Hey, welcome everyone. We're gonna go ahead and get started. Thank you for joining us. My name is Megan Hostetler. I'll be your host for today's event. A little bit about me: I lead a content marketing team at SAP. What that means, I get to dive into research on marketers and consumers from really across the globe. Helps me to understand the trends and what's shaping the future of CX. But enough about me. We have a lot to discuss today, so let's go ahead and dive right in. All right, so today I will be joined by a panel of experts from SAP, Google, Infosys, and this gentleman, the gentleman here. I'd love to get to know you guys. So if we could, let's do a couple of quick rounds of introductions. So Lucas, I'm gonna pass it to you first. Could you tell a little bit about yourself and your role at SAP? Yeah, absolutely. Thanks, Megan, Hey everyone. Lucas here out of Berlin, Germany. I'm heading our strategic ISV partnerships department here at SAP Engagement Cloud, working quite closely together with Google and other partnerships. And here today specifically to explain a little bit on why, what makes AI at SAP so unique and what's the additional value proposition you can get of utilizing our SAP AI capabilities. Very much looking forward to it. Rouzbeh, over to you from Google Cloud. Thanks for being here. Thank you for having me. And yes, my name is Rouzbeh Aminpour. I am one of our solution engineering leads here at Google Cloud. I heavily focus on, as I oftentimes call it, the translation of the core products that we deliver to the market and the real-world use cases that you as marketers care for. So attaching the sum of all the SKUs together and ensuring that those use cases can come to life at the highest level of quality that's needed. My main focus at the moment is our media generation stack, so anything that has to do with image, video, audio, or music production. Looking forward to the session. Thanks for being here. Arjun, do you wanna wrap us up here and tell us a little bit about yourself from Infosys? So I lead SAP Services, which basically means I am the one standing between a great roadmap slide and the customer's messy data. My job today is very simple, to tell you whether whatever we speak actually holds up once it's live, and to tell where we have already done this elsewhere. Looking forward to the discussion today. Awesome, thank you guys for being here. We have an amazing panel here. We're gonna get into all of the good things with AI and bringing that to life. Before we get started, just wanna give you an overview of what to expect in the next hour here. We're going to dive into what opportunities AI will bring, but also the urgency behind it. We're going to talk about how to harness that potential of AI, and we're gonna bring in the help of Rouzbeh from the Google Cloud side of things. We're going to talk about how to apply AI with proven models thanks to partners like Infosys, and then how to accelerate AI adoption and think about what actually is going to stick and move us away from that pilot mode into scaling AI across your business. So a lot to tackle in an hour, but I have full faith that we're gonna have a great dialogue with this group. To set the stage, though, before we dive into that panel discussion, I want to just leave you with a few things to think about before we dove in. When we were thinking about this webinar, what we were going to shape this, what kind of content we wanted to include, we wanted think about our own customers at SAP. What kind of things are they facing, and how are we helping them? And what we see time and time again is that these businesses are operating in a landscape that is full of volatility. When I say volatility, I mean things like economic pressures, technology, and behaviors that are constantly changing. So for example, we're seeing a shift in consumer behaviors, of course, when it comes to AI. 30% of consumers are already using AI agents to make decisions and act on their behalf when it come from buying from a brand. That's huge, and they're already moving to 30% of actually making those decisions. So consumers are moving quite fast when it comes to AI. So not only are organizations adopting AI, but consumers are quickly, if not more so, adopting AI faster than businesses today. So what we mean by this and why this is causing the urgency of AI is now is the time to take advantage of what it can do for you and how you can meet and exceed expectations. And in this complex environment, what we'll talk a lot about today is content and engagement and what that means when it comes to that reliable source of truths, if you will, to win customer loyalty with that AI-backed engagement strategy to build relationships that have potential to grow even stronger when it come to AI or risk a fast decline if we're not adopting AI fast enough. And we know this can be a lot of different factors that come into play of why people aren't adopting AI. So what we wanted to do and what my team focuses on when it comes to content marketing is research. So we thought about why engagement, why now? Why is it going to be so reliable? We wanted to some global research that explores engagement maturity to help you understand this complex volatile landscape. So what this research report does it introduces a new engagement maturity score to determine that potential to embrace engagement across your business and the impact that that would make on your outcomes. So what I wanted to do is share a few main takeaways that we uncovered to really set the stage for this panel discussion. I'll give you my quick takes, and then I'll invite our panelists on for an interesting dialogue around all things AI. So. What we found from this research is that marketers, you know, we've been doing this for a while now. We've been trying to find the right message, the right customer, the right product, at the right time. This has been really the North Star for a really long time, but it's been conceptual until now. Finally, AI has put all of this within reach, but what's holding us back? There's gotta be things that are keeping us from moving and scaling AI across our business. And it isn't for a lack of trying. In fact, we've seen a lot of marketers go into pilot mode and trialing AI where it makes most sense in their organization. But what we're seeing is that marketers are being forced to move fast, and that's not new, but the intensity of that fast moving adoption is there, all while they're operating in systems that are working against each other. So things like a 7 