Jewel, you are Sarah, you're maybe an SAP, a US. employee, right? Of course, if you ask a question, of course, for a travel policy, for example, of course, you expect a very different answer that me as a German employee, right, would would get.

So you then now need to connect this actually with your master data, like, where are you located, in which country are you? on of, and and under which payroll are you actually, which taxes apply to you, and so on. So all of a sudden it becomes a very, very tricky problem. Same with MCP, like last year, oh, everybody could build an MCP server. It was so super simple, right to to to hook up your MCP server and do amazing things with it.

But that becomes ver. Like, for ten API is not an issue, hundred, you get already, uh, context, blowed and all these these challenges, but we have 20 zero APis, Oh my goodness, right? So it becomes this like, yeah, it's like, because it's so huge, right? Because there's so much things. So it becomes this problem of scale, right? And doing this really end to end for the customer. Because what we also build is really an integrated experience across.

So you can ask a finance question, you can ask your HR question, your supply chain, you can correlate that. Like, this is, this is the biggest challenge to bring that, so to speak, to to, to together, yeah, and design it then really for the uh, for the, for, for the right outcome. Yeah, and you said this also, what's interesting, and another interesting thing is, um, from my perspective, is I. You had recently this other podcast, I think with Andre, you had it, uh, on the.

The most important thing from a development perspective is actually that people start writing their evals. That is with this like, I was on this tour for a very long time now. Um, because, um, the the problem. And why does agentic coding work so well? Sarah is, um, of course, you can verify the outcome, right? Uh, you can either say, Hey, is the program compiling? or that? are your unit tests, right? Does it work, et cetera? And and of course, combined with a little bit of taste, uh, and a lot of hard engineering work and tropic and openai built these phenomenal code generation models.

The problem is, if you now want to build a reliable outcome in finance and so on, you need the data that say, hey, with this input, that's the output, right? In order so that the coding agent can validate that e and assert that against the reliable outcome. And that's something where there's a mindset shift in terms of how you describe the right boundary conditions to your coding agent, right? The harness like, like, uh, all the boundary conditions need to be true from a security perspective and from a data privacy perspective.

All the code qualities, because you also still wanna maintain that code on on day two and day three and day four, not just get get the 1st version byte coded. And then, of course, these evals, right, that then tell you, hey, this agent is actually doing what it's supposed to be doing in a, in a, in a, in a variety of ways. Do you still? And you sometimes need to laugh, because do you still remember in the when?

I was a computer science student, uh, where the Google guys came in in a in a lecture and they said, like, Hey, I can go home at five, zero pm. Because I wrote my tests. And of course, this was non. You still remember that, like test 1st or test driven development Of course, yeah, it's and so on.

coming back.

I, it's coming back. It's like the the. The reality is, nobody did it, at least I never did it. Because

no, it was not popular.

a, it was so much more fun, it was not very popular at the end. Why was it? Because a, it was a so much more fun to write the code 1st, right? And then b. Of course, usually the product manager gave you a very messy requirement. It was very hard for you to to write the test actually 1st. So you, whi, while you wrote the code, you kind of iteratively discovered what the system, how the system would actually behave, right?

Now, the behavior and the writing the code is so much automated, right? Because now you can write almost software like completely on its own. But of course, now you need to describe the right outcome, what you want from this thing. And so that changes very much how the developers, of course, also, uh, now need to work specifically. Now that the models, uh, have the steps changed since December last year,

it's really hard, uh uh. Or I think it's not obvious how to picture, um, if there's a a version of agents and, uh, models powering those agents in enterprise systems like SAP, getting better in a compounding way, the way they have in, uh, generic code generation. Do you think it's possible in terms of verifiability or the ability to, um, go understand and evaluate against that intent? Because it is much, um, I don't know if I would say it's more diverse in code, but it is, uh, uh, it's not.

