AI is changing every part of customer experience. The technology is real, the outcomes are measurable, and the teams moving now are already seeing the difference. But the question on CX leaders' minds right now is the same. What does this actually mean for my team? And that is what we are here to explore today. And more importantly, we're going to show you. Today I am joined by Sara Richter, chief marketing officer for SAP CX Engagement Cloud, Kollen Glynn, senior vice president and global head of SAP Commerce Cloud, and Sanjeet Mall, senior vice president and head of SAP Sales Cloud and SAP Service Cloud. I want to hear thoughts from all three of our panelists, because the shift we are seeing today in Autonomous CX is that humans guide and AI executes, right? So Sara, perhaps you could just start us off and say what you think the most exciting part of this development is. So there's a lot to get excited about, but for me, it's that businesses can now react to signals from customers when they actually happen, rather than finding them out after the fact. Mm. And the reality is that there's too many signals. No team can go through them all manually and react accordingly. And then, of course, we're now seeing the shift in terms of AI trust. Some of our recent research showed that a third of consumers are okay with an agent making a purchase on their behalf. That's interesting. It is. And 78% are okay with routine purchases made regularly on their behalf. So this is a really significant shift that's exciting. And I think what it's going to mean for businesses is that when they think about growth and loyalty, it's really going to come down to who can be the most efficient and effective in that moment. The great news for our customers is if you're already an SAP ERP customer, you have the advantage of this incredible wealth of operational signals, orders, fulfillment, inventory, and loyalty at your fingertips today. Very exciting indeed. Kollen, what are you thinking about? I think the thing with commerce that I'm really excited about is that we've really thought from the front-end customer experience to how customers discover products all the way to the merchant side and how things get optimized and fulfilled. That whole journey we've reimagined, and there's something for everyone all the way along there. I think people will be super excited about it. Very much so. And Sanjeet? You know what I'm excited about is in the area I work in, sales and service, the top line is very important stuff, right? And for the first time, some study shows that 70% to 80% of the time that sales or service rep are spending on is actually bookkeeping, which has no value whatsoever, right? No value whatsoever. However, it's required. Otherwise, how do you document this? How do you do this thing? Now, with the advent of AI, and you'll see some of that in the demo, is all the minutiae is taken care of so that humans can concentrate on what the real value is: building customer relationships, solving customer problem. And that's what makes me excited that first time in humanity's history we can do 70%, 80% of real work. So what does this all mean, Sara, for the way that marketing teams are operating today, day to day? And I hear you might even have a little demo for us. So before we do that, it's really a big change. Today, marketers are spending so much time building. They're creating campaigns. Mm-hmm. They're pulling lists. They're orchestrating journeys. And they're reacting to performance data that could actually be up to a week old. Wow, that is old. Yeah. And apologies if it's loud to those listening at home. We are on the show floor. On the show floor, aren't we? Things are happening here in Vegas. Absolutely. So but the beauty is, with autonomous marketing, what marketers can do is they can focus on the business outcome that they're trying to achieve. They can use Joule to help them define lists, to define KPIs and parameters, all the things that really matter. And then the assistants and the agents in SAP Engagement Cloud, well, they do the work. Mm. So omnichannel orchestration, campaign optimization, all of these things now happen continuously, fed with context from real-time signals that come from the customer and the business through SAP ERP. And that's really exciting. And what we're also excited to do is to partner with Google to really allow marketers to move even faster, to get better access to personalized imagery, messaging, recommendations. At the end of the day, being the customer feels understood, not targeted. Mm-hmm. So I think that's all pretty cool. Hopefully everybody agrees. But rather than leaving it there, why don't we have a look at the demo and see what it looks like in real life? We are going to roll that video. A global sportswear company is launching BestRunX, a new lightweight, sustainable product line across multiple markets. Global marketing manager Maya has two weeks to launch and opens Joule Work. She tells Joule she needs to reach the most relevant audience. The Campaign Assistant pulls live signals from across the business, ERP inventory data, re-engagement from past sustainable buyers, and rising purchase intent mapped in seconds. The Audience Identification Agent surfaces eco-conscious athletes as the highest value match. Joule recommends building on the Propel campaign, a proven foundation adapted for BestRunX and this audience. Inventory varies by region, so the Campaign Configuration Agent structures two coordinated journeys: immediate purchase messaging for in-stock markets and a pre-order nurture path for constrained regions. Maya reviews the campaign and confirms. The