Adobe Marketo Engage Fundamentals for AI

Transcript

All righty. Well, thank you everyone for joining us. This is, as everybody could tell from the pre-chat, a very popular subject, and I’m very happy that we’re able to come together and talk about the fundamentals of AI for marketing and Marketo.

So we can jump to the next slide, please. So a couple of house rules before we get started. So there’s no self-promotion or pitching a marketing of any kind during these groups. It’s meant just for sharing of ideas. And please don’t contact the people outside of this communication. I’m sure that many people, I won’t want to speak for you guys, but as we are champions, we would love to talk to you. But please make sure to email us first to make sure that it is okay.

But Beth and Karina, please feel free to state otherwise if you don’t mind being slacked or LinkedIn requested.

So please don’t use any of their use cases or anything like that without their permission and consent.

Next, please.

All right. So as you also have seen, this is a recorded session. If you do not want to be a part of a recorded session that is completely okay, please feel free to drop now as there is going to be, not you Karina, please stay.

There is a recording that will be shared out and you can view this after the fact. So please feel free to drop if you do not want to be a part of a recorded session.

All right. So if you are interested in other things that the Marketo user groups are providing outside of the Deep Dive, please feel free to join the Marketo office hours, the champion office hours. So you can find that by going to mugs.marketo.com and in this case, the Deep Dive so that you can find any recordings including this one. We have a backlog full of fantastic content. And if you don’t have one, go ahead and create an account so that you can have access to not only the Deep Dives but all the other champion content.

All righty. So speaking of the champion content, we are currently accepting people to sign up to try and become a champion. So please, it’s not too late. I highly suggest it. Everybody else on this call can highly suggest being a part of the champion program. It is a phenomenal program. And the worst that can happen is being told no this year, and you can keep trying again and again each year. There’s a lot of people out here willing to help to try and make sure that you can eventually become a champion. But I highly suggest application still open until, I think, end of next week. Correct me if I’m wrong, Brad. But here’s the QR code. And again, can’t suggest it enough. It is a little lengthy, so I highly suggest. Oh, yeah. Right there. June 5.

So please, please go ahead and start that process as soon as you can.

All righty. So I think we passed this one. Yeah, so go ahead. And you can check the champ content if you’re unsure about that process. We do have a champion program, a mug that we did, talking specifically about it. So please reach out and have access to that content as well.

All righty. So the AI Forum is coming up on June 2 in Columbus, Ohio on June 16. So go ahead and check out those AI Forums if you’re able to join. They’re a phenomenal experience. So please check out this registration here. I know that there’s a bunch of content online as well if you’re curious as to what it enthrall entails. But yeah, go ahead and check out this QR code for the AI Forum.

Speaking of AI, next user group events, we have the application special.

And now we’re talking about the fundamental mugs on May 26 and then the virtual mug on demand expansion in Portland on June 3. So as you can see, there’s a huge amount of content coming through. Please go ahead and join virtual if you can. And then you can go from there.

Yep, so here’s some more in-person contents in Singapore, depending upon where you are. Japan has a couple of in-persons. So there’s a mixture of things coming up. I know that there’s a mega mug that they’re thinking about. Yeah, so we have the London Summit coming up on June 17. So a lot of stuff going up around the world. Again, checking out the mug content will tell you where your local one is versus all the upcoming topics. So as you can see, where AI and beyond is coming up next. All righty, so as you can see, we have quite a full agenda going in through all of this. Karina, I don’t know if you want to talk about this, if the introduction slides are coming up, but leave it up to you.

We can continue because… There it is. Perfect.

Awesome.

So Chris Kelly here. I’ll be the moderator for this wonderful group of champions. Incredibly happy to have such professionals talk about such a fantastic topic. I know that a lot of people are interested in it, so I will do my best to answer questions in the meantime as these two will be presenting on it. Karina, would you like to introduce yourself? Yeah. So my name is Karina Vidal. I am currently a manager at Bonterra. I work on marketing operations, run a global team there. I’m also a first-time champion within Marketo Adobe Engage. And then in addition to that, I am an AI enthusiast. Love it so much, love operations, and I’m here to share all those cool, fun things with you guys. So I’ll share it, pass it over to Beth.