It's not obviously verifiable, as you pointed out, like, do you think it can be?

that that's exactly the point. That is where the the starting condition is great, right? I think in terms of like two lanes, the 1st lane is, of course, you have the system of record today, right? You know exactly in the system. Hey, given this or that instruction, right, what is the outcome, right? Because you can see it in the database, right? And then you can construct, hey, if the order to cash process runs like this, then you need to expect right that the that, uh, and uh, the the cash, like the accounts receivable, needs to come in this way, right?

With the following taxes and so on and so forth. So that gives you verifiability. Now, the challenge, of course, is rather, that is never enough, right? Because, uh, if you just look into the system of record today, that data is insufficient for this grand vision that everybody has. That it becomes this autonomous enterprise. Or, like the agency of these agents, is increasing right over time. So to at the beginning, the agents, of course, are coming back to you.

Some people call this human in the loop or whatever, right? So they need to come back to you. Like, also still with cloud code or Codex, and still ask you some clarifying questions. Hey, I have now, I could now go this way, I could do that way. And with that, what, what? What you want to design for is that you start to capture more of that context, right? I always call this the tribal knowledge, the stuff that is not in the system, stored somewhere, that just lives in people's heads.

Or maybe in Slack channels, maybe in teams channels, maybe it was just a discussion on the phone, right? So it's not stored anywhere. So how can you drive a decision from that? And then? So the question is, the agent needs to come back, ask you for input. Now you wanna store that, and now what we do? In the past, we called this process mining, now we call it agent mining. Because you record all these decision traces, these context of what the users are entering into the system.

And then you can either use it to say, like, Hey, wait a minute, this is actually an anomaly. The folks in, I don't know, in UK from our company, or the folks in Australia shouldn't do this because the standard operating procedure is this. Or you say, like, Oh, that's actually a very good improvement. And then you can elevate this to be the new standard operating procedure. Maybe not just for Australia, but maybe for the rest of the world or more countries to run your company more efficient.

Because now you learn something how the organization behaves, because it can go two ways, right? It could be either good, uh, or it could be a bad thing. And then you maybe wanna, uh, streamline the process, how people then actually conduct the process in a different way. And that then leads to this kind of, I call this, then this data flywheel, so to speak. So because with every trace, every input a user gives you, with all the observability that an agent writes you.

You have new data sources that can then lead to new evals, right? Where somebody says, Yes, that's a verified output, uh, so to speak, that I want. And then, of course, you can optimize the system more towards that outcome, uh, depending on which which data you gathered.

Do you, uh, have a strong point of view today as to whether, uh, agents operating against these business processes, uh, within SAP or otherwise in enterprise software? Do you do you think it's going to be, um, computer use? Do you think it is all, you know, code and tool use on APIs?

It's an interesting question. Uh, I have not a a very spec, uh, a a finite answer yet to this. So I think, um, given, of course, also how clunky Uis are, and so on and so forth, and knowing the challenges also from UI automation from the past. I mean, it's phenomenal what they can do already today, quite frankly. I mean, it's still a little bit slow, right, and and and so on and so forth. But I still believe, for the most part, uh, it will, the the majority will live with tool calling, right, and and uh, agents running in the background and so on, right?

Because you also don't, uh, you know, uh, uh, uh. Maybe want to have the browser open all the time. Okay, we can do this with headless browsers and so on. But I mean, if you can do this right with a more structured approach from an integration point of view, I think that will be the preferred method. But then, of course, there will be always kind of things where an API is maybe not available, or you have a legacy s uh system for a time being and so on.

And then, of course, these computer use approaches and so on will nicely tie in, so to speak, um, as as well.

If we zoom out a little bit and just think about, um, uh, agents and um, automated business processes, in what domains do you, uh, hope customers will see that be most effective? 1st

Well, I mean, we need to be clear, right? I mean, it has been, for the most part, uh, very productive in what I call the unstructured world, right? Because let's face it, I mean, large language models are very good in the unstructured world, right? Text and the images and stuff like that. And so, of course, everything where unstructured data is concerned, for the most part, like in services and in support, and, uh, and maybe sales, right?