campaign opens in SAP Engagement Cloud. Maya reviews the automation flow, triggers, timing, and journey logic already configured with ERP inventory signals live at every decision point. She opens the email. Messaging and product recommendations already personalized for eco-conscious athletes, grounded in brand guidelines and live inventory signals. Maya approves. The Campaign Orchestration Agent activates delivery across all five channels. Weeks later, she returns to Joule Work. The Campaign Insights Agent surfaces the results, reach, engagement lift, and revenue impact, and feeds those learnings back into Joule, ready to shape the next campaign. So it was so interesting there, Sara, to see how that looks in marketing. Thank you so much. And Kollen, when you're looking at commerce, obviously the customer feels this most directly, right? Yes. So how does Autonomous CX actually change commerce from how teams are running it to how customers are experiencing it? Yeah, if we start off on how customers experience it, I think everyone is aware that the AI search engines, ChatGPT, Gemini, Perplexity, Copilot, is driving a lot of traffic now to storefronts. But if you're not showing up there, I think everyone's heard the term that GEO is the new SEO. So one of the things that we're doing in commerce is we're coming up with an AI merchant center. We're doing direct UCP integration into these AI search engines. And even we put out an MCP server, so if you want to come in that front door via your own agent or AI, you can do that as well. If we move on to the shopping experience, we've come out with a Shopping Assistant that allows natural language queries, multimedia type queries. You can upload pictures, video, whatever. We understand that intent. We turn that into proper results. You can apply filters and all sorts of things. If you move down the pipe, we're looking at order execution. The reliability agents that we put in place automatically resolve conflicts where orders might normally get stuck. And we do revenue reconciliation to make sure that from a point of sale or front-end e-commerce actually matches what gets posted to the ERP back end. And then even autonomous optimization. So you remember I said we have the AI merchant center that we're pushing catalog to these AI engines automatically for you. Well, part of that is also making sure that you're matching the schema that Google expects or OpenAI expects or Microsoft expects. And that can be challenging and very tedious for a merchandising team to do. So one of the things... Some of the agents that we put together are around content scoring and optimization. So we can automatically match your catalog against a schema, look for weaknesses, and suggest optimizations automatically. So what would have normally taken a merchandising team maybe weeks to go through manually can now be done in minutes. Wow. And it can automatically keep up to date with that as new products get added to the catalog. So as part of that, I'd like to show a demo that shows what that back end can look like for a merchandising team. Absolutely, let's take a look. Hello everyone. Meet the SAP Commerce Merchandising Assistant, an AI agent built into Joule. When we ask Merchandising Assistant, "What can you do?" it responds with two core capabilities: product content scoring, which evaluates your catalog against a target schema like Google Merchant Center, and product content optimization, which generates AI-powered enrichment proposals for missing or incomplete attributes. Let's see it in action. We ask the agent to perform a health check on our B2C sales catalog against Google Merchant Center schema. The agent immediately kicks off a background validation task. When we check the status, the results are clear. Out of 25 products, only one is feed ready, three are AI fixable, and 21 are rejected. The catalog needs improvement, and the agent is ready to help. We navigate to the health report. Here, we see all 25 products with their individual health scores and classifications. We select the products we want the agent to fix and trigger the optimization. The agent queues five products and begins working in the background. No manual editing required. The optimization results are ready. For product ZB-FF-026, the agent has generated AI proposals across 11 attributes, including condition, description, material, and more, each with a confidence score. And here's where human-in-the-loop control matters. Nothing is applied automatically. The merchandiser reviews each suggestion, edits the value if needed, and explicitly accepts or discards it. You can go one by one or click "Accept All Proposals" to approve everything at once. Only after your approval does anything get synced back to your commerce system. Fabulous. And Sanjeet, the sales and service is often where the customer relationship is won or lost, right? So for the teams that are on the frontline of that, what are we seeing change? And I hear you also may have a couple of demos for us. I absolutely love to show the demo, but before going there, the things are fundamentally changing. It's not a little change, it's a massive change, right? And as I was saying before, this frees up our assistants and agents in sales and service both. They are designed to achieve a goal, right? And the goal really is that everything from nurturing the lead to fulfilling the lead, whatever can be automated, they kind of made the assistants do that automatically. Autonomous agents, right? So what you can do as a sales rep or a service rep can invest your time in the things which matter, which is really relationship building, solving customer problems, finding