Yeah, hi, everybody. My name is Beth Corby. I have been working in Marketo for many years. I’m also a first-time champion this year. It’s a great program, so if you want to apply, please do so. And I’m also an AI enthusiast. I’m really excited about the AI features that are coming out. I’ve been playing around with them, and I’m looking forward to presenting this to you guys today.

Fantastic.

Alrighty, I can’t take on this slide, Chris. Yeah, go ahead. But we’re all here because we really want to understand how we can thrive in an environment with AI. So we’re going to open up to a lot of experiences, and we’re going to go over some of the use cases that we think can be applicable, some unique ones. I think things that are going to work well, but ultimately, this should get you guys going, moving and grooving and understanding what do you need to set up, what do you need to start organizing within your instance, so that way you can have autonomous agents thriving in there. So as we move into the next slide, you guys will see that Marketo has always had AI capabilities. It’s not something that they recently created right now, but originally, we had to do a lot of integrations. So for those who are familiar with webhook, REST APIs, or even using IPaaS systems, this is really a way how we were able to connect some of our AI agents externally, and then connect it to Marketo, so that way it could process and do some activities Now we’ve advanced, Marketo has the MCP, so for some of those who are on beta, who are leveraging the instance, you guys have the opportunity to automatically connect to Marketo, and you’re able to navigate and view what’s happening within your campaigns. But did want to call these out, because there’s different ways of how we can use this when we connect them. Your webhooks are your foundational pieces in terms of when you’re going to push data into one of your LLM instances. In addition to that, you can also leverage APIs. So APIs have a similar functionality. If you’re one of these people who have multiple processes, or you’re a bit more advanced, and you have more agents, you can also leverage IPaaS. This could be your middleware for it to connect into one of your LLM models you’re using. So if you’re a GPT or a cloud company, these are methods of using it. And then your Marketo MCP can be directly connected to it. I will just explain though, because when people hear MCP, it is a still fairly new term being used. But in MCP, I like to use a metaphor, so that way people can visualize it. I always say think of a big, large extension core that you can plug 10 different applications to it, and it powers all these technologies with electricity. And MCP works in that same way as a model context protocol. So it’s just a process of how it responds within the environment, and how we can push data out and ingest data as well, so that way you can leverage it, especially with your agents, so that it can take certain actions on your team’s behalf.

As we move into the next slide, we’ll see here what this connection really looks like. So currently Marketo, we’re expanding from your traditional automation and making it so much more robust, where now we can have reasoning models. Usually right now, we’re very logic heavy at the moment in terms of you set up a program, you have a trigger, it listens for an event, and then actions are taken, or you schedule it. But now with the ability for us to connect AI to it, now we can actually leverage AI reasoning. So this logic is becoming so much more robust in terms of how we can process and the information that we can ingest, so that way we can take away some of that tactical component and give your team back some more strategy. This is just a quick waterfall of how the iPad, whether you’re using iPads, MCP, or web hosts, how they work. But there’s so much more to it, and we’ll talk a little bit more. One of the things you guys will notice as we continue moving on, we’ll start talking a little bit more about how we connect to Marketo, how you can push this information or data that you want into your MCP, your webhook, or your agents. Then you have your agent layer. So just want to be clear that in the moment, there are a lot of beta AI experiences that we’re also going to talk about here. But in order for these agents to work and thrive, we need to ensure that we have structures. We’ll go over what those structures are looking like. But once we have this method set up, then you have your response processing, where your agents can then do their reasoning and then function autonomously. However, this autonomous action is also going to be very heavy on your lead fields, your tokens, and campaign decisions. So we’ll talk more about what that architecture should start looking like if you don’t have it currently set up in a certain manner. But we’ll give you guys ideas and pitch to you our recommendation of what your architecture should start aligning towards so that way you can have your agents functioning autonomously. And I’ll say that over and over again.

So if we move on to the next slide.

