[Anjul Bhambhri] Hello, everyone, and thank you for being here this morning.
In yesterday's keynote, Shantanu and Anil described the Adobe AI platform where data, models and ultimately agents support all your marketing, productivity and creativity endeavors. Now, that technology comes to life in Adobe Experience Platform and all its associated applications, really unleashing Agentic AI to deliver unified customer experiences.
Now, last year, we introduced the AI Assistant. It's a technology that allows you to use natural language, to seek insights from your customer experience data. Going forward, we are supercharging the AI Assistant with agents. So I'll spend a few minutes describing what are agents.
Well, to start with, they are powerful software artifacts that are both intelligent and autonomous. You continue to engage with them through conversations, but they go beyond answering questions. They can actually do things for you.
I mean, you have to tell them what you want. What are your goals? What are your constraints? And then they get to work, often in the background, and they'll proactively make suggestions, filling in gaps. And it turns out that Adobe Experience Platform applications are a natural fit for such agents. So sure enough, we tightly integrating them into all your favorite applications, like the Real-Time CDP, Adobe Journey Optimizer, Customer Journey Analytics, and Adobe Experience Manager.
Now all of these capabilities will be surfaced through the AI Assistant. So thereby you have a uniform interface across the board. Behind the scenes, the AI Agent Orchestrator will manage all these agents for you. So maybe let's take a look at a few of the agents that are on their way. The first is the Audience Agent. So as the name suggests, this is your ally when it comes to audiences, and it is tightly interwoven into Real-Time CDP.
And with the familiar AI Assistant interface, you'll now be able to create and manage audiences. So let's take an example of an online travel website, and say an audience specialist goal is to increase car rentals through cross-sell.
The specialist can ask the AI agent to find customers with flights, cruise or hotel bookings in the next six months and then filter the ones that have a high propensity to add car rental. So if you see that the directives are pretty simple, that the practitioners are providing to the AI Assistant.
But what happens behind the scenes is a lot more complex. Now the agent begins by analyzing your stated goals and constraints. It reasons about your intentions, and then it generates an orchestration plan, picking up relevant attributes and events from unified profiles. It may even build some descriptive and predictive models on the fly, to estimate audience sizes. And eventually the agent explains the plan to the practitioner, seeks adjustments, and then based on your feedback, it can update the plan and then go into action.
So in effect, you as a practitioner are specifying what you want, and the agent focuses on the how. And of course, all of this is managed by the Agent Orchestrator behind the scenes.
So next up is the Journey Agent. Now this is embedded within Adobe Journey Optimizer. So just like the Audience Agent, this also accepts your directives in natural language to create and manage journeys.
So it can handle basic things like, Hey, what are my top performing journeys? Or, Stop all upsell journeys. But it does go further than that. So if you take travel and hospitality, it's an industry with loyalty tiers like gold, silver, platinum, and when customers risk losing status, brands have to bring them back into the fold. So how do you incentivize them to stay on track? That's the goal. And the constraint is that loyalty programs run on an annual cycle. So when this problem, what are these goals and constraints, is proposed or stated to the agent, it will design a journey to entice certain customers that are off track with relevant offers. It will look up the past travel behavior. It will look up their loyalty status. It will combine that with industry knowledge as well as with brand requirements. And then it will recommend when and how to nudge customers. Deciding on the right offers, the timing as well as the channels. Now these recommendations, with explanations, will be presented to the practitioner and they can define them iteratively. So I think you'll see a pattern here that you are always in the driver's seat and you have an autopilot that you can switch on. So now that you've built your audiences and crafted what you think are the perfect journeys, that's awesome. But we know that experimentation in marketing is an always on exercise. As customers change or markets evolve, you want the ability to play with new ideas. And to that end, we are building the Experimentation Agent. Now the Experimentation Agent, it lets you quickly cycle through ideas, deploying the ones that prove effective. So again, let's take an example. Say you want to increase the new subscriber engagement. You present this to the agent. In turn it gathers insights from your customer experience data as well as prior experiments and it'll generate multiple ideas or hypotheses. It even predicts results for each hypothesis. So let's say one suggestion that the agent makes is to incentivize subscribers with small rewards for completing certain tasks. You like the expected lift, you choose to run that experiment, and at the end of which, it's going to summarize the results in a way that is pretty easy to comprehend. So at that point, based on real world validation, you can either productize that experiment or you can continue to iterate further. So with that, let's look at how Marriott is using agents to enhance their Bonvoy guest experiences. With me on stage, I have Rachel Hanessian, our Senior Product Manager for Generative AI. Over to you, Rachel. [Applause] [Rachel Hanessian] Thanks so much, Anjul. [Applause] [Rachel Hanessian] I'm excited to show you all a host of new capabilities that are coming to Adobe Experience Cloud that will help brands deliver unified and even more personalized consumer experiences across all touchpoints. Before we get started, I'd like to thank Marriott for allowing us to use their brand for this vision demo, and all data and workflows that I show are fictitious, to protect the privacy of Marriott guests.