new ways to do thing which adds value to yourself, the company, the customer. And the first one, the first demo I'm showing is really showing the way a new modern salesperson looks like, how his day looks like, how he's being having multiple assistants so that he can focus on his customer. Let's see the demo. Liam has a customer meeting coming up with Maggie from a healthcare organization. Traditionally, preparing for this meeting would mean opening multiple systems, searching emails, reviewing CRM notes, checking past activities, and researching the customer online. But instead, Liam simply opens Joule and asks it to prepare him for the meeting. Joule automatically pulls together previous customer interactions, email history, opportunity details, customer activity, public web insights. It even identifies that Maggie is the primary decision-maker and highlights why this opportunity is strategically important. Liam doesn't spend time researching. The insights come directly to him. Now Liam joins the customer call with AI assisting him in real time. Think of it like having a highly intelligent sales coach sitting beside you, quietly helping throughout the conversation. As the conversation happens, the AI automatically captures qualification details, identifies buying signals, detects budget authority, updates the opportunity in real time. At the same time, it recommends the right products, the right pricing strategy, the optimal discount range to maximize both win probability and margin. The call ends. Before Liam even closes his laptop, the AI has already created the opportunity in the CRM. Next, Liam generates a quote. The AI automatically recommends the best fit products recommended pricing suggested discounts. Liam simply reviews and accepts the recommendations. One customer conversation, just a few clicks. The entire sales cycle accelerates dramatically. Finally, Liam receives a notification. Maggie has signed the quote, and now Liam doesn't need to manually process the next steps. Joule takes over. The system automatically creates the order, sends confirmation emails, updates the customer, enables order tracking. Sanjeet, so interesting to see the modern sales associate. Yeah. So the second demo I have is about the service. Now, what we are going to see there is a massive optimization of how a service rep can do things quicker. Multiple systems are involved, but none of that matters. Don't think about system boundaries. Let's roll this. Let's take a look at how the Case Management Assistant helps service teams resolve cases quickly and get customers back to business faster. The service rep uses Joule to quickly find out which cases need her attention first. Instead of navigating through menus, she simply asks in plain language and gets an instant list of prioritized cases. In this example, she's working on product registrations. The Case Management Assistant takes care of the heavy lifting, gathering key information, preparing the case, and processing it. This significantly reduces manual work and minimizes the risk of errors. The service rep gets the entire case processed in seconds without ever having to switch screens. In this video, we're walking through each step of the process, but this can be fully autonomous so that cases get processed automatically, and the service rep only steps in when her expertise is truly needed, guided by clear action recommendations. No more waiting for customers, no more time lost on routine tasks. Just faster resolutions and a better experience for everyone. Now we will see things through the service manager's eyes. The Service Optimization Assistant helps you build trust through proactive service management. On the home screen, the service manager gets a quick overview of how the team is performing, what the key service metrics look like, and how many cases were handled by AI agents versus human agents. The dashboard gives him more insights and proactive recommendations. Here we see how the assistant has detected an escalation risk across several billing cases and recommends reassigning them to the specialists with the most experience. The service manager accepts with a single click. The cases are reassigned, the escalations avoided, and the rest of the week just got a little calmer for the whole team and the customers. Thank you all for these insightful demos. And just before we go, let's do a rapid-fire round. One last thing that you want everyone to remember in just one sentence. And Sara, we will start with you. Fantastic. Focus on the business outcome. Pick a process that matters to your business and your customer, pick the part that matters most to your customer, and zero in on that. Love that. Kollen? In the same vein, I would say just take the step with commerce. We've got you covered no matter B2B, B2C, D2C, everything. Take the step with us. We've got something for you. Fantastic. And Sanjeet. And finally, I agree with both of my colleagues here. And just to add, don't think about system boundaries. Think about process. Let the agent figure out what system to execute on. Beautifully said. And thank you so much for these takeaways. It is amazing that what we've seen today isn't a road map, right? It's already real, it's already working, it's already available. Sara, Kollen, Sanjeet, thank you again for joining me. Thanks, Clare. Fantastic. Thank you. So we began with one question, what does Autonomous CX actually mean for my team? And hopefully you're leaving with a much clearer answer and a few ideas for where to get started. If you still have questions, there's plenty more content to watch on demand. Thank you so much for joining us.