The new thing, which is one of the most exciting pieces we’re talking at the top of the hour, is Mercado now has the ability to use AI directly into Mercado itself. So sometimes thinking about AI and agents, it seems like a lot of technical pieces, a lot of technologies and solutions. But now with some of the AI features that we have currently in beta, some members who are Mercado clients, you guys have access to it, beta has closed. So if you don’t have access to it, it’s still good because it’s a great opportunity to understand what’s soon to roll out into your instance. But we now have the opportunity to actually leverage agents directly within Mercado. So no more external creation of agents or processing. We have agents that are readily available that we can start using and leveraging. And we also have the ability to also leverage some of the basic functionality that we use for chat on external LLM models. We can use it directly within Mercado, whether it’s because we want to do some QA, we want it to build programs for us. If we want to understand some of the campaign recommendations, we can have that same communication, but it also builds for us. So we’ll talk more again, but just wanted to let you guys know what we currently have at our fingertips and what are the current possibilities that we can do if you’re an in-house builder, that you want to build your own agents or have it in a centralized solution within your team. Mercado also has the opportunity to create these things and connect them right there directly.

Next slide.

For us to start leveraging, however, all of our AIs, you guys cannot forget to sign your Gen AI consent. This allows you to really leverage AI, not just using the AI assistant, but you can leverage this beyond other assets. So if you’re using email designer, you have the ability to also use AI to create content. You can use it to create images on your company’s behalf. You also have the ability when it comes to interactive webinars to also leverage their AI features. But in order for you to get ready and using AI within Mercado, it doesn’t just start in the foundational and architectural component of your system. It also starts with signing the consent. So that way you guys can fully leverage all the features that they have available and that’s continuing to roll out. And then as we move to the next slide, nothing is perfect unless we focus on the foundation and updates that we need to do to our instance. So for those who have adopted instances that are 10, 20 years old, for those who are creating a new one, these are great things to have you start thinking. But in order for AI to thrive in any environment, whether it’s out of the box provided by Mercado or you’re setting it up yourself, we just need to make sure that we have consistent processes. We’ll talk about tokens, our naming conventions. We’ll talk about even lifecycle and program templates because after all, you want your AI to be able to work autonomously. I always love to use the phrase that AI is an extension of your team. It does not be all be all do all solution. It will only work and thrive as much as as much guardrails and context that you give it. So the most easiest way to not have to be on top of it is to ensure that you have repeatable processes, you have clear naming conventions, and you have a setup where it’s clear for it to do its own reasoning and research and know what campaigns it can leverage and which ones it can’t.

So now I’ll hand this over to Beth.

You’re muted, Beth.

Sorry about that. So yeah, that was a great tie-in, Karina. Thank you for that little slide there. I want to just talk a little bit here around why structure matters for AI. So there’s one big misconception when it comes to AI, and if we go to the next slide, we’ll talk a little bit about that.

AI does not understand the context to your business unless you teach it, unless you let it learn. So if you don’t have that structure that Karina was just talking about in place, and you have inconsistent naming conventions, random folder organization, duplicate logic, unclear lifecycle stages, broken attribution, and AI confusion overall, AI is not going to understand your business automatically. It’s going to get confused. But with that structure, that nice naming conventions, predictable hierarchy in your folder structure, reusable operational frameworks, overall a good lifecycle in place, AI learns from that, and that learns from that operational structure. So overall, AI does not create efficiency if you don’t have that strong foundation in place. So it’s super important now more than ever, it was important before, but it’s even more now that you have that structure in place, and we’ll talk about all of these individually in our presentation today. The gold star there, that’s the gold standard, that’s what we want to get to. And then we can hand it back to Karina, and she’ll talk about naming conventions.