Now, Marriott's business travelers are really as loyal as any in the hospitality industry. And empowered by new Adobe capabilities, they're finding smarter ways to inspire Bonvoy business travelers to extend their trips. I mean, how many of you all are sticking around in Vegas after Summit? All right, a couple of you. Well, let me show you a glimpse of the experience that Marriott can deliver.
I've just finished booking a work trip to Las Vegas.
After booking, I receive a notification to extend my stay. Now, I'm just thinking business right now, so I'll swipe it away. But later, when I'm in my confirmation email, I notice that there's an opportunity for a points bonus if I extend my trip.
When I tap through, I see discounts on local Vegas activities. And what, is that a discounted helicopter tour? I'm going to extend my trip with this activity. And now I'm all set for my Vegas work trip.
Now that Marriott upsell experience was delivered to guests at scale, leveraging several new Adobe capabilities. Why don't we check in to see how we put it together? I'll start by finding the right audience to target for my goal of extended business trips. I'll do this using Audience Agent, a new goal-oriented way for practitioners to create and refine audiences in Adobe Experience Platform. I'll start by asking Audience Agent through AI Assistant, help me expand my extended stay plan with additional audiences.
It looks like Audience Agent came up with three audiences based on my goal and looking a bit more closely, luxury travelers will play really well for this campaign, so I'll add that to my plan.
Once I'm on the canvas, I can actually simulate engagement across the different channels using machine learning on the fly. And it looks like I have a pretty good opportunity for email promotion. I'll definitely be capitalizing on that later. Last, I'll optimize my audience just by focusing on profiles that have a high propensity to convert.
And with that, I just got a pretty sweet upgrade. And I didn't even have to pass by the front desk for it.
Now that my audience is optimized, let's sculpt the journey that these business travelers are on.
Rather than starting from a blank slate. I'll just prompt AI Assistant to create a journey to incentivize business travelers to extend their trip to include leisure time.
Now, AI Assistant is using past journey performance data as a reference to help me come up with some basics for my use case. It even identified the two audiences that I just optimized. I'll indicate further that I need a FOMO-inducing push notification, a personalized confirmation email, and discounts on local activities.
Fantastic. That saved me a lot of time. With just a few short prompts to AI Assistant, I was able to publish a draft of my journey all the way to the canvas. Easier than room service. But we're not done yet. We now need to continuously optimize the journey that these business travelers are on using experimentation. And to do this, I'll use Experimentation Agent.
Now, the metric that I'm trying to move is booking extensions.
It looks like Experimentation Agent has already come up with a pretty promising hypothesis for me, that promoting discount on excursions instead of a points bonus will drive more booking extensions. Using this information, I can add an additional treatment to my journey, and then, based on simulated performance, I'll either roll this experience out to everyone, or I'll go back and find another hypothesis to test. But for now, I'm ready to pack my bags and go. Back to you, Anjul.
[Anjul Bhambhri] Thank you Rachel. [Applause] [Anjul Bhambhri] That was awesome. [Applause] [Anjul Bhambhri] We are truly thrilled by the uptake of our AI capabilities on the AI Assistant. An overwhelming number of you have adopted it, and we've heard from you that you are seeing significant productivity gains. Here is a quote from a customer that is quite representative of what we are hearing. "AI Assistant has saved us hours troubleshooting, data, journey and audience management." Even on agents, the early feedback has been equally promising. We've seen great success at AAA Northeast. They saw 165% increase in car rentals, while targeting just a fraction of their prior audience sizes. So in this case, less is more.
