AI Agents in AEP: An Overview

Transcript

Hi everybody, welcome to the webinar. We’re going to give it a couple, maybe like one, two more minutes to let a few more trickle in. And then we’ll get started. And just an FYI, in the chat we’re going to drop the link to the last webinar of September coming up, Enriching Adobe Commerce with Adobe Experience Platform, if anyone is interested in registering for that one. That’s the last one this month of September. And then of course we’ll have more come October as well.

Before we get started too, just everyone please feel free to leverage the chat for questions that you might have as Ben goes through all of his wonderful content that he’s prepared. And at the end of the session he’ll do his best to address as many as possible. Thank you Katie for dropping that link in there. And again, we’ll get started in maybe one, two more minutes.

All right.

Thank you so much for your time and attendance today. Just please note the session will be recorded and the link will be sent out to everyone who registered afterwards.

All right, Ben, we’re at five after. Do you want to go ahead and get started? Definitely. Let’s go ahead and kick this off. So hey everyone. So I’ll do a quick intro. So my name is Ben Joseph. I am a principal strategist on the field engineering team. My background is across data management, activation and media. So really from a solution standpoint, I’m really aligned to AEP, particularly real-time CDP. And I’m really excited about this webinar today. Obviously the topic of AI, especially agentic AI, it’s so important, not only within Adobe, but really the overall industry as a whole. As far as the agenda goes for today, it’s very straightforward. So we’re going to do a quick intro into Adobe’s history, evolution and the current state of our AI. We’re going to talk a bit about some of our guiding principles, how the AI works, how other customers are really seeing value. We’re then going to jump into the major focus of this meeting, which is the new AI agents that live within AEP and really how you can go about using them. We’re going to run through a quick demo as well, just so you can see those agents in action. And then we’re just going to close out with how you can go about getting started. So hopefully this all makes sense that it aligns with your expectations for today’s session. And again, as Amanda alluded to, feel free to ask questions in the chat pod. And then at the end of the webinar, I’ll try to get through as many as I can. All right. So kicking things off. So I want to start with sort of a mission statement for this webinar, if you will, really around the goals and the intention and what we’re attempting to cover. And as I’m sure you’re aware, the Adobe AI and agentic strategy, it is evolving and very rapidly. If you follow our product announcements, you likely saw that several more agents were released about, I think it was about two weeks ago now. That’s on top of the agents that already exist. And there will be more agents released in the future as well. And I want to be the first to acknowledge, these can be confusing to track, right? Agents, they work across different Adobe solutions, both within AEP and outside of AEP. There are a bunch that align to some of our newer AI first applications as well, some of which you may have not heard of yet. And really the intention today is to focus exclusively on agent application within AEP. So really to do that, we want to make sure that you understand the distinction between agents that live across AEP, really where they fit in across the broader customer experience strategy. And we’ll talk about use cases that are tied specifically to those agents. What we are not going to cover. So agents and use cases that live across other Adobe solutions. So even though agents that live, let’s say across AEM or Workfront may play into our overall customer experience strategy, that’s not going to be the focus for today. And again, the same goes for agents that are part of the newer AI first applications, which I’m listing at the bottom right of the screen here. So again, hopefully this all makes sense and really just aligns with expectations for today.