to 14 day lag between campaign execution and then actually being able to see and use these insights. Or 60% of CMOs say they're unable to act on business signals in real time. Or only 18% with access to real-time operational truth. So while leaders are now facing the pressure to solve all of these issues that we're facing, like fueling loyalty, driving sales. Plus the pressure to learn and adopt AI along the way keeps them stuck in this pilot mode. So the good news is if marketers get this right, we know that enterprises have the potential to successfully scale this across the business. And those that are successful see three to five times increase in higher ROI. And marketing and customer engagement teams expected to be two times more likely than other functions to report early ROI from scaled AI initiatives. So that's why we're all here today, guys. That's why invited these panel of experts to get you from experiment mode to fully fledged implementation across your business that's gonna make an impact. So imagine launching a fully personalized multichannel campaign in hours, not days. Connected customer data, generative content, intelligent agents, all working together while your brand is guardrailed and stays intact and your team stays in control. That's not some future state that's over here in La La Land. It's what we're gonna show you today. So for SAP customers, success with AI isn't new, it's already happening. So for us, it's not an experiment. It's the baseline. Over 62% of our customers are already using SAP Engagement Cloud's AI powered features. And nearly half have gen AI embedded directly in their workflows. And when it comes to AI as a baseline, the opportunity shifts from being an automation, which is kind of the corporate standard now, to autonomy, where assistants and agents don't just execute on these tasks, but they coordinate outcomes together across your different systems. So with that, I hope this set the stage and got you so excited for this conversation. So let me go ahead and invite our panelists back up to this virtual stage. We got SAP, Google Cloud, Infosys, who is going to share what it actually looks like to move from AI experimentation to enterprise-grade impact, including some live demonstrations to help you bring this to life and make this possible in your organization. All right, so we have a couple topics we're gonna dive into. The first one is talking about harnessing the potential of AI. So Rouzbeh, I'm gonna bring you on first. So welcome to the party. You work on Gemini agent platform, generative apps and agents. What a tongue twister. And you work at Google, which is very exciting. A lot of cool, fun stuff that are happening, right? And so for most people who mostly know Gemini as more of a consumer product for the day-to-day user, what does the enterprise side of Google AI Stack actually look like today? Yep. Thanks, Megan. And one thing to call out, as you mentioned, the consumer side of Gemini and saying how AI is here today and now, just a fun side story is it's not just here for us, but the entire next generation is learning and experiencing it. And that's the world that they understand. And if you think of marketing as a mechanism for us to communicate, right, as humans cross-communicate with each other. That is the language that the next generation really truly understands well. I look at my own son and how he interacts with AI to actually learn things, explore things, choose what things they would want to buy, like fishing pole or whatever, and the amount of research they can do. We just talked about Legos, just want to call out your son's Lego collection. Legos too, exactly. So there's a lot of interactions that happen on that consumer platform. However, when you want to look at Gemini from an enterprise perspective, there are some differences, and those are very important differences. One of them is each of your companies have different needs, every company is different, the way you interact with technology is different. And so you demand that level of flexibility from these technologies to be able to deliver against what you would like to deliver, whether it's your brand promise, whether it is the way you like the messenger, your content, the style, and also the workflows. And so if you look at Gemini from a consumer application perspective, it's made to be as seamless as possible and give you the most cookie-cutter approach to experiencing AI, meaning a six-year-old can experience it just as much as I can as an older person. But at the same time, you want those SKUs to be available to your company so you can build those experiences that you like. And so that's what we focus on heavily is to bring the building blocks of generative AI, meaning having SKUs like Nano Banana for image generation, Veo for video generation, more recently Omni, which allows you to bring audio, images, as well as video in and have any modality of your choice. That actually stands for omnimodal, so image, video, audio in, and image, video, audio, music out type of a conversation, and then wrap those SKUs with the types of infrastructure and database platforms that are needed for you to be able to achieve those results that you're looking for. So a multimodal experience with those Lego blocks that are needed for us to achieve the level of fidelity and results that an enterprise would expect, and on top of that, with an enterprise, not only is the experience different from a business perspective, but from an IT standpoint, there are some needs as well. Data protection, data privacy, grounding of the information against our enterprise. And so those kinds of security, safety rules are also set in place. For example, in the media generation side of the fence, you would want to make sure that your AI capabilities that you're using in your enterprise don't expose any kind of videos that would have, say for example, violence or any kind of like pornography or any of those kinds of things that would maybe get you into hot waters with your customers, but also copyright protection and making sure that whatever videos you're producing are actually made with the types of data and the infrastructure that gives you copyright protections. Outside of that, you're also looking for the types of experiences that are more sophisticated. You're not trying to just send an email or one campaign for one video as the consumer side would be. You want to scale these things so you can actually hyper-personalize to that promised land of hyper-personalization that we've been talking about