Alrighty, we all use naming convention. It is like the foundational standard to literally generate a program because you cannot without a name. But naming convention during this era of AI is taking a significantly different use case. The naming conventions that we have currently is really used for marketing purposes. We can use it for a data point if we want to understand the acquisition program. But I want to reframe the way we’re looking at naming conventions because this is technically one of the foundational languages for AI. It’s really a way for AI to be able to classify assets, if it’s looking for related programs, if it’s seeing that there’s duplications of certain programs, or if it can leverage an existing campaign so that it can run a specific action. This is how I can start thinking autonomously without you having to be overseeing or having a significant amount of human in the loop. This is an example within this image where we’re using our internal LLM provider. But here is just a quick question that I had asked it and it searched for a program, gave it very little context aside from what I was looking for. But because of the naming convention, it was able to not only give me the name based off of how I communicated with it, but it found the exact program because of the naming convention that we have that did mention the keywords that I asked our internal LLM. In addition to that, you can see the Mercado link. This is one of the features or capabilities you have, which are Mercado MCP.

But this is a great way of how naming convention, the AI agent was able to do his own logic, his own research without me having to dictate exactly where it’s at to take actions. As we continue into the next slide.

As we in this slide, these are just examples of naming convention. I always say like, remember that these use cases can be widely different. Your naming convention can be totally different because some organizations are international, they have various markets, products, or just multiple business line. But this is just an example of how we can see how naming conventions are becoming the metadata for AI.

Recommendation always try to have your character count under 80, especially if you’re a Salesforce user. After 80, if you connect it to a campaign, it’s not going to have enough character counts to include all the names. But as you can see here, we have the traditional naming convention for our programs. However, for some who haven’t experienced or leverage naming conventions within your operational program, this is a great opportunity. Because as I mentioned before, if you connect your AI externally, so you’ll be using webhook and iPads, your MCP.

This really allows the AI agent to make calls on your behalf. So it can find certain campaigns, whether it’s a requested campaign, an executable campaign, or if it’s an operational program. With the naming convention, not only will it understand what type of campaign it is, it’ll understand its function and it’ll understand its action. So that way, it can take those recommended action on your behalf without AI going rogue and creating 50 campaign for every single prompt or anything that you’re asking it to create technical debt for your instance.

Again, some good recommendation as well is try to make sure you use one type of delimiter. So if you want to use a period, that’s one of my kettles out of the box. If you guys have noticed campaign names have assets at the end, it’s always period and then the asset name, so list.email.continue.

You can use that. You can use underscores, recommendations, underscores, just in case you have those external connectors for JSON scripts. It’s much more easier for it to read and consume. Always try to stay away from using a lot of cool funky things such as period, underscores, dashes, because that will throw off the AI. And if we move to the next slide, we’ll go over some of the explanations of what AI actually sees. Here’s just an example of what a naming convention can be. Using period as the spaces, but though we’re looking at it and we’re understanding what this naming convention is, where people are coming from the channel, AI doesn’t read it the way we read it from a human capacity. What it’s breaking down is a categorization of what each component is. For example, in the beginning, it’s seeing VAN as a brand. It sees 2026, though it’s numerical. It sees it as a year. As you continue, you can see how it breaks down things into months, quarter, channel, driver, campaign, and type. Again, if you’re communicating, you’re writing prompts, or you have a context document in terms of what is the duty or the role that your agent should be playing, this is the information that is going to use and leverage to digest it. If you have unique naming conventions or not naming conventions within your operational programs, that’s where it becomes very finicky, and you’re going to continuously have to be that human in the loop to approve every single action. I like to say, if we move into the next slide, I like to say that you have to manage your AI as an extension of your team. The way you create some type of run book or documentation for a new person to come in to read it, it’s the same way you have to treat AI, at least in the foundational stage when you’re getting your instance prepped for it. Again, if you guys have different naming conventions, that’s totally cool. You don’t have to copy exactly what we shared, but these are five core components that we highly recommend. When you’re creating your naming convention strategy, try to consider the who, what, where, when, and why. This allows not only your team, but AI as well, to understand what is the business unit or the product, where it’s located, what type of campaign or the channel, also like the timing and why in terms of what the campaign is for or the executable campaign, why is this created, what action does it take. You can play around with this, build your naming convention, but these are the core components we recommend. When you have these core components within your naming convention, in the next slides, you get to see what the abilities are within the AI agent. Again, focusing on whether it’s out of the box within our Marketo agents and the AI assistants or an external, your AI agent, when it has a great naming convention, has the opportunity to classify, retrieve, recommend, and review patterns. If someone does go outside of the naming convention or something isn’t working, it can quickly notice and find these campaigns, and then it can alert you if you have dashboards or if you set things up. It can alert you when something is not correct, if something is wrong, and ultimately, it helps with governance.