So with that, I want to jump into a quick intro just for context, really into Adobe’s history with AI and how it’s evolved over time. So Adobe, it’s been incorporating AI into its solutions and workflows for over a decade now. In the beginning, it was really natively integrated into our solutions and helped power specific features. So think things like anomaly detection in analytics or personalized recs in target. And then really AI as a service came into play, right? Where AI worked more horizontally across experience cloud and the open ecosystem. So think use cases like building a propensity score with customer AI within real-time CDP and then sharing those segments with AJO for activation. And then it’s been a few years now, but you obviously started seeing a lot around generative AI and Adobe Firefly. So really the ability to start prompting certain requests and then receiving a generated output. So whether that’s customized content, personalized ad copy, tailored responses to specific questions. So thinking about things like gen studio or AI assistant. So that leads us today to today in the latest evolution, which is agentic AI. And I’ll start a minute just going into what that is and really how it differs from previous AI iterations. And really we see agentic AI as sort of this more sophisticated iteration of generative AI, right? There is still an output that’s based on a prompt, but there’s really this greater level of understanding and intent. And there’s a multi-step reasoning process. And what this leads to is really this greater layer of autonomy and decision-making, and then the ability to take actions that can be based on goals, right? So from a use case perspective, how might this play out? You might see things like goal-based audience creation or journey creation, or maybe insights visualization based on very nuanced prompts. And this then leads to the question, how is this all done, right? How are we bringing the idea of agentic AI to life? And really the simple answer here, this is where AI agents come in, right? And help execute on the tasks that we highlight as possible with agentic AI. And agents really act across three different areas, all of which I think are equally important in bringing the AI to life. So one, the agents are able to interact. They process a language prompt, understand what you mean and what you’re asking, and they can respond in different intelligent ways. Two, and we’ll get more into this part, there’s a more advanced reasoning process where agents can think through problems and really make decisions more autonomously. So maybe you enter a prompt like, hey, help me understand why recent sales are down. The agent could then analyze the data, find the root problem, and then maybe provide three solutions. And then from there, and I think potentially most importantly, it can act, right? And actually do pieces of the work for you. So maybe use it to actually create an audience or a campaign or an email. And we’re gonna get more into this, but there’s always a level of human direction and oversight as well. So it’s never gonna act or create anything without approvals. So really, I think in summary, it becomes this amazing support tool that’s really able to enhance your workflows in really a really large number of different ways. And this is very important here, right? I wanna chat through really our approach, or Adobe’s approach to agentic AI, really with regard to some of our main principles around responsibility. So one, I just mentioned this, I think this is super important. Humans are always in the loop, right? While there is a level of autonomy in terms of how the AI reasons and gathers information from different systems and creates a response, humans are still in control in terms of oversight, in terms of review and approval mechanisms. There is a level of explainability. Oops. You can always reference how the AI came up with an answer, where that information was sourced. So there is gonna be that level of transparency. Quality is very important as well. So accuracy and trust, they are a huge focus. So feedback loops and monitoring frameworks are in place to detect bias and errors and such. Governance and security. So this aligns with our prioritization around data protection and privacy controls to make sure security standards are upheld. There are access controls, so we make sure our customers are always in control. So it’s our customers who decide if and how they use generative and AI, agentic AI capabilities, who within their org has access to certain features. And then finally, training and customer data sharing. So we always wanna emphasize that customer data is not used to train foundational AI models. This has always been a guiding principle from us right from the beginning.

Okay. So with all that, I now wanna sort of transition into the specific agents that are helping power our customer experience orchestration strategy. So if you are familiar with Adobe’s agentic strategy, you’ve probably seen this slide before or some variation of maybe at Summit, maybe at different Adobe meetings. And I really like this slide because I think it does a great job of designing the many different agents that we have and aligning them to our orchestration strategy. But I will caveat, I also think there’s a lot to explain and clarify on this slide. And this is where I personally like to start going a layer deeper, right? But at its whole, and we’ll go into these different components in detail, this slide is conveying that we see content, data, and journeys as the three key pillars of experience orchestration. And then across these pillars, we obviously have all our different Adobe experience solutions, but we also now have this agent orchestrator and all these different agents that you see in that middle row there. And why I like to provide clarity here is because all of these agents aren’t necessarily specific to the applications within AEP. You might be a real-time CDP practitioner and see a great use case for the workflow optimization agent, but then you might learn that that agent is actually aligned to Workfront. Or maybe the same thing with the content production agent, which is actually aligned to Gen Studio. But that’s not to say that you can’t actually use all of these agents. You just need to have licenses for the solutions that align with them. So even if you do have all the licenses that would unlock these agents, obviously AEP, it integrates with our experience cloud solutions. So all these agents would and are designed to work together. So again, this is really just a point of clarity. So while the agents can all work together across the experience orchestration strategy, individual access for these agents will depend on what solutions you you do currently have licenses for. Okay, so with that, what we’re looking at here is all these agents or all those agents on the prior slide and really where we see them aligning with the customer experience lifecycle. So we look at the practitioners, the users, or the customer experience team as really being the center of this circle. And our vision is that all these agents or all the agents are able to supplement and support and even take on many of the tasks of the customer experience team across all these different functional areas. So all of these agents are obviously important. They essentially take on different roles, but again, not all of these are specific to AEP. So what I want to demonstrate here highlighted in yellow are the specific agents that align to AEP. So if you are an AEP practitioner or user of the platform, these are the agents that you’ll likely be focusing on at first. And it really makes sense, right, when you see what pillars they’re aligned to. So things like audience management, journey orchestration, performance analysis, these are all core areas of AEP. So it should make sense that the agents fall into these pillars. And a quick caveat before we do progress further, you will see asterisks around the data engineering agent and account qualification agent. That’s just because while these do live across AEP, they’re not going to be a deep focus for today. Our intention is really to focus more on the core agents that have been announced recently.