in the marketing segment for years. You can now achieve those. And so the infrastructure and the AI capabilities that Google is developing for cloud customers and enterprise essentially bring the types of capabilities and feature sets that are needed for you to deliver those kinds of promises and expectations. Yeah, I just love that we're continuing the Lego conversation. We'll just continue that this whole time of building those blocks for a foundation for AI across your company. So speaking about AI across your company, there's a lot of different industries out there that are probably thinking about AI in different ways. So tell me about when you think about horizontal capabilities that could work for any kind of industry, whether it's retail, financial services, healthcare utilities, what does that look like reusing some of those capabilities here? Right. And this is one of those areas where Google really excels, which is in order to have a very horizontal and very complete AI capability, you need to have a solid foundation of data. And this is where Google, and you know, it's very impressive when you look into Google's Knowledge Graph, for example, and the sources by which Google has been acquiring data just so we can build the best of models, whether it's licensing, whether it's data donation, web contributions, generation of data. There's a lot of places. I'm sure we've all seen, for example, those videos, Maps is a very visual way of understanding the concept and the power of Google's Knowledge Graph. But you've seen humans that are walking with backpacks with the cameras on top, climbing the Himalayas, for example, or you see these vehicles that are driving around the streets trying to map the world around us. That same type of a mechanism is happening and that significant investment is happening at the Google site to make sure that we do have data for all sorts of horizontal applications, but also meaningful information that applies to vertical specific. Um, just last week I was at a conference, and it was around security, and someone mentioned executive security, and I had literally no information about executive security. How do you protect a leader? And I was able to go and interact with Gemini and actually pull the right information. And I'm sure you're all using that at your consumer side. What plant is this? What is that white stuff on the leaves of my flower that I love so much? That information helps us build capabilities that can span across industries. When you look at it from a marketing standpoint, marketing isn't just producing an image. Marketing means that you have to have and be context aware of what goes inside the image. Marketing means you want that image to be as realistic as possible, right? And so when we build these tools, we make sure that we have the right amount of data that helps us produce that image to as much realism as is expected from a marketer, but also in most recent releases, say, for example, we released Gemini's Omni model, that actually brings the laws of physics into the videos that we produce. So when you pop a bubble and you give instruction, "create me a video of a bubble that bursts," the model actually understands the concept of a bubble bursting and how gravity affects the tiny molecules of that bubble, the water that's bursting in, and what effect it has before it actually falls to the ground. That is what's needed for us to produce the level of quality of videos that every single one of you as marketers expect. And so we focus a lot of time into making sure that we can produce the best of class of those multimodal models. But also as time goes by, what we're finding is those models are great, but the expectations for marketers continues to go up. So we say, fantastic, you are able to produce that photo-realistic video in 10 minutes, and it's polished and ready to ship. Well, it doesn't stop there. We don't get to go and drink some more coffee and have hallway chatters. What that means is, can we get closer to hyper-personalization? Let's repeat that experience for every single customer of ours. And so now you have to be able to scale. And to scale, you need agents. These agents are basically your assistants, your workers, and you can build those, and this is where we're spending quite a bit of time on our agent platform, and one of the user experiences that we're bringing forward is Gemini Enterprise, for example, where you can develop your agents to be your worker bees that actually help you achieve far more than what you could do if it was just you by yourself. And yet the expectation has continued to rise as far as, produce more marketing campaigns, increase the ROAS, and you're not getting additional head count. And then the last part is these agents are now coming in with far more capabilities as far as reasoning, meaning understanding of the world around us so that they can actually deliver better results to you. So when you're doing your market research, say in a travel agency, you want to be able to say, I want to micro-cluster all the customers who are like Rouzbeh who do spend quite a bit of their income on Legos, for example. And yes, Rouzbeh in Seattle would like some sunshine in the middle of February. If spring break is coming up in a month, let's hit him with advertisements to San Diego with Legoland. And that advertisement would go only to Rouzbeh and nobody else because it's made just for me. And that level of reasoning and understanding of the concept of rainy season, sunny, spring break coming up, the combination of all of those, and those agents having the capacity to actually support micro-creation of those advertisements becomes the next wave of the innovation that we're seeing in enterprise. It's, I mean, it's an exciting time. And if you think about marketers that are working for these enterprise, they're thinking about that attention to detail, that bubble example that you mentioned. I don't think people are thinking about that every day of all the details that go into creating that. So, you know, when you think about these enterprises that might be skeptical about what's real and what's not, what's still an experimental, what's like enterprise ready. Could you speak to like where we are today? Yeah, and I've been part of the AI world for close to two decades now. So what I say, I say it in a very humble way from a personal experience. This is the first time in my career where I can stand and speak to business leaders and know that the technology is actually available and ready and can deliver the results