SAP Sapphire 2026: post keynote Q&A for media and analysts

Please welcome to the stage SAP Chief Communication Officer Monica Shaw.

Good morning and welcome to our fourth keynote Q&A here in Orlando. Thank you very much for joining us here in the room and of course thank you everyone for joining us virtually. I really hope you had the opportunity to listen into our keynote today and hear about our exciting news. If you haven't, the format today is really designed for you to ask questions and with that let's not waste too much time. Housekeeping rules: we are going to live stream the session as always. We are of course on the record and you will have the opportunity to ask questions in a second. But before that, let's welcome today's panelists here on stage. Let me kick it off with our CEO and chairman of the executive board, Christian Klein. Muhammad, member of the executive board for product and engineering. Sebastian is a member of the executive board and chief operating officer. And last but of course not least, Philip, CTO and extended board member.

So before we kick it off with questions, how did it feel?

So basically for everyone who hasn't had a chance to watch, what are the three most exciting moments for you?

Thank you very much. So Sebastian, we got different questions in the same direction. Let me summarize it one way or the other. So the question is, with AI and on-prem customers and us having said we only offer AI to cloud customers, there are questions: is there a confusion? How are we changing strategy? Why have we done this? Would you mind explaining it to us shortly?

Yeah, for me, look, there's no confusion at all. I mean first of all, be super clear: our Joule assistants and agents are designed for the cloud. I mean this is the destiny and the destiny remains absolutely unchanged. However, we also want to meet our customers where they are on their transformation journey and that's why Philip and the entire teams have put in a lot of work technically but also Thomas and the service team to enable the majority of our Joule systems and agents also on premise, on ECC and S/4HANA on premise for customers that have already committed the majority of their landscape to the journey as an interim solution so that they can benefit from AI while they are modernizing. I think that's absolutely the right thing to do and I'm excited to see customers taking advantage of this.

Thank you. I'd like to open it here in the room and if you have a question, we have two mic runners. Please raise your hand. Also please say your name and the company you work for. It's pretty tough to see you with the lights so be patient with me. I think Kakan, right?

Yes, I waved to you. Hi, Kakan Kunduraci with H&M. My first question is for Muhammad. You've described a future in which enterprises can automate far more processes with the autonomous enterprise, but I think many customers have already experienced gaps between the demos that you can see on stage and the reality in their companies. So what makes you optimistic that you can really deliver the results and are there any data points to support this? And also, what are the main obstacles that you see in companies? And a question for Sebastian and Christian: there has been some M&A activity recently in the past few weeks. Give us some more color on whether you're planning more deals in which areas. Thank you.

Yeah, thank you for the question. I think it's a very important question, right? Because there's a lot of feedback around our Joule and what's in the market today and the gaps that are there. But I think that's the thing that hopefully showed well in the keynote today: we started with the platform because we had to go build that and this is the platform that we built the autonomous suite natively on. Obviously, we feel really good about the innovation payload that we just talked about and on most, if not all of them, they're in beta now that we're working with customers and partners. And I think there's a slide that Philip showed as well where hundreds of partner developers have actually experienced it. The demo that was done was also not by SAP, it was with a partner on our real product truth that was there and then we built the agents on top of it. And then the customer stories that we shared aren't stories that we made up effectively, right? These are stories and customers working with us on the early release of the product as well. So what we feel really proud about is the product truth that we talked about in the keynote is beginning June, it's going to be in early access which effectively for us is general available bits if you will. We just want to continue to learn as much as we can with now a broader set of customers before we officially declare GA. But there's nothing that we showed you today, unless we explicitly called it out, that isn't ready to be in the hands of the customer today. Now it is fair to say that hey, there isn't public knowledge about the new Joule Studio 2.0, what we're doing with the new Joule spaces and the new conversation layer with the cloud-based harness if you will in the autonomous suite. So the feedback that's there now, at least in the public domain, is based on obviously what customers know. But from here on out, as Christian pointed out, we're going to go really aggressively with our customers to make sure they can try and scale up, getting their hands on the keyboard. Because one customer sort of told me this a couple weeks ago, which I think was the mindset that as SAP we're operating in, and when we talked about our financial close assistant that hopefully all of you saw in the demo and we said, 'Hey, this is something here that we can go deploy with you.' And the customer response was, 'Well, is it really there? Because if it's in 3 months out, I'm just going to go build it myself.' So that's the sense of urgency that this new SAP that Christian talks about, we're all working with. And that's why what we didn't want to go do is get on stage and say, 'Hey, this is coming in Q4 or this is coming in H1 of 2027.' Hopefully you saw none of that in the keynote. All of these things are things and the customer stories that you saw are what we're working with now on the platform that we just outlined in harmony to you guys over the question.