This allows you to work with your AI as that extension of your team, and think of it from this perspective. You wouldn’t micromanage your actual human team member. You definitely don’t want to repeat that process with an AI because, again, the focus of AI is to move us to being strategic rather than being tactical and having to continuously oversee its activities or what it’s doing because that’s just going to put more pressure and it’s not really the best way to leverage AI if we’re trying to move quickly, be strategic, and create fun, cool, new innovations within our organization.

So now I’ll pass this over back to Beth.

Yeah, so I’m going to go over folder structure and overall governance. These are important components. Folder structure naming conventions falls into the governance realm, but overall, you need predictable organizational context in order for AI to help scale. So if we go to the next slide, we’ll talk a little bit about the folder structure first. There’s a screenshot here of a demo instance that I use quite regularly, and it’s showing a typical naming convention aligning to the channels that are in the environment. So that is the one on the left, and then on the right is your high-level folder. So I always typically recommend keeping your folder structure limited. So I would not want to drill down four or five levels deep to find what I’m looking for. So I try to keep it as simple as possible, starting with the channel and then building on from there. If you want date folders, that’s fine. I’m okay with that, but just keeping it consistent and simple as possible.

This is going to help AI, because AI is going to rely on overall patterns and how things are organized within the Marketo instance. So make sure you create a predictable pattern. So having something very simple such as this is going to help AI tremendously. And it’s going to help AI overall just find things. So overall, AI has a harder time to recognize patterns if you just have folder structures all over the place. I remember the first time I implemented an instance long ago, and I was just throwing folders everywhere, and it just ended up in the long term where I couldn’t find anything. So again, come up with a structure. You don’t have to follow this exactly, but it just should be very simple so you’re not drilling down.

And if we go to the next slide, this is about governance. And everything we’re talking about today ties into governance. But these are the main areas that are critical in the AI era. I mean, all of it’s going to be. So organizational standards, so that’s overall how you are telling Marketo to build programs, assets, all that good stuff. And AI is going to rely on, again, that consistency and the pattern. Pattern is going to be a word you hear a lot today. User roles and permissions. So AI is going to be great. It’s going to be easy to do things. But in my mind, that’s a little bit dangerous at Marketo. So you really want to have a good strategy around who has access to what and who can do what with the AI. One of the things I’ve learned recently is the QA and review process is going to be a game changer with the AI tools that are coming out. It can really do a good detail validation around some of the things that you’re building. And it has some default rules, but you can also upload a QA document and have it do run through that document and make sure your program is set up the way it’s supposed to be. And this is overall going to just make sure that you’re following compliance, operational consistency, brand alignment, all that good stuff. And overall, it’s going to reduce risk. So you’re going to get things out more efficiently, more quickly. AI is going to help you do things faster.

And the bottom line is AI is going to execute faster than humans, but that governance that we’re talking about today is going to ensure overall that it executes responsibly. And if you go on to the next slide, I’ll hand it back to Karina.

So let’s talk about the type of campaigns. If we move into the next slide, we’ll see that we have different type of campaigns that we currently leverage within Marketo itself. However, we’re going to focus on two that may not be as popular.

We’re going to focus on executable and request campaigns. If we just click one more time.