that's expected. And yet, for the first time in my career, I'm seeing business processes and organizational adoption of the technology to be lagging behind the technology. If you look over the past decade, it's always been marketers ask for hyper-personalization, email campaigns that they could send out to each one of those customers like the example I just gave you for the travel use case, but AI was always missing something. Even if you go back to two and a half years ago where the AI models would generate a video of a human with a third arm sticking out or something was off, and we've come a long ways. These models are now available, and they're ready. I think what needs to happen is a lot of focus onto process, operational changes, adoption of this new way of thinking of what do we do with AI and how much can we speed up our operations? And then the last part is, we, from a hyperscaler perspective, produce the core SKUs, going back to our Lego example, if we have any executives or marketers from Lego, you now know one family has quite a bit of income and the advertisements work. I think there's one in my bathtub somewhere, don't worry. One of the most important parts is, a lot of companies don't necessarily have the technical resources to adopt AI capabilities. So what they're looking for is solutions that actually adopt AI in them, so seamlessly, I can just go use that tool and the next morning I see a lot AI features available. And so what you see from our perspective, because we are building these technologies as horizontal capabilities, the way for us to succeed is to actually partner with the right folks who have the right level of adoption and understanding of the different vertical experiences. And in marketing, for example, we've had the great pleasure of partnering with SAP, where we can actually bring the Lego blocks to the table and Lucas and I, for examples, have had a lot of good interactions as far as how do we make it into, for example, our Engagement Cloud experience. And it's humbling when we actually see the Lego blocks come together with companies that have the capabilities and the competencies to actually adopt these tools, but also more importantly, fuse them into the existing experiences that customers are used to leveraging. And that helps with the adoption, and that reduces the pain from our enterprise customers who want to adopt these technologies. The intent is there. It's just the change management that needs to happen. And so when companies like SAP come through and they deliver these kinds of experience, seamless and the existing tool sets that our customers are already used to, it makes for a great experience overall. Well, love to hear that. I think what we'll pivot to next, if that's cool with everyone, to how we can apply this AI. So, Rouzbeh, you just walked through the breadth of Google's Gemini agent platform. Lots going on there, a lot of exciting things. And now Lucas, let's bring you back up. Rouzbeh set us up really well of that partnership between SAP and Google, but why does that matter? What does SAP add to that foundation that Google is building? So, Lucas, welcome. Yeah, absolutely. Thanks for being back on the stage here. Yeah, I think Rouzbeh called it two really important aspects there. One is the data component that Google is already using to build some of the foundational Lego blocks, how he referred to them. And he also said all these businesses out there and probably all the businesses from all the attendees on the call right now are quite individual and unique to itself. And, as we all know, within the realm of talking about AI now for the last couple of years and decades, there's always been this concept of the quality of data that you put in very much determines your output and the results. And Rouzbeh also talked about everything being very much context aware. And so I think what ultimately SAP is then bringing to the table when partnering here with Google on some of these key capabilities, not just uniquely within marketing, but even the broader customer experience space, is really the business-relevant context and data components that SAP delivers on behalf of the customers working with some of our products. And so if we look at some of the key components of quick partner plug here to kick things off on my part is what we really kind of released early on this year within April is the three different components of how we really partner together with Google. There's the component of the foundational layer, the data aspect of it, where if you look at all the different data products and the business-relevant context that SAP generates throughout the breadth of applications that our customers use, whether it is ERP, it's the different components within our customer experience stack when it comes to commerce, marketing, sales, and service. But then also partnering with the side on Google to then say, we can exchange the different data sets in a zero-copy way to then also either get customers access and combine the different data touchpoints and products from SAP with the non-SAP data products that would typically sit on their Google environment, things like advertisement KPIs and detailed information on trends, on search, weather, and even geo-like information. So even having that exchange to really have consolidated data foundation that your AI can work off from to generate the best possible results to move into that hyper-personalization that Rouzbeh already also mentioned. The second component is then really kind of like the channel component where we've been partnering with Google for many years now, across things like mobile push for in-app, and push advertisement, we have mobile wallet for customer identification in-store promotion management, but then also the entire advertisement space. Then lastly, we also rolled out RCS as a new channel for marketers to really communicate in a two-way way for their consumers. And all of these two components are then topped with the AI layer where we're really bringing that into action, what Rouzbeh just explained perfectly in the infrastructure and all these technical capabilities to ensure that enterprise can really run those AI use cases and activate them in a way that they're already aligned with all these enterprise requirements that he spoke about. And so we're utilizing the technology, let's say under the hood, but we've built them into some of our key components in a way that we're not just purely adding the business context, but from also understanding the workflow and the processes