Okay, so the actual agents that align to AEP and that we’re going to be focused on the rest of the session are the following. So we have the product support, the data insights, audience, and journey agent. And the first thing I want to call out, because you’ve probably seen this before, is really just a quick explanation around what the AEP agent orchestrator is. And what it is, it’s the central technology for building and coordinating functional agents across AEP and the overall experience cloud. So really think of it as the overall, almost container, for leveraging agents. When you’re looking to execute a certain task, you don’t separately bring up the audience agent or the journey agent. You ask a question and then the reasoning engine within agent orchestrator will decide which agent to call to then carry out the intended action. When you set up agents, you’re not setting up these agents individually. You get an overall skew for the agent orchestrator and then individual agent access will depend on solution ownership. As far as what a job is, so these are essentially just the defined workflows that agent orchestrator will process, coordinate, and then call on what are more agents to go out and execute. So what are these major jobs? So for the product support agent, it’s really around troubleshooting and creating and tracking support cases. For the data insights agent, it’s primarily analyzing and visualizing data for insights. Audience agent, it’s really about managing, creating, optimizing different audience segments. And then similarly, journey agent, it really revolves around creating, analyzing, and optimizing journeys.

And then finally, hopefully to bring these to life a bit more, what do the agents look like, right? And how are they accessed? So hopefully you’re already familiar with AI Assistant within AAP. And that’s the answer right there. Agent orchestrator and subsequently all the different agents that can carry out individual jobs, these are all directly accessed through AI Assistant. And it’s worth to note AI Assistant, it has been reimagined a bit with a whole lot of newer updates if you’re not familiar. So once you do have access or once you access AI Assistant on the right side rail, it can now actually evolve and expand into different formats. So you can see full screen, split screen, interactive cards, and the whole experience is really built for improved conversations, right? There are improved visualizations, streaming responses, personalization with more long-term memory. And then at the end of the webinar, we are going to go into more detail on how you can get this up and running and get started within AAP with all this.