And then, of course, in anything related to knowledge work, right? That deals a lot with documents. Of course, this is where we see, like, just do it for consultant product. I've mentioned, right? This is a lot of unstructured information. This is, of course, where, you know, it was the easiest to get quickly to the to the, to the, to the return on investment. It was harder now to kind of combine this. Also, you mentioned tool use, for example, right?

I mean, the models had to learn, of course, 1st of all, and got needed, got better, right, on how to use the tools. And then you need to build orchestrators, right, and disambiguate. Oh, what does an order actually mean? you mean a maintenance order, sales order, purchase order? The the many, what, what order is a very overloaded term, it's very, uh, uh, very ambiguous, and um, so, and that's, of course, this orchestration log logic, that's what is a hard thing to build.

Yeah, and um, and so I think, uh, overall, but now that it has become gotten better, right? Now you can do things like chat with your data, right? And instead of going to the data analyst, business analyst, that curates you some dashboards. And if it's 80 of the cases, that might be a good enough dashboard. But for all the other 20 of the question, you always need to go back to your IT department. No, now you can just converse in natural language with the system, it pulls the data, right, natural language to SQL or whatever have you pulls that data.

You converse it until you have that point of view of the data that you wanna have. And then you just pin it and say, like, Okay, that's actually my problem. Now I wanna manage that problem for the next, I don't know, two, three weeks, until the problem maybe has disappeared. And then, of course, you you move on, maybe then you delete that tile, uh, and so on, and so forth, like, so, this, this kind of um, combination of the structured, unstructured world, which is required, right?

If you wanna go into the tabular world, right? Because lots of data in finance is stored in tables and sales and the supply chain and so on, and so forth, right? Um, unlocking that took a little bit of time, but, uh, now it's actually. We are seeing through, for example, the knowledge graph, the SAP knowledge graph that we've built, which is kind of the glue between natural language and the structured data in the system. To to to really bring this together.

that actually leads to one of your, um, I I think like, I don't know if it's unconventional, but it's certainly not the the dominant narrative in AI right now, which is your interest in, like, predictive and tabular um models. Uh, can you? Can you talk about, like, y, you know why LLMs aren't the be all end all here, or why we can't just use, um, uh, tools and um, uh, calculation external to the model and com in combination with LLMs to achieve what you want to achieve?

yeah, now, 1st of all, from a business motivation, it's a great question, right? Sarah? I mean, 1st from the business motivation point of view, right? Again, LLMs unstructured world, that's all good, right? But most of the time, if you, if you wanna plan forward, right, if you wanna make good decisions in a company, you need predictions, right? You need predictions in terms of, oh, what's my demand, right for? oh, is this depending on the seasonality effects and so on?

What's my demand forecast? Maybe right for my products in the retail store, or what's my demand, right, uh, for my products? So I can plan accordingly my manufacturing, right? If I'm a manufacturing customer, or you wanna predict your cash flow, right? Uh, you wanna pre. And that has a bunch of input variables, like, oh, what are actually my day sales outstanding, right? And that is determined based on are customers paying yes or no? That's a classification question.

And if you then say, okay, if a customer is not paying in the uh, within the payment terms, what's the payment delay? A classical regression question, and so on and so forth. Now, the problem is, of course, still today, if we look at these predictive questions, right? And and then you wanna maybe do a what if analysis from it, right? Now, if you want to do these predictions, quite frankly, then the challenge is. Large language models are not made for this, right?

The way how they, you know, generate just one token after another, essentially in a sequence to sequence modeling, I mean, they're language models, right? So that that then they do this phenomenally well. But if you still wanna do these predictors, where you have to go back to these classical machine learning approaches, right? You use XGBoost or AutoGluon and and many of these, uh, AutoML approaches, right? That that might be, uh, that that are still out there.