of how our users really use our tools day in, day out. We are on our track to really then ultimately giving them a chance to have those processes run completely autonomously. I think that was also a great analogy when we spoke internally, and kind of preparing for this webinar, where Rouzbeh said it's the same with autonomous driving. We've seen all these little milestones in between to really get us there to have self-driving or fully autonomous driving cars. And I think we're still in this phase right now where obviously we're releasing those agents that ultimately can take on requests from marketers and solve them autonomously in the background. But at the same time, we have those embedded AI capabilities that need to be fully context aware. And just because this is such an important thing, we've also prepared here another slide that really just speaks to making everyone from the attendees here understand the breadth and the depth of all the different context-relevant data and information that SAP brings to the table that makes the output so much more unique and so much more tailored to not just the specific business of our customers, but ultimately the end consumers. And so from the breadth of the ERP back office data from things like finance, fulfillment, pricing, inventory, onto the specific products where Engagement Cloud also sits within the customer experience spaces when it comes then to marketing engagement to say, what are all the different KPIs and engagement touchpoints from consumers while they're engaged with all your digital marketing channels, right? Relevant information for AI and agents to understand in order to fully personalize and hyper-personalize the outcome of it. You have your campaign history, knowing what has already performed really well in terms of editorial content, visual content elements, hero images, product recommendation. These content elements also, when it comes to the whole orchestration and what channel works the best, all the way to loyalty, product, and sales information. And the output is quite, let's say simple, right? It's still the same vision that when we talk about hyper-personalization within the space for quite some time, but specifically also when we still look at some of the, let's say very labor-intensive, more manual processes that still exist in the day-to-day from marketers and CRM teams out there, we try to explain the value and the benefit here within one slide because the context-aware data components, data products from SAP, combined with the great technology that Rouzbeh just explained to us, really helps marketers and those teams working with our platform to translate a very generic template and content block into all these unique and fully aligned to the specific audiences I want to engage with content elements. So in this case here, it's very much the same shoe we're trying to promote, right? Because there's still merchandising contracts, there's the production, there are still these marketing calendars which we'll need to align to, but we can do that in a way where with really just a couple of prompts and understanding the context of my specific marketing operations, I can very quickly translate that into an urban style type of content layout and eco-friendly. I can make sure I'm aware of not just regional-specific, audience-specific but then also product-attribute-specific content. And so we're very excited because also now within our Q3 release, we made this fully GA, some of our fully embedded AI capabilities. So without further ado, let's have a look at the demo as the next step here. Okay, so let's take a look at the demo portion of today's session. We're going to kick it off here straight within Engagement Cloud. Joule, is already open and we're going to look at deploying things. We're gonna use some of our autonomous agentic capabilities to create an entire campaign from scratch, and the content using our agents in the background across email, but then also in app. And then we're gonna look at how we can also still execute and get access to those AI capabilities right inside the tool. So not every time we have to trigger a fully autonomous, agentic use case, but we can combine the best out of those two worlds. So I'd start by inserting the prompt to kick things off with. I've already prepared to one to say, we want to prep for product launch by giving some information about the product itself, but then also giving instructions to make sure it's focusing the right segment, and regional nuances, and so on. While Joule now is analyzing my request, defining the intent and then operationalizing it by accessing the different agents in the background, there's much more happening than just a pure content creation that we're focusing on today. So we're just gonna skip a couple of parts and gonna go straight to the email and in-app campaign creation portion as a next step. And so this is the entire program that was now created by Joule in the background. I now have the possibilities to go straight into the email and to not just review what content and editorial elements were created by Joule in the background, but I can also fine-tune still the email using some of the embedded AI capabilities. So upon my review, I can quickly check and see, okay, like in this case here, there's not a perfect alignment of the images that were used. And so typically also for marketers, very manual labor-heavy tasks. Either I would have to go back to the merchandising team, request a different image, I mean, that could take weeks if we don't already have it ready from the shooting before, but this is, I think, one of the prime examples of how some of the gen AI capabilities can be used. I can click into the image, quickly select the content composer to then get access to not just the context of my current campaign, but I can therefore then also use a prepared prompt here to request the alignment and the look and feel of this product image to be changed to really align this to the white and green sneaker in this case. It didn't edit or change the shoe itself. It really just changed here the look and feel and the direction of the shoe. So really a quick task which might have taken hours or weeks researching whether or not this image was available before, and then we could simply do just in a couple of clicks. All the content that is generated by those AI capabilities are stored within our overarching channel media database. So even if I now, in this case here, go into the in-app, which was already created by the agent, I can see that the same