Okay, just one more visualization here. I want to clarify what the reasoning engine is. Since you might hear this term or see it in some of our material or documentation, and really you can think of the reasoning engine as the brain within the agent orchestrator. When we spoke earlier about some of the agent skills, so understanding natural language prompts, activating the right agents, keeping the conversations in memory, coordinating the execution of tests, this all happens because of the reasoning engine within agent orchestrator. And when I speak to coordinated execution, reference the visual on the right there. We’ll see this in the live demo in a bit, but it’s essentially the reasoning engine that creates a plan, breaks a prompt into goals, and then creates tasks and actions that can be utilized. And again, you’re going to hear me say this a lot, it goes without saying, all this can be modified by humans, right? If you want to take out pieces of a plan, modify, or even go in and do the pieces yourself, this is 100% guided by you, the user. Okay, so now let’s jump into the actual agents. So the first one we’re going to talk about is the data insights agent, and this is something that’s available if you have a license for CJA. And what this agent does is it automates insight generation with the ability to analyze both customer and journey data. And really by doing this, it’s able to provide a variety of insights and subsequently take proposed actions. These can be from helping diagnose root causes to providing optimization recommendations. So really when we think about some of the value drivers here, primarily it simplifies insight discovery, right? For both analysts and non-analysts with the ability to self-serve data. It allows users to ask very specific business questions in natural language to get quick answers, right? Really without needing to dig through data. And when we think about the creation aspect of agents, it can help build insight visualizations and it does this right in the platform for you. So what might this look like in application? So you can ask prompts like, what is my revenue by plotted category in the last 30 days, which aligns with the data analysis and visualization use case. You can be more specific or add in ad hoc components to your prompts, right? For example, show me daily website conversions, but only across this specific region. There is a visualization adjustment component based off of natural language. So if you have a given visual, you can say something like, please change the date range to only first half and you will get an immediate revision. You can leverage the agent to help ideate on root cause analysis. A question like, why did my website visits decline last week? And then lastly, you can ask questions around summarization or insight simplification. Help me summarize the insights from this particular report. So I think really in essence, there’s a lot of opportunity here and a lot of different ways you can go about using the data insights agent. Okay, so the next agent we’ll touch upon is the audience agent. This is made available for users of Realtime CDP and AJO. And really this agent, it was designed to tackle challenges that marketers face when really building and activating audiences, especially when you’re thinking about things like configuring complex segments, rules and logics, right? Which can obviously be a time intensive process. So how does this agent help? One, it helps expedite audience management, so exploring, finding audience attributes, which again can be a very manual process. In a case like this, you can quickly make a prompt to surface relevant attributes to streamline that audience building process. You can retrieve audience insights faster. So things like audience sizes over time, changes to help make quicker decisions and audience optimizations. And then finally, there’s the actual process of audience creation. I’ll call out that this is on the roadmap, but ultimately this will allow you to create segments with specific prompts. And this includes segments that are based off of campaign objectives or propensity-based goals, which I think is personally just really exciting. So let’s go through really some of the prompts that you can use that would utilize this specific agent. So we talked about audience insights and really uncovering big changes in audience sizes. So maybe you ask what audiences have changed significantly in the past 24 hours. And then when you get the results, that affects how you treat the segment at an activation level, right? Maybe you’re taking a spiked audience and you’re raising frequency caps from a paid media perspective. When you are building new segments, maybe you want to run the segment logic against similar audiences. You can ask, do I have any audiences with identical segment logic, but different names? This could obviously help with things like segment hygiene, duplication. You might ask something like, what’s the size of my loyalty member base who have made purchases in the last 14 days? And again, uncovering quick insights at an audience level can really help inform how you then want to reach them and it affects how you message them, maybe from an email or paid media perspective. And then again, the use cases that will be available soon, but those revolve around creation. So you can ask things like, build me an audience of new prospects who browse specific product pages, or if you are wanting to make it goal-oriented, create an audience to maximize purchases across paid media tactics. So again, there’s obviously a lot you can do here, all of which can really help from an efficiency standpoint when you’re thinking about the overall topic of audience management. The journey agent is next. So this one aligns with this supports the building and optimization of journeys. And what does it do? Where does the value lie? So one, it can help with the identification of audience overlap and scheduling conflicts. This can be really helpful in preventing message oversaturation. There is drop-off analysis where the agent can detect the key points of a journey where users may fall off and then provide recommendations to boost engagement. Within that realm, it can help spot conflicts with journeys as well. I will call out as of right now, the users would still need to make the adjustments themselves based off of the agent’s feedback, but more agentic workflows with conflict resolution, they are on the roadmap. There’s prompt-based journey insights that revolve more around operational workflows. For example, show me all live journeys. And then again, this is coming, I believe very soon on the roadmap will be journey create, where you can use text prompts to generate multi-step journeys. And that’s complete with events, conditions, and events.