Thank you. Philip, I'm going to be with you in a second question here, but I think I had one question here in the second row.

Hi, Maribel Lopez, Lopez Research. Thank you. You've done a lot of work on the AI platform and the contextualized part of that and you've had acquisitions. When you're talking to customers or when you're thinking strategically about how you're going to advance the contextualize and reason portion of that pillar, what are the things that you feel you have to deliver over the next year to give your customers a best-in-class contextual data experience?

I can start. I think listen, there was this slide that we used for those that saw the keynote, right, that said there's a public model and those are getting smarter every day, right, we use the best of what's out there. And then there's the context that SAP provides, a third circle that says your company context. So in this middle one we continue to sort of put our knowledge graph, the deep understanding of our Joule, so the knowledge graph, that's the semantical model, the richness, the authorization into on top of what the public model is. The process knowledge part of it is out there as well. We've also now talked about Joule 1.5 as well as our intent to acquire WalkMe that continues to sort of make the middle layer much, much stronger together with the public LLM. The third thing that we're going to continue to build on and some of this you're going to see over the course of the year is the knowledge graph that we have is of course the canonical knowledge graph of our products as well. What we want to make sure is most of our applications customers extend as well. So it needs to by default out of the box understand the customer knowledge graph that has been extended, added to with the partner solutions on top of it. And then you saw company memory, that was probably one of the few ones that's in beta, not in early access beginning June, but we believe that's going to be a phenomenally powerful one as well because that continues to build the customer context as well into the whole context. And then of course I think you can continue to expect a lot more progress on our predictive models. That continues to expand our context layer in combination with the public model to be able to build human ability.

And maybe just to add, of course clearly, I mean that's more for me the kind of modeling side of the house, right, where also with all these additional agents we are building in BDC helping the users, right? Because if you have SAP, non-SAP data, it needs to get harmonized. Relative of course for the master data is one answer but of course there's also so much more data, right, where we're ready to model your specifics and so on and so forth. And a lot of investment goes to making that much easier. Also of course using AI and AI enrichment to harmonize and semantically bring the data together. And then for sure, I mean we are still in, we said Q3, we are aiming with the closing of Truvio and of course not just the modeling part at design time but then at runtime, right, this federated lakehouse architecture that allows you to keep the data where it is, have full compatibility. That also then, like in the technicalities, right, where you don't need to map between different data formats and so on because the whole industry is going to Apache Iceberg and we are doing the same as well for full interoperability to make it much easier that also no matter where physically the data resides in the enterprise, right, doesn't matter whether it is on an Azure data lake or sits in S3 or whether it's just a good old SQL server, could be a good old Oracle database that sits on prem, right, like Dremio can federate all of that together into BDC with a very competitive price performance. And I think that's also a very important piece of the whole puzzle.

Thank you. I'm going to take one question now. We received one for you Philip and then next question. First row, second row and then back. For Philip, you have introduced a new SAP API policy. Critics say this is just another way to restrict access and push customers into SAP-controlled integration or AI stacks. What's actually changing for customers and what is your principle here?

Yeah, well there's much to be said, right? We have a very long FAQ that goes into all the colors and different variations of that. But look, at the end of the day the API policy in its essence is just what everybody does in its industry standard because what we have been saying is on one hand side yes, we need rate limits for example in certain areas and then many of our applications, many other applications in the world have these rate limits already because you need to protect it and of course with the ever increasing workloads that are coming in through agentic workloads and so on. So that's just something that is expected because we have seen outages, we have of course seen also certain attacks which we could prevent but there's always something we need to harden in that regard because there are new workloads which we need to deal with. And then on the other side yes, clearly there is also misuse of certain APIs that can be used that have never been designed for that purpose partially and also sometimes APIs because we just need to move on. Certain APIs, we have this with the clean core policies already in place forever because you know you need to move on to new protocols like Iceberg, like A2A and so on because these are the new channels, governed, well-defined channels and not like the old stuff that keeps just everybody awake if it does not produce any meaningful goods. Things need to move on, right? And I mean it's just normal in software that some old APIs just die out and then we move on to new ones. I mean this happens on an iOS platform, it happens on any Java platform, this is just software as it progresses and evolves.