gen AI content elements were used in this case here for the eco-friendly audience of the sneaker, promoting it, the same product images, and so on. Everything happened autonomously in the background upon my request to the Joule agent from the very beginning. Now let's check how we can create a similar look and feel of the campaign, but for a slightly different audience. Let's go back therefore to the automation program, and we're gonna do a couple of changes into the program manually by making sure I can include a different switch, like focusing on a completely different audience, saying, okay, in this case, it's not just my eco-friendly, I want to target my urban audience. And then I can further on and use the same replicated email campaign to then go fully into editing mode here. Now let's take a look. So we can also select the hero image, open the Content Composer. We can always manually add additional images here in this case for additional context. We're gonna give him the white sneaker, select the image, quickly insert the prompt to say, "Look, I wanna change the background in this case to look and feel and do not touch or edit the shoe at all. Just want a simply different background similar to the style of this eco-friendly promotion." And I can do so again with just a couple of clicks here. This doesn't just work for the images. So if I go out here and I click into the editorial section, we can also see that whenever I work on text together here with the AI-assisted Content Composer, what's really important to call out is the additional context. So I have the entire email content for context automatically when I select this check box. I have access to all the different image property for additional context of the same email campaign. I can select my AI profiles, which are directly connected to the brand governance model that's controlling and really making sure everything happens within those guardrails. It already understands which audience we are targeting with this campaign. I have all the product information here for reference, but I can also, if I scroll here to the very bottom, add additional context from well-performing campaigns from previous launches. So I can quickly search for Urban. I can select my Terra Urban release, select the campaign, and now it has the full context of that previously well-performing campaign to then also consider this when creating the new editorial for this product promotion. So if I quickly insert my prompt at this point here, close the context menu, click on Generate. Now within a couple of seconds, I have the editorial section I can create. But we don't just stop here for the hero image and the editorial section. If I go further down here, I can then also not just make use of our AI-assisted Product Finder by prompting my way through my inventory and product information to really find the relevant products that are gonna resonate the most with my target audience. Click on Search. Get instantly the different and right products based on my input, and I can insert them relatively quickly with just a couple of clicks, making sure not just the images, but the names, the pricing, and everything is completely correct. Now I can go ahead and insert an entirely new block into this campaign here and actually make use of those images I just inserted by now also using again the gen AI capabilities. I'm going to add a couple more images that we've just inserted. Deselect here the hero image that I just used from the content block template. I'm gonna select the images, so the products I want to specifically promote for this audience. And by inserting my prompt to say I'll create an extra shot using these three products, within just a couple of seconds, I really can create a more contextually relevant promotional banner for those products, making sure I create a virtual model. I can make the model wear all the different products I'm about to promote as part of the campaign to really just already give all my consumers a much better example than just kind of like a static product shot of how these products will look, how they can combine them together, how they create a completely new look, and it creates an entirely new shopping experience, always making sure we can leverage all the business-relevant context, in this case, audiences, product images, previous campaign examples, and KPIs, to then generate newly created personalized content that helps me to better target and sell better to my targeted audience in this case. So I have to say, I'm not as old as Rouzbeh, do not have two decades worth of AI experience, but I have a little over a decade, and that was one of the coolest demos I think I've ever seen when it comes to AI that's real, that's live, that here. So excited to continue to see this going. Okay, thank you, Lucas, for the amazing demo. Just to wrap up this section, I wanna bring this to life with Arjun from Infosys. So. Just to set you up here, for most marketers, getting AI off the ground can be a bit daunting, but curious how you're thinking about AI right now, and is it daunting, is it not? But if you're like 78% of the people that said in our recent research, AI is essential for retaining customers in 2026, you're definitely not alone. There's a lot of others that are thinking it is essential, but they might still be in that pilot mode to get it actually implemented. So Arjun, let's bring you up and bring you on to the panel here. So you've implemented AI for a lot of customers. This is real. This is happening. And curious from your perspective, does it actually work in the way it's supposed to be? What's it looking like for you? Of course it does, Megan, we have built a very exact set up for a real client. Picture it like this, our Dubai fitness challenge, it's a 30-day challenge which is announced months ahead of time, and you know where exactly you are going to get the demand and what is going to be the demand, the shoes, the yoga mat, the hydration gear, and you now which of the specific neighborhoods that you are gonna get this demand from. It shows your marketing engine can actually see the live stock that is available and the fulfillment at the same time. Stop guessing which campaign to run and act on the data and make it work. That's exactly the pitch. Let me show you what actually happened when we built it for a real customer as we go along. Awesome. Yes, thank you. I love to see, you know, feet on the ground, the people who are in there making this happen. So let me switch gears here. Let's accelerate some of this AI adoption. I