The problem is just it doesn't scale, right? So we haven't seen in the predictive space the same level of democratization, right? You still need to hire a very good, talented data scientist, right? And then if you, for example, if you are a large company, we did this, for example, at a pharmaceutical company. If you just want to solve the payment delay prediction problem, I've mentioned, right? You have they are. They are running in 90 countries around the world, and they need these two models.

So you end up with 180 models. You need to train, you need to curate the data, you need to train the models, figure out, right, what the right model, uh, is feature engineering, like the classical, uh, uh, machine learning kind of approach, right? That, uh, that, uh, that was used in the in the past. And what we said all the time is, okay, look, we have all this data in stored in these tables, right? Hundred, thousands of tables, right, where all this information is stored.

Can we not apply the same idea that large language models or multimodal models did for the unstructured world, actually for the structured? In order to start predicting things? So you can just basically provide a little bit of context, a small amount of data, not a large amount of data, because that was always a problem, small amount of data. And then starting making high accurate predictions, so to speak, in that domain. And that led actually this was two years of research.

We published that also in Europe and a bunch of other conferences. Uh, we call this RPT1, so rapid ones and for relational pre trained transformers. It's still based on the transformer architecture, but with a very different, uh uh architecture. Uh, we released this and we see some, some, some, meanwhile, some very, very good results from that. In in various domains. Where, as I said, classification and regression, sometimes time series and so on, uh, are concerned. And we believe this will be huge because it obviously will allow way more people from a business impact to, uh, to, to make these predictions, which large language models have a really hard time with.

hm hm. When you, um, uh, think about the gap in, uh, I don't know. I think you described it as like hype versus adoption within the enterprise customers like, uh,

The innovate innovation race versus the outcome race. Yeah, you're right.

innovation race versus outcome race. Um, uh, it's a it's a good framing, like m. The change is happening very quickly, that's hard for companies to absorb. Where, where do you see, um, challenges for the enterprise in adoption today? And where are customers making the most progress with with you, or where they most excited?

Yeah, that's a good, good question, right? I mean, usually I say the the the. The primary problem, as I said, is is is the problem of a data, right? Because most of the time, the data is, of course, very disaggregated in a in a, in a company, right? I mean, for a variety of reasons, right? Either because you made certain decisions, uh, how you purchase, uh, solutions in the past, or you did an M. A, right?

So you acquired a company. Naturally, of course, they bring a very different IT system landscape as well, and so on and so forth, right? So you have this aggregated information now. And the problem is, of course, that limits, uh, the potential of what you can actually do with AI, right? And and then the question next is, how do you integrate this safely? And what I see is clearly customers who did that kind of homework right now.

Of course, uh, it's not a new topic. We're discussing this for 1015, maybe more years, right? The ones that did their their homework, they of course have a much easier life right to then also reap the benefits. Uh, for example, of AI, right? The 2nd one, as I mentioned, is already is the problem of scale, right? The bigger, the more complex the landscape and so on, right? Then, of course, also then bringing this together in a unified experience is a challenge.

And then finally, of course, everything around then security and so on and so forth, right? Because then there's always then this gap between, oh, there's an amazing innovation. Take OpenClaw, for example, right? I mean, amazing, uh, what, what, what this has brought to the world in terms of further ideas. And of course, I mean, from a security perspective, and so that's a problem. You don't want to run this like, just like it is there and get up and deploy this in your organization, do some.

I mean, this is, this is, uh, nobody would ever do this, right? So then, of course, you need to bring this, make it secure. I mean, we have seen with Light LLM, how long is this now ago? Two weeks or something? You probably saw it, right? Like with this vulnerability that still all of a sudden steals all your keys and credentials and so on and so forth. Like, and that you don't want, right? If you're the chief information and secure the security officer in the company, right, you don't have a job anymore, right?