And similarly, what can we do with this agent, right? Where do certain prompts align with strategic use cases? Like we’ve seen with all the agents so far, we can help use them to detect and analyze audience changes within journeys. So maybe a prompt like help me understand how many profiles entered my journey in the last week. You can leverage insights and pair them with recommendations. So for example, can my journey see a conversion lift if I used a different audience or maybe a different channel. And then we get into some of the use cases from a creation perspective. Again, journey creation, it’s going to be a big one here. Help me create a journey welcoming all target users who just signed up for this webinar. Maybe you want a dynamic message adjustment. This will be available with that same release on the roadmap. So you use a prompt like swap winter sale email with spring collection due to inventory changes. And then on the previous slide, we talked about ultimately using it for the detection, but then the actual resolution of conflicts. So maybe a prompt like help me understand and adjust the overlapped audiences in my next scheduled journey. And then finally, we’re going to conclude with the product support agent. And this agent was really designed to reduce friction as it pertains to various support workflows. So ultimately the idea is to make it faster and easier to track, to identify and support different issues. And this is done from a few different angles. So one, it can be called for quick troubleshooting. So common support questions where the agent pulls from documentation pertaining to the products is a very typical use case. And if you’re seeking an issue that you want resolution on before going to support, you can use a prompt like provide me troubleshooting workflow without opening a ticket. And I will say there’s also an element of this that looks beyond product documentation to troubleshoot and combines this with what it’s seeing in your instance. So in a case like this, a prompt like why isn’t my streaming audience executing or were there any issues in my latest data ingestion workflow? These would look at the available data specific to your instance when it goes about formulating a solution. If you are finding that you’re not able to solve the issue through these methods, you can then leverage the support agent to initiate a support ticket for you. And it will also capture all the relevant context and the data to try and accelerate the resolution for you. Adding to that, it can also be used to track the status of all your tickets as well. So common prompts here we might see that align to these use cases, pretty simple, generate a support ticket for me. Maybe categorize and prioritize my support tickets based on severity. Provide me the latest updates on my open tickets. So hopefully this can be a really helpful tool for those that are in the solution every day. And again, something that can really help support workflow efficiencies and time for resolution. Okay, so with that I would like to jump into a quick demo. I will call out, so this demo it covers a lot of different areas, right? A lot of separate agents and functions across an entire marketing initiative. In real life there would obviously be more steps, approvals. I wouldn’t recommend ever creating and launching a campaign in the span of five minutes. But really the idea is to give you a broad view of what the agents can do and just how they might look in practice. So let me jump over.