know we're getting a lot of pressure from leadership, from the world right now to dive into AI. So one stat, again, from our research here, according to the Engagement Index: 54% of enterprises can't access or use real-time data. In fact, 60% suffer from dark data, which is data that is collected but can't be used. So I'm curious, in your opinion, what's the biggest blocker to that real AI adoption right now? Is it data, trust, skills, governance, maybe something else, maybe a combination of all of those? But Arjun, let's bring it back to you to kick things off there. I agree for me it's data. The model side is already solid, and we saw great work that has been done in Gemini, Vertex, etc. So most enterprises haven't connected their own data yet. So what AI does is guessing instead of knowing what it should do. So once you get the data part right, I think AI will be able to turn on its magic wand. Yes, I love that, I need one of those. Also, don't tell my boss, but I want that job that Rouzbeh was talking about earlier with the Google guy with the hat to hike in the Himalayas. So if there's a job description, just let me know, and I will sign up. Okay, next up, so when we're thinking about agents, when agents are going to be sending those personalized content and they're gonna do this autonomously at scale, how are you gonna make sure that every message stays on brand and compliance? So Lucas, let me bring you back up. You shared an amazing demo, but what does that look like for you? Yeah, I think it braces also back to the context that we previously discussed, right? Obviously, the brand governance framework needs to be in place. I feel like it was one of the probably most given bits of feedback when we had the early, early adapter conversations with some of our customers here. Similar on how we said, like, we're all aware from how we can utilize AI on the consumer side of things, but now I'm doing it for the business. And so, even for some of the output to be aware of just some of the data touchpoints that we talked about that are relevant for the consumer to make it feel very personalized and individual to them, there are all these guardrails that obviously a brand specifically within the enterprise space also wants the output to be framed in, right? Because ultimately there has always been a lot of work going into things like making sure brand CI, brand guidelines are fully fulfilled across just the different channels, just so when we look at the look and feel of let's say email templates, website design, mobile native apps, and so on. And so those are all very important aspect that we're trying to align the AI output as well too when it comes to some of the capabilities that you've just seen across the demo, to then make sure also when it come to AI, similar to some of the brand colors, you know, like font, like some of the basics that we know from visual brand guidelines. We're helping brands already to define those same guidelines when it comes to the AI output, to say what's the tone of voice, which we have the editorial example there from an email to say, like, there's a specific words that can never and should never be used, there are specific terminology that identify also the brands in terms of how it communicates the value of its products and its DNA. And so these are all things that we're taking care of on top of the fundamental things that also Rouzbeh talked about, where we talked about explicit content that should never show up when it comes to the output that some of the image or video generations, it's then the next level where we're trying to align brands with their guardrails to really make sure they can also get this through the legal and compliance team internally as we talk about AI adoption. There's a lot of things to think about and governance obviously is a big one. So Arjun, back to you, someone who's working in these live accounts, can you share an example of what governance looks like when it comes to an actual customer example? The case in point, it hasn't happened to our customer, not where I am actually working still I have my job, but I would have named the brand where this exactly happened. Earlier this year, a well-known global retailer ran a promotion that landed on a locally sensitive historical date. Part of the campaign was AI generated, and it went through a six- or seven-level approval process, but nobody caught it because it was the AI-generated wording. Everybody went through the style, the look and feel, etc., but they didn't pay attention to the wording. But it resulted in stores getting closed. The country CEO was dismissed, and the company was asked to run mandatory sensitivity training for all of their employees. So every marketing team there will tell you, "Oh, we have enough check and balances in place. This will not happen to us." The global retailer had the check and balance as well, but it wasn't just built to catch that AI-generated line. So that exactly is why a dedicated check, something like a brand governance agent that Lucas was talking about, is mandatory. It's not nice to have, but a must-have, even before you start using AI to run your autonomous contracts. So building that foundation, getting that governance and check is so important because it could happen to even the biggest of biggest retailers. I know you can't share the name, but it could happened to them with seven different checks and probably multiple different teams. I think it could to anyone. So getting that government's piece right is crucial. So thank you. Rouzbeh, back to you, welcome back. So from the Google side, how does Gemini Agent Platform think about trust and safety as more of this execution shifts to that agent-to-agent workflow? It's exciting. Yeah, I would say it's front and center, probably one of the most important aspects that we focus a lot of time on. Outside of quality, that would be probably the most. If you think of Google and why customers come to Google, especially enterprise, it's because of trust. It's because they know that we put that as a priority and make sure that we set all the structure in place to make sure security is in place so your information doesn't leak out outside of the wild garden of your enterprise. Make sure we have privacy in place, not just for external, but also within. So if you look at, for example, solutions that we have like Gemini enterprise, the same types of rule-based access controls that you can give for your personal Google Drive, where you share a file with someone, and etc., that need-to-know based