And that's, and that's, of course, another big challenge from an adoption perspective as well.

what do you? S uh, think the um, uh, function of a finance or an HR, or a supply chain team that would have been operating out of SAP in their day to day work? Um, you know, a a year ago, what do you think that looks like a few years from now? If, if they're successful, if you and your customers are successful with the um, AI transformation,

Yeah, I mean, 1st of all, it's very simple. They will get rid of a lot of the mundane work, right, like collecting information and preparing powerpoints for decision making and so on. So what we're going to see is a much, much faster way of making decisions, making better decisions, right? And then, of course, automating the mundane work. So what the people will do is they will run more scenarios, they will run, get better, deeper insights in a much faster way.

In order to then really think about, we always call it this more strategic thinking, right? And kind of in a way, Sarah, if you will. For me, this is the same way, like, like everybody who works today, maybe in a finance shared service center, right? It's for me the equivalent of a junior developer today with cloud code. So now they actually become they get one level higher, right? They're now not so much anymore. Uh, um, tasked with then writing a lot of the code, right, with with, uh, with codex or with copilot.

But they actually then start supervising the code, give feedback, right, and capture, of course, the essence of what the code should look like. And then, you know, do much more review. And then rather think about what to build next, right? Think about the next requirement. And how is that actually differentiating? So it will, like every, every role, every level will kind of get up leveled, so to speak, right? Because the the work that's being done today will be pushed down to, will be pushed down to these agents, right?

And thereby, therefore, I I think we, I believe, in general, what we will see is that, um, people will just achieve so much more. Because there is a lot of intelligence baked into the system that gets rid of of many of the things that we're that we're doing today. And that are actually, uh, well, at least in many cases, not a lot of fun.

I, I must admit my ignorance here. I don't, I'm. I'm thinking about this and I wanna talk a little bit about the impact on the business, if you're right as well. Um, I don't actually know how SAP prices broadly today, but the question would be like, how do you price? And if you are, you know, delivering more outcomes for customers or serving them, you know, service as software in a different way. Do you think that changes the business model for SAP?

it does absolutely. I mean, there's, there's no question, and we have prepared for this already. So for me, it was always very clear. I mean, for the most part, SAP software is seat based, licensed, uh, uh, uh, today, with a few exceptions, like a Concur or a Field Glass, for example, or the business network. Um, but, you know, very clearly, with AI, it was very clear for us that, you know, step by step, it will go towards this consumptive world, right?

At 1st consumptive, and then maybe in the next step, uh, once we have more verifiability in the system, then also towards maybe an outcome based, uh, license model. Uh, to, for example, what Sierra is do doing and so on and so forth. Um, uh. But the reality is also it is today for us. It's a hybrid model, it's consumptive, but it still has a certain element of seats in there and so on and so forth. Because also it's a joint journey with the customer, because the the customer is saying they are not yet ready in many cases.

Uh, uh, for a purely consumptive model, right? Because they need pre 1. Predictability, right? And then, of course, they are not yet fully, also everywhere, trusting the outcome, right? And or know then also, of course, is the value already there. But then they are afraid of that. The costs may explode from a consumptive perspective, et cetera, et cetera. So what we so what? At the end of the day, what we have designed is a hybrid that is basically ready for this consumptive world.

But actually meets the customers where they are today. Knowing that they demand still a lot of predictability in the enterprise space in order to cost control. Uh, the the whole thing for themselves as well.

That makes sense. It's unclear how I. I also believe that transition's gonna happen, it's unclear how quickly it will.

Exactly, no, absolutely, I agree. I mean, nobody knows this. And at 1st, you see customers that are more, they have have a wide range, right? They are also of opinions, right? And of course, some customers are a bit more forward leaning already, and then and oth others are more uh, um, uh, still, uh, asking or demanding, uh, a classical model, so to speak. And so therefore it's a, it's, um, it's a journey.