Yay! A red box. We’ll focus on executable and requested campaigns because, again, we’re talking about how we can set up our instance for growth within Marketo and AI. So I think a strong theme within this webinar is to ensure that some of the ways that we’re thinking about how we’ve used Marketo is definitely changing. And it’s evolutionizing, which is very good because we have to start thinking more forward in terms of how we can leverage our instance and how we can optimize it within this era of AI because it doesn’t seem like AI is going to slow down. So we definitely need to get either on par with the curve of where it’s moving and ensure that we are also setting ourself up for success, not within the next six months, but within the next few years, because AI definitely is changing super fast. So if we think about executable campaigns, think about this as your orchestration layer. And then if we think about requests, we can look at this as a decision layer. So anything that’s agent-driven automation, these are going to be some of the best use cases for you to set up some guardrails with your agents as you’re functioning within your Marketo environment. So in the next slide, we’ll talk about executable and request campaigns. So for those of you guys who haven’t leveraged executable campaigns, or maybe you’ve seen in one of these older instances and are curious exactly what it does, where it lives and how it functions, executable campaigns are technically need to be set up when you’re setting up a campaign itself. So you’ll see a little tick mark and it’s executable. Once you created something unique about it is you’ll notice there’s no smart list. Executable campaigns are called by other campaigns within the flow step. Some of the benefits of having an executable campaign is that you can have a lot of processes within its flow step. So that way it processes it before it moves on to the next flow step of the parent campaign that requested it. Think about this in terms of your AI, it can be a metric for it to be able to automatically execute certain activities that you know are very common to your business practices. I’ll use demo requests as an example. So if you have a demo request and programs are going to trigger it, you can set up demo requests so that way it routes to the correct teams that you have within your instance. This is one example of consistency that you can leverage at scale. So regardless of the type of agents that you’re creating, if you expand it and grow much larger, you know that your agents can easily find these campaigns for the same process and execute, again, without creating technical debt of having to create multiple campaigns for every single action that it needs to take. In addition to that, we have the request campaign.

The request campaign gives you the ability to launch any decisionings within the flow step as well. I think pairing request campaigns and executable campaigns are one of the best. Because again, if your AI agent has to run through multiple process and it can again, leveraging naming convention, know exactly which campaign it can trigger for another waterfall workflow, then this is one of the best situations where you can leverage it. Again, you call upon these within the flow step. Request campaign is different from executable campaigns because you do have to set it up within its campaign so it listens for the trigger of when another campaign is requesting it. But this could be part of the workflows that you leverage for decisioning. If we move into the next flow.

Awesome. Here’s a great example of how these two can be mixed together. So when you have executable campaigns or request campaigns, your AI has the infrastructure of where it can work within. So in this example, I’m just using one of our AI agents that we use externally so that way it can review our inboxes. It can review inbox responses and then categorize people based off of their marketability. But within here, we have request campaign. So instead of having one campaign shooting out or a trigger listening for a change of an update, we’re processing things within one area. But again, this is so hands off. The agent is the one doing its own reasoning, it’s reviewing the inbox, it’s categorizing them, and then it’s calling upon Mercado to trigger the actions within the workflows themselves. And then it also goes through the request campaign, and the request campaign itself runs through its own flows, and it triggers the correct executable campaign depending on the situation of the email that the AI used for its own reasoning model. So this is just one example, but you can obviously multiply this within your architecture. It’s a great way for you to have these guardrails if you don’t want your AI going rogue, and it’s a great way for you to take that human of the loop component outside so that way you can allow your AI to continue processing on behalf of your team.

Now if we move to the next slide, we can talk about center of excellence and the benefits of combining this with your campaigns.

Yeah, so the center of excellence is your framework, and your business is developing programs. And also again helps AI see those patterns, see how things are built, how you’re using them. So if we go to the next slide, we have an example here of a standard center of excellence that has every type of program in it, and it’s built in a way that’s going to, that people can clone and easily make things repeatable. So the idea is creating that consistency across your team so when they have an email send, they all use the same program template, and it’s all set up the same way. And it’s also going to help AI, if we go to the next slide, based on many things. So AI is going to be able to see these repeatable structures and look at these structures, and when you ask it to create a program, it’s going to create it the way that you want it to be created instead of just building something random. It’s going to make it overall reusable, so a reusable framework across teams. So with that reusable framework, it’s going to make your team aligned. Having that center of excellence helps the overall way that you’re using Marketo be documented in a way, so it will provide that operational guidance to your AI.