Amanda, can you see my new screen with the demo? Yes, I can. Perfect. Okay, so to preface the demo, here’s the context, here’s the background. I am a marketing specialist at a large outdoor company. I am planning a new campaign surrounding weekend hiking. And really my first goal is to highlight products that customers may not have purchased. And I’m hoping the products can be used for both backpackers and day hikers. So my first step is to decide which products to include in my campaign. So I’ll want to see how they’re performing. So what I’m going to do, I’m going to type show me the revenue by product subcategory for camping in June. I want the top 20. So I will send over that prompt. And immediately I’m going to see that the Data Insights agent pulled the relevant info from CVA and even created two forms of visualized data in order of revenue. But I also see that it’s a relatively long list, right? And I would like to narrow it down a bit further. And because I want to target backpackers and day hikers, I’m going to say from the table above which ones can be used for both backpackers and day hikers. So I send the prompt and you’ll see that I get back a very helpful list of five product subcategories, all with detailed explanations as to why they might be suitable for both target audiences. So now that I know the key products I want to use, I want to start building my campaign. So I will ask it, help me plan an email marketing campaign targeting weekend hiking. And I call out that I do want to experiment with two variants. So I’m going to go ahead and send that prompt. And really this is where I think it personally starts getting fun, right? First you’ll see that it calls out the reasoning logic. And this basically shows that it understands the prompt and outlines the different parameters of the ask. Then what it does is it outlines the actual plan, which includes four different phases. So audience creation, journey creation, content creation, and performance monitoring. You can even go about and you can expand this plan and you can actually see that it includes a bunch of different subsets as well. So in this example, I review the plan, but then I do realize that it’s missing some things, right? I want to make sure, let’s say that I’m targeting at least 10,000 profiles. And also I forgot to add in my logic that I just want to target people who have viewed Backpacks Intents, but who have not made any recent purchases. So I again, I reviewed the plan. I’m going to adjust what I’m looking for and I will send that prompt. Okay, so the plan updates and now you can see that it has these different constraints around audience size and targeting. And then it asks me if I want to proceed. I review the plan one more time, but then I decide that I have stakeholders within my company that I want to consult before setting up monitoring. So I’m going to ask it to please remove the monitoring phase for now. I will do that later. And the plan is updated. I review it. It’s asking me if I want to proceed. I decide I want to move forward. But before I tell the agents that I do want to move forward, I know that it’s going to start actioning on the steps when I give it approval. So when I’m reviewing the plan, I notice that the first step is building the audience sentence. And the sub steps here include analyzing relevant XDM fields, converting requirements to segment definitions and calculating audiences. And I want to highlight the last step there because it’s important that says creating the audience after user approval. And again, I’m always going to call this out. I think it’s so important. I call it out again because what it’s doing, it’s reiterating that crucial principle that’s embedded into our AI that there is always that human oversight. So with that, I say, yes, let’s proceed with the plan.

And sure enough, I now have the audience fields and the definitions and an estimated audience size of 17,000 with a call to action of confirming the creation piece. Would you like to proceed with creating this audience? I say everything looks good. Let’s proceed and create the audience.

Okay, so perfect. I now have confirmation that the audience segment has been created. I have the audience ID, the name, the description, and everything. And then if you reference the right side there, when we go back to the overall plan, phase one is now automatically marked as complete. Okay, so now onto phase two where it asks me if I want to proceed with creating, configuring, creating the email journey. I say, yep, let’s proceed.

And here what I get is a comprehensive draft. It’s complete with entry triggers, audiences, conditions, flow logic, channels, and all the detail that I need. I review, everything looks pretty good to me. So I give the confirmation. This sounds good. Let us create the journey.

Okay, and then there it is. The journey is created. You can see that phase two on the plan is complete. And one thing I do want to call out here, this goes for the audiences too. You can pause in these workflows, right? You can exit out. You can go and check out the audiences and journeys. You can review, edit, modify. We’re obviously moving fast for the sake of this demo. But I definitely do recommend in real life you should be doing that. But again, we are moving fast here and it looks like we’re ready to proceed to the content piece. Another call out here. This is a really good example of how all the agents work together regardless of what solution it’s utilizing, right? Because in this example, the workflow, it’s around content creation. So it’s calling the content production agent, which is then executing the action through Gen Studio, right? So even though this isn’t an AEP app, again, it’s still integrated with AEP and it’s showing a cross-solution workflow. So in this case, I say, yep, what content variations should we use? So the agent thinks for a second and then it suggests two content variations that align with the themes and audiences of our campaign. And it calls out or called out similarities to other approaches and pieces of content that have performed well in the past. It asks me if I want to generate, then go ahead and generate these two pieces of content for the email journey. I say, yep, let’s generate those new pieces of content. And then there we go. I get the two different variants. Everything looks good to me. And I conclude. And you can see that on the right side, everything is complete per the plan that it put together. So obviously this is a very clean end-to-end example of a lot of different major pieces surrounding campaign planning and creation. But again, the idea here was to just give you an idea of how some of the agents come together, how they operate, what they can create. And again, super important, all while the practitioner provides oversight and direction to the different agents. Okay. I am going to go back to the slide where Amanda, just sorry to ask you one more time, does everything look good? You can see the screen? Yes, yeah, good. Awesome. Thank you so much. Okay. So with that, I want to wrap things up and just outline how to go about getting started. So the first thing I would do, I would recommend becoming familiar with what use cases the different agents handle, right? And understand how those align to solutions within and outside of AEP. So the ones on the left are the ones available through AEP and the given application licenses. The highlighted ones in green are the ones we reviewed today. And again, the data engineering and account qualification agent, they will be available soon. And then on the right side there, these are agents that you’ll likely see and hear about. They’re embedded into our customer experience strategy, accessed through the agent orchestrator, but they’re only available with licenses to those aligned solutions, right? And again, while they’re not part of AEP, AEP is able to integrate with our experience cloud solutions and hence the agents. That’s why you were able to see the content production agent demo when we asked it to create content variations. So again, that’s the first recommendation. Again, just make sure you’re familiar with what it is you have access to given your current product ownership. And then let’s just go through how to get started. So first, you’ll need to agree to and sign specific licensing terms to leverage generative AI features. If you are already using AI Assistant, it’s the same language and you should already be good to go. But then to access the agents, and remember, these are all accessed through the agent orchestrator, you will need to add an agent orchestrator SKU. So this is something we ask you to work with your account team on. They can outline the specific steps and they can get you started with that whole process. So really overall, we really do hope you take advantage of all this. I personally, I think it’s a great time to start using these features, testing, experimenting with them, and just really learn how you can better embed them into your current workflows. There is obviously a lot of value, so many different areas of opportunity. So again, we really do recommend talking to your account team about getting started if you haven’t already. And that’s it. That’s what we have for today. I hope you found this hopefully informative and that you do continue getting value out of AP and really just all the exciting AI possibilities that are out there right now and that it offers. And with that, I think we can go about checking the chat pod and I will go ahead and launch the poll we have as well.