What do you? Um, uh, let me let me rephrase that for a 2nd. When when you look forward, uh, and think about SAP's position five years from now. And you compare it to the uh, broad market pivot away from SaaS and software, uh, in terms of just how investors are valuing these businesses and their enthusiasm about their durability. Um, my own opinion is like, the challenge is real. And yet it will affect these, uh, the the incumbent software companies like, very differently, right?

There will be winners and losers versus like, universally, like everybody's market cap should come down. What do you think? W is going to be characteristic of a winner? Or why does SAP get to in endure again?

I think at the end of the day, it's all about adoption and the outcome you bring to the customer, right? I mean, the technology, look, the reality is, for most companies, the technology doesn't matter, right? I I always tell tell to my developers all the time. Our job at SAP is to make the technology disappear, right? We need to get the outcome in front of the customer, and of course, not just the value itself. Of course, you also be able to produce and price it in a way so it's a win win situation for the customer, right?

And And of, and of course, the vendor, right at the end of the day. So what? Uh, what we are really trying to do, and this is also why we, you know, from an architecture, we are so flexible, right? We said, like, we don't over index on a specific, uh, we p. We have partnerships with all of them, right? And and and really, als only invest in the things that are actually differentiating for our customers versus the things that anyway will likely get commoditized.

Uh, in the in the tech stack. And then try to make sure that we, of course, bake the enterprise qualities in and the integration is there. And the customers can turn these capabilities on almost instantaneously in order to benefit from it. Why is this important? Because in order, because if you take a lot of time to reap the value, then your return on investment is essentially gone, right? Uh, or, or it becomes the the business case becomes harder, right?

And therefore, what we are really focusing on is to to deliver these outcomes for the customers. And I think that will differentiate the winners from the losers at the end of the day, to really focus on the business outcomes for the customer at the end of the day.

As we wrap up, I want to ask you, uh, a few quickfire questions, including a little bit more personal one. Um, uh, our listeners always wanna know, like, what do you do all day? As like, a CTO of SAP, like, can you just describe how you spend your time?

well, I spend most of the time reviewing, uh, the progress with the teams, right? And we're thinking along, you know, from all the layers with the teams, from the database to the models, right to the UI. Review the progress, give guidance, feedback, learn something new, study, of course, what happens outside. Uh, do a lot of prototypes, right? Where we speak here, I have a bunch of command line interface instances running here, prototyping a bunch of things, right, trying things out, uh, see what works, what doesn't work, right?

And then give this also as kind of inspiration back to the team. And then, of course, you know, um, yeah, work a lot with the teams, think about, of course, uh, the how we connect the vision then to the execution, and of course, work a lot with customers as well. So I love working with customers because, um, they always, you know, give you, they keep you honest in terms of, uh, what, what you can do already and where you can still get better.

Can you give us a flavor of? Just because you have such an amazing, like, global customer base, like something, um, challenging or interesting that a customer's asking you to do right now?

Um, oh, well, there are plenty of things, right, as I said, right, the predictive one is just immediately comes to my mind, right? Like, if you can improve the accuracy of the demand forecast by three, four, right, that is huge for them, right? This is this is multi million value immediately that you can deliver to the customers. Or I have many customers, of course, that come and say, like, Oh, Philipp, can you actually help me? Look, the commodity price is all through the roof, right?

Oil and gold and silver, right? Like, if you're now a jewelry customer, right, you, you, you with jewelries, right? So you, you have a challenge, right? You have a challenge because now your your entire business model, you need to think about this. And then you discuss with the customers coming from the business challenge. Hey, how can we actually help you to address this? Because now they need to change their manufacturing, right processes and come up with new products.

And then the question is like, how can you help me research the right products, find product market fit? And then, of course, out of that, then determine, uh, how to actually, you know, source new suppliers, right? Because, um, uh, because maybe I need to source different materials, uh, from from different suppliers. So, like, these challenges are real in this ever changing world, right? And then for some people, it's, it's, it's, it's other things, because they are challenged, uh, on on others.