permissions even goes down to level of detail with every single file, being context aware of who should have permissions to that type of an information. And so there's a lot of focus on security, safety, privacy, but also the ethical aspects of AI of, are these being used for the right use cases that actually benefit society and not. So I would say it's probably one of those things where, as a customer, you should expect companies like ours to actually heavily focus on those because we are building the foundations for all of these use cases to come out, and so we take that responsibility, not lightly, and we've put a lot of emphasis into that. Great, okay, well, let's bring us back. I know we're getting short on time here. So the crux of this conversation is really moving from pilot mode into scaling this across your enterprise. So Arjun, from your experience, can you walk us through maybe what separates some of the companies that are getting in this pilot mode from ones that can scale their AI? Yeah, probably three things that would separate the ones that are scaling beyond the POCs that they have been doing on Android. So first one, every demo looks great, but what's hard is making that same AI demo to fit into their customer actual SKU, the store network, and fulfillment rules, and whatever they have. That's where the project starts. So how do you make sure that the model and the demo that works also works in the customer messy data? That's the first part, getting the data right. The second one is a Dubai example that we spoke about a little while ago, right? So if you are able to connect that predictable front office demand to what's actually in your back office, the real-time ERP stock and what's about to run out, and then the marketing can start acting on the actual data signals instead of making a guess. We have done this exactly with some other customers. A global beauty brand connected real-time inventory and fulfillment signals into their marketing engine, and the campaigns only promoted when there was an actual stock in a nearby vicinity. During a major stock shopping event that ran, it grew 84% year-on-year jump in the daily sales during this event in the cosmetic division. That's not a demo number, that's a real-time proof point on how you can effectively use AI if you are able to connect to the data. Third one is the pattern that we see across every customer. Don't roll everything out on day one. Take a single use case, one signal. It could be a stock. It could a specific seasonal spike. And prove it works end to end. Make sure that you're able to connect all the data across your PIM, where your product resides, your CRM data, where you're customer data resides, and the actual operational data within the ERP, so that you are able to get that one single customer personalized view and then expand. So we run this along with Infosys Aster, which is a CMO offering suite that we have in terms of how the marketing teams can amplify their productivity using AI. The ones who try to do everything at once are the ones who are still stuck in pilot. So if you're asking what separates the pilot from production, it's not the model. It's always that you need to start with the clean data, put the governance in place, and don't try to pull them afterwards. Get the data and the governance right, this all will fall in place. Great examples here. And then for those who are making this leap, can you come in and share? SAP has a lot of these features, capabilities, functions that will help with governance, help us to scale. Maybe Arjun, I'll bring it back to you and then Lucas, if you wanna chime in. But what are you seeing as far as the impact and the results? Like, why are people doing this? Why are they scaling? Yeah, so Wella drove somewhere around 25% channel revenue from AI-powered push messages. And PUMA probably saw a 5x revenue from email. So these are the customer proof points that we have done together with the SAP Engagement Cloud. Adding on top of some of the examples that Arjun delivered in terms of engagement KPI, so even the revenue generated from some of outputs, we can also speak to some of just workflow and operational lifts and shifts that we've seen to just even decrease the amount of time that it takes teams to really complete the creation process of these either full journeys or individual campaigns from sometimes, just if it's a couple of hours here and there for each of those messages to reduce the time there is what we've seen quite a lot from some of the early adopters to really say they can just effectively cooperate much quicker on a day-to-day basis because some of things there are, they had to rely on other teams, whether it's graphic design teams or with the example that we had in the demo going back to merchandising teams ask for different assets, they can now fully operate and execute on their own behalf. Awesome. Okay, Rouzbeh, I'll give you the final remarks here. So from Google's perspective, what is telling you that enterprises are actually really adopting this and not just running demos? Not that we don't love your demos, Lucas, but bringing this to life. The short answer, I've never been busier before. Now, as you go into different engagements, the conversations are changing from what is gen AI, how do I use it? And you're starting to hear about, "I'm applying it in this approach. What is the best way for me to actually solve for it?" And then the question is the next thing. So we're hitting edge cases. "Can we change it this way? Can we have mechanisms for AI to check its own outputs, right?" You're hearing that next chapter of questions as opposed to the introductory. So we're well past the one on one. We're now seeing people widely adopt it at scale. And I think from a public standpoint, you're seeing the results like every quarter, and Google's recording the numbers. So when I, there's a little bit of seriousness behind the joke of we've never been busier, but it's actually very real. A lot of customers are looking at this. These are board mandates. These are at the director level, at the employee level. Everybody's experiencing AI and seeing what they can do for themselves and automating pieces that they don't want to do manually. And so that adoption is happening very much like we've never seen before. It's unprecedented. Well, thank you all for being here. Thank you all for attending this webinar. Thank you to our panelists and I hope everyone has a lovely rest of their day and is excited to bring AI into their business. So thank you all for joining.