And overall, just make things scalable, make it easy and quicker to do things like an email send or set up a webinar program or something like that. So that consistency is needed for AI-assisted operations, so it’s really important. Center of excellence is one of the main things I always push for organizations to do. And then we can hand it back to Karina, and she will go into tokens.

So we love tokens, right? I know at Adobe Summit, they were passing out shirts, and one of the phrases I really wish we had was, what are you talking about? Because tokens are literally the coolest things, and now they become even more cooler as we leverage them for AI. So let’s go to the next slide. In the next slide, you guys will see an example of tokens. So again, going back to structure, going back to context, the big frame idea is how can AI be identified? How can it be autonomous? How can it take your team out from doing a lot of tedious tactical work and really focus on strategy, really think of developing as our teams continue to innovate and revolutionize how we’re being seen within an organization. This is, again, one of the best low-hanging fruits that you can leverage, especially for those people who have been dabbling into creating AI-generated copy.

This is a great opportunity. So here again, we’re trying to make sure that we’re setting up AI to have reusable processes, and it can understand the information we’re providing it, and it knows what to look for. Little human in the loop. This, again, is super helpful going back to request campaign, executable campaigns, a thing I forgot to mention about executable campaigns that you can use tokens from the parent program.

Again, think about structures, how you’d want to set up, especially for those who are dabbling into AI-generated email copies for your email team. This is a great opportunity of how you can set up your programs to ensure that it’s very clear in terms where you want to inject certain AI-generated copy, whether it’s the subject line in this example, or if it’s the copy itself within the email body, or even images if you have images that are readily created. This is an area that you can expand it, but it’s so much more because when we think about tokens, I want to say traditionally, some may have experienced it differently, but when we think about tokens, especially if you’re adopting a Marketo instance, the first tokens we think of is personalization and then program tokens. It’s the easiest ones to use, it’s the quickest ones to leverage, and we understand its value. However, as we move to the next slide, you’ll see that there is a plethora of type of tokens that we can leverage. And within these tokens, we can use my tokens, which is the one we’re most commonly used to, your person tokens, we have member system tokens, but one of the tokens I do really want to call out because it has a lot of value when it comes to AI is the Velocity Script tokens. These are really one of the tokens that you can leverage. Again, if you want to create something identified, an email copy, create an email with under five minutes, even if you’re using Marketo’s in-house AI agents or the AI chat for a marketing brief, these tokens can be leveraged so that way it can pump this information that’s being generated and added automatically for you on your behalf into your programs themselves, the copy or the assets, wherever these tokens exist. If you have your executable campaign or request campaign structure, this again can be tokens that it can leverage to either send out an email or conduct the process that you need leveraging these tokens. So now we move to the next slide. We’ll talk about how all of this comes together. So just to help any of the visual learners or people to understand the processes of how this mechanism can work like, you have here for an example, I’m using the Marketo MCP, but you have the my tokens that you can set up. You can have this within your MCP so that your MCP understands the program the agent needs to clone and the tokens that it needs to pump information for. If you have an agent that creates copy for you, if you have an agent that does A&B testing, or if you just have an agent that’s doing QA within the copy to make sure that it has the tone and voice of your organization, within the MCP, it can easily call upon these programs and it knows exactly what tokens it can leverage. Again, you can use Context Docs for this, or if your tokens and your naming convention are very clear in what it is and what it does and how it could be used, then your AI can use that reasoning model to be able to decipher and leverage these tools at its disposal. So once it moves into an LLM agent, again, whether internally, externally, then it has the ability to make the updates as needed. And then the final stage is good old Marketo. We can have Marketo execute on our behalf by having the agent actually turn on the campaign or schedule the campaign, or you can do that final human in the loop review for the final QA before anything is turned on, triggered or sent. Then we’re going to talk about another example for people who may want to start off with Marketo first. I’ll let Beth talk a little bit more of what that experience within out-of-the-box solutions of Marketo’s look like.

In the next slide.