So the the world is very, very diverse in terms of business challenges that customers have out there.

yes, and I mean to your uh, uh, so something you mentioned earlier, uh, you know, being the partner that folks come to and say, you know, uh, how does my business change? Given what's happening in the state of Hormuz? Is a hard question, right? Make it better, make it, make it easier for me to navigate around it. Um,

now, but but that's an interesting one, right? I also look, I go into every, even if I'm the CTO, you know, uh. And usually people then t expect from me. I tell them a lot of technical things and so on and so forth. What I learned actually, and I did this mistake, probably more than anybody else in this world, is to kind of pitch the technology, right? And this is completely wrong. When I sit together with CFO or a CIO, the 1st question is like, Hey, what's top of mind for your business?

What are, what are current? What are your current challenges, right? And then work backwards to the technology, right? And that always, I always found that this is the most useful approach.

hm hm. And then, uh, last question for for you personally, outside of everything that you're doing at SAP, uh, just as a technologist. Like, what else are you interested in and paying attention to, or in in tech or AI? Or like a a belief you have about something that's gonna happen?

Well, I think, I mean, obviously AI is the dominating thing because it's so, uh, pervasive and ubiquitous. But um, no, I think what we also are very, uh, very excited about the work we are doing. Um, finding an in in the quantum computing space, um, trying, like to find new algorithms also, um, there, because, you know, programming quantum computer is very, very different, uh, compared to everything else. Uh, that's in a very early research stage, but that's also that's super exciting as well.

A And why is that like, uh, commercially relevant to you? Like, what's the, you know, to your point of solving backwards from the the business problem?

Let me put it this way. The hypothesis is that, of course, once the hardware matures right in the quantum space, there are certain problems that you can address, uh, that are hard to to address today. Uh, what we are focusing on is the optimization domain, so obviously. And then if you go into, uh, think logistics, traveling salesman problems, knapsack problems, like all these kind of usual, uh, hard problems in computer science. These are interesting problems where we, where some people, where we believe that could be interesting for the future, uh, for maybe different kind of computing paradigm to solve for.

And we try to be, uh, a learning, so to speak, and we will be very hardware agnostic in that sense, right? Uh, like, like, SAP always, always ran on different computers, uh, in this world, um. But what is important to us is that we find early on already, uh, and of intellectually, but then also find, uh, these, uh, these new algorithms that then can propel that forward. Because if you can obviously load your trucks right and and do the route planning even more, right?

The the outcome is same, the emissions go down, right, and you save a lot of money and so on and so forth. So there's a lot of things you can still opt that, that, that today, of course, with the limitations you have, you depending on how large the problem size is. You can only, you can only approximate, uh, and then need need to live with the with the best solution you can get in any finite amount of time.

yeah, I think that's a great note to end on, just because, um, it reminds, uh, everyone, including me, that actually there are like interesting computer science problems everywhere in the enterprise. Uh, where, uh, I I think a lot of people, um, if they haven't worked on these problems of scale with real customers, they, uh, they might assume that. You're like, you know, like, building a CRud application is pretty easy at minimal scale. In 2026, it's actually remarkably easy with coding agents.

And it is, oh, absolutely, and yet there are, um, there are problems where we are limited by algorithms and computation everywhere, right? And it takes some imagination to just go attack them.

and it's also, to your point, right? Yes, a CRud application is like, that's, that's, that's a solved problem, if you will, right? But most of the time, uh, it's not just building a little CRud application for a data object in the in a database. I mean, that's just, uh, uh, software is a little bit more complicated than that usually.

Hm, hm, uh. Thank you so much for the time, Philipp. This was great.

Yeah, thank you for having me.

Find us on Twitter at No Priors Pod, subscribe to our YouTube channel. If you wanna see our faces, follow the show on Apple Podcasts, Spotify, or wherever you listen. That way, you get a new episode every week, and sign up for emails or find transcripts for every episode at no dash priors dot com.