Wegmans. They also saw 3x higher engagement rates using AI-powered audiences for mobile campaigns.
Now, just as marketing practitioners have the AI Assistant, your brand's customers will also soon have a similar technology at their fingertips. And we are calling this the Adobe Brand Concierge.
Now, we are all familiar with the power of commodity LLMs. At Adobe, we are taking this technology a step further. Brand Concierge is really the fusion of generative AI with customer profiles and your brand's asset portfolio. So much like the AI Assistant, this is also conversational, it's multimodal, it has insights into your customer's buying patterns and preferences. And it also knows about your brand's offerings. It knows about the inventories of products and other assets. So all of this really allows for personalized engagement and decision-making. No more generic answers from commodity LLMs. So when my daughter goes online shopping for tops, the Brand Concierge can recommend the colors and styles that match the pants she bought in-store a few weeks back. So see what I just did there? Now, if my daughter hears this, she would want all her favorite brands to have the Brand Concierge now. Over to Rachel to see this in action. [Rachel Hanessian] Thanks so much, Anjul.
So, many of Marriott's premium properties can offer a personalized concierge experience in the lobbies of their hotels. But that type of experience isn't really scalable across all hotels in their portfolio. Or is it? If done digitally. Jump ahead with me to the last day of the conference. I've now shifted from business to leisure mode, and actually, these sunglasses help a lot with the bright lights. At the end of the conference, I received a push notification about a celebrity-chef dining experience. And can I just say, I love that Marriott knows that I'm a foodie. It's in my consumer profile.
When I tap through, I enter a multimodal conversational experience that's powered by Adobe Brand Concierge. And will you look at these brand assets? There is nothing generic about this experience. It looks like the dinner is tonight at Marriott's partner, Cosmopolitan Hotel. Let's get a little bit more information about it. Can you tell me more about the chef? This looks fun, I have to meet this guy. Let's book an 8 PM reservation for four tonight. Anyone want to come with me? Great, well, Marriott was able to confirm my reservation instantly by linking me to a Booking Agent that's linked to Adobe Commerce. Well, back at Marriott, with Brand Concierge deployed to over 9000 properties, they need a really simple way to analyze across all of the guest interactions. They'll of course do this using Adobe Customer Journey Analytics' new Data Insights Agent. I'll simply ask AI Assistant, analyze Brand Concierge requests. This is looking across all 9000 properties. I can even filter down just by business travelers. Seems like everyone's really interested in these chef experiences. Last you may have noticed this real-time Refresh button at the top. And yes, it does exactly what you were hoping it does. When I select this, all data and visualizations update to show in real-time. There's no better way to get a pulse on what's happening right now across your entire consumer experience. Well, it's time for me to turn in my key card and check out. Back to you, Anjul. [Anjul Bhambhri] Fantastic, thank you, Rachel. [Applause] [Anjul Bhambhri] Thank you. So in addition to the Brand Concierge, what you also saw is the evolution of Analytics. Conversations are the new way that both the marketers and decision makers, they draw insights from data. And as you heard that the Data Insights Agent is enabling that shift. Now it certainly covers your basic scenarios like, what's my revenue by product, category, this month, or in a certain region? But it really shines with sophisticated capabilities such as forecasting, anomaly detection, causal analysis, and remediation. So to give you one example of its power, it can compare audiences against stated goals and explain what are the potential reasons for any of the differences across this. Now, these are difficult things to do without this kind of a technology available to you. Now, this kind of an analysis can be applied to whether it's a purchase, customer onboarding or order fulfillment funnels and really enable your teams to take action quickly.
So I'm sure you can appreciate how AI is transforming marketing. This technology will allow you to create more personalized and impactful customer experiences. We really look forward to seeing how you will leverage these innovations through our applications to drive unified customer experience. Thank you so much for your great partnership.