I think the thing that I would add, right, that I think the narrative out in the public domain misses is, you know, there's a line I said in the keynote that says not all SaaS applications are created equal, right? So if you look at sort of the level of complexity that SAP applications have just from a logic perspective and a compliance perspective, auditability perspective, statutory compliance and different regulations as well, if you just go access that, you don't really understand sort of the engine of what it's doing behind the scenes. And what we're saying is listen, just to be able to sort of connect data, do some integrations and things like that was the purpose of the APIs that were out there, not to sort of now build systems of reasoning on top of it by directly calling something that you don't understand into the black box what it is. What we're saying is the better way for you to go do that is if you want to go invoke that is through agentic orchestration because what we're bringing is that logic, that reasoning in a safe, certified, compliant way that you can rely on to be able to connect with any other agents that are out there as well. So if you think about it from that lens as to, hey, SAP has a responsibility to make sure that the compliance, the auditability, statutory requirement, all of those things work, then the safer way for you if you don't understand the 50 years of logic and the code that's out there in some industry complexity is for you to access it through the orchestration layer because now we're surfacing that through Joule in a way that you can rely on the same way that you can rely on just connecting and integration from here to there. That's what those things were for. So I think there's a fundamental difference between sort of a level of responsibility we feel to be able to make sure our customers are still complying. They can rely on what comes out of the agents and so forth. This is what Christian talked about, like you can't have three years on your financial report because that's not going to go too far and you're going to be out in public domain. You have to be accurate when you go do your flows. You can't just be purely probabilistic in it.

Thanks so much.

Hi, Alexander Zahra from itnews.com in Australia. Early access starts next month but in this age of AI everything's time compressed. So if we look at Sapphire 2027, you know, all of this will have been in place for a year, the autonomous enterprise will be humming along in lots of different places. So what is it going to look like then? I mean obviously you can't give Sapphire 2027 announcements now but you know, on the Starship Enterprise space was the final frontier but with the autonomous enterprise what is the equivalent? Or does AI eliminate the idea or the concept of anything being final ever again? I mean, you can't spell final without AI. So, what are your thoughts?

I mean, I think in general, I say this always. I think in settings like this, I tend to resist the urge to look into a non-existent crystal ball of predictive future like some of the other tech pundits do, right? I mean, it sounds really good. You can get a headline out of it, but the world changes. Where we are, listen, we have a vision, we have a view. We did talk about where we see us as SAP investing a heck of a whole lot more is I think the question that Liz asked which is this context layer because that's really where the differentiation and the value generation is going to be. So you can for sure expect to hear a lot more advancement on how we bring the data, the process, the code differentiation that we have alongside predictive models and things like that that now unlocks far more newer value-based scenarios for our customers. So that's a given if you will. But then just from a time frame perspective, the thing that I'm excited about, again, going back to Kakan's question, listen, we're excited that the things that we talked about are goods that are on the truck now that our customers today, even as we sit here, are touching and playing with. And with broader access we give you in June. But I'm also excited about the things I hope and expect Christian's going to launch via LinkedIn over next month and the month after and the month after because I think the world around us is changing in a monthly and a quarterly way. That's why we didn't want to announce things that we think might be worth waiting till July with a LinkedIn post from Christian as opposed to cramming everything here for a year. So I do think we're taking a planning horizon and making it actually very fast, very innovative with product truth that's there for customers to go touch on. So I'm excited about those things. I'm not a prolific social media user as some of you might know. I mean, I apologize for the disappointment you might have had in following me, but I think Christian does post a lot. So, please stay tuned to him. I might like it. So, if you're following me, you might work on that.

I mean for a minute more, I think Christian talked about the portfolio is sort of a new SAP, right, the way the agentic experience in the platform is all new as well. But I think there's a big part of how we work is also a new SAP. So most of what you saw here is largely a lot of innovation that's been pulled together in a number of months, a small number of months, in 3 months. That to me is far more important because the world around us is moving so fast and we need this ability to react to say, hey, something new that's come out needs to be in the platform and how can we combine that with our context very quick to even far more value. And that confidence, that new Joule, the how part of it going back to Kakan's question, to me is far more powerful. And this is where you're going to see kind of SAP now changing differently to say we think this is now possible with the tech that's out there and here's SAP's answer to it with our context there on our platform for our customers to be able to rely with the strong governance and enterprise-grade qualities that they expect. I think that to me is probably sort of the subliminal but underlying message here in this event that that's how we expect to operate going forward.

Did you want to add? So, front second row, blue, and then on that side.

All right, I'm Alves from Estadão in Brazil. Well, we have seen in the introduction video in the keynote that new technology they use to face resistance but seems that artificial intelligence is quite different. Everyone is rushing to adopt that and sometimes, well maybe as a side effect, companies are still struggling to make real money or make any money from that. So I have two questions. Why artificial intelligence is different from any other technology before in this speed of adoption? And how SAP can help companies to properly use this technology and move away of just piloting or exploring, trying to find ways to use that to make real money and transform the business?