Okay, the poll is being launched right now. Hopefully you guys can all see that.

And then let’s jump to the chat.

All right, so I see a lot of questions around how can we find out if we have access to these agents, how do we know how to get access, do we need to buy different SKUs. So hopefully the webinar, hopefully these questions were a little earlier on and you’re better understanding that now. If you had AP and you add the Agent Orchestrator SKU, you will have access to those. Again, assuming you signed the Gen-I Rider as well. And then again, I recommend familiarizing yourself with the agents that align to given solutions because there are differences there that was all outlined in the stack. So hopefully that is a little bit clearer now.

Some of these questions, let me see, can we access the Agent Orchestrator from outside Adobe? That I don’t have an answer to right now. We’ll have to get back to you on that one.

Same thing with the API access or SDK access to these agents.

I will get back to you on some of these questions. Some of these I will need to confirm.

How many journeys and campaigns are supported for measurement? That is going to eventually play into the SKU that you have for AI and Agent Orchestrator and what you’re essentially coming up with in terms of the SKU that you have.

And then I think a lot of the other questions were just really revolving around access. So that was in the chat. Let me go to the Q&A.

Okay, so only two questions in the Q&A. So that first question was also in the chat. So just about how many journeys and campaigns are supported. So again, that’s going to be based off of the SKU that you work out and you ultimately have for Agent Orchestrator. And then my organization is worried of data privacy.

So if there are data privacy concerns with how AI and AI agents are, you know, and how they work, I would recommend speaking with your account team at Adobe. We can get the right folks either within product or legal involved in conversations to really answer any questions you have about, you know, some of the privacy and legal concerns that you might have. But we definitely have people on our end that can help support those and answer any questions that you have.

So I definitely recommend getting in touch with your account team from that perspective.

And that’s it. That’s all I see from a Q&A perspective. Amanda, did I miss anything? I’m not sure if there was anything else that came up in the chat.

No, I think that you answered everything.

Well, I’ll gather a list of the questions and I will try to get those out to all the recipients within the week. And then if there’s anything urgent as well, please always reach out to your account team and we can expedite responses. That’s all I had for today. I do appreciate everybody’s time and I do hope this was helpful. And we’ll, you know, follow up with there’s recordings and access to content. So again, thank you guys very much. I enjoyed speaking with you and hope this was helpful.

I’m going