Yeah, so this is a little snippet from a webinar I did earlier this week on the new AI tool in Marketo. And what I’m doing here is I’m creating a program by prompting through a conversational, and I know it’s going fast.

And then once I did that, I uploaded a campaign brief with the tokens I wanted to use in the email, and it updated the email for me. And then finally, I ran a QA prompt where it essentially QA’d my, for instance, so it was 100% failed, but it showed me all the things that I needed to update in that program before I would send it.

Yeah, if you want to check that out, just look for me on LinkedIn and you can find it. But basically, it was a great, it was a demo overall of how that AI agent works in the Marketo builder. So it was fun.

Awesome. So on the next slide, we, again, want to make sure that we’re setting up Marketo and any AI capabilities to be identified. Because again, this is going to allow us to not be fully in the loop of every single process or decisioning. If we just focus on the setup that we have or the foundational, whether it’s an AME convention, our tokens, architecture from the campaigns or the folder structure, then we sit in a place where we could have our agents running and doing processes on our behalf without us having to continuously be reviewing, updating, refining all of the processes, just to make sure we can get one specific outcome. Again, this is all for scalability to enhance and scale, but this all becomes that metadata that your AI needs, so that way you can function autonomously. And now I will transition this back to Beth, so we can speak about lifecycle.

Yeah, so I’m going to talk about channel and lifecycle configurations. These are kind of two of my favorite areas in Marketo. And I think it’s really important to AI to be able to see these things within your org so that you have overall a good way to help AI make more reliable decisions.

And if we go to the next slide, we’ll talk about channels first. So channels is a setting in the admin section called tags. And these are your overall marketing initiatives. So these are the at a high level, the things that you’re going to be tracking in Marketo. And within those channels, you have statuses. So like for an email send, you’re going to track when that email’s opened, clicked, what was the success and the unsubscribed. The important thing within those statuses is tracking that success.

And that’s overall those statuses and those channels are going to give AI context around how people are engaging in your different programs. So when AI can see that, it’s overall going to be able to be like, hey, this is working and this isn’t maybe that great. Maybe we should optimize it. In tandem, another thing that’s going to be important, if we go to the next slide, is your lifecycle.

So this is an example of a simple lead lifecycle. And being able to see where people are within your funnel and having this foundation in your Marketo instance is to me a game changer because you’re going to speak to these people differently based on where they are in the funnel. So if AI knows where somebody is, it can give recommendations around what messaging you may want to give somebody in the customer journey and what should happen next. And then to tie it all together in the next slide, overall, having these two things, channels will explain to AI how someone is gauging, and the lifecycle will explain where they are in the customer journey. So ultimately, you’re going to support things like smarter orchestration, be able to identify customer progression, and improve your segments and who you’re talking and how you’re talking to people. So very important items to have a structure for is your channels and lifecycle. And then I will pass it back to Karina on the next slide.

Oh, I’m sorry, this is mine.

The future of AI in Marketo. So we talked about a lot. We talked about structure. We talked about really base things in Marketo, but they’re really important. They were important before, but again, they’re even more important now. So if we go to the next slide, we kind of have a vision. We’re looking here at traditional marketing automation, how things were very manual. We had static workflows. Everything was rules-based automation. And with the recent Adobe Summit announcements and the roadmap, we’re seeing a transition phase. We’re starting to see AI-assisted workflows, conversational interfaces that you saw during the tokens area, smarter QA and validation, connected ecosystems, and AI-supported operations. And that just takes us to the future vision. We’re going to have a lot of AI-driven orchestration, autonomous operational support, connected AI ecosystems. But one thing that’s important to mention here, and I hear it a lot, is that a lot of people are scared AI is going to replace people. I don’t think that’s necessarily true. I don’t agree with that. I think the future of Marketo AI is not replacing marketers. It’s reducing that operational friction and making things smarter, faster, and giving more scalable execution overall. So I hate to hear that people think that AI is replacing you. You need humans for these things, for sure.

So if we go to the next slide, we’ll talk about the roadmap. The most exciting sessions within Adobe Summit is the one I always look forward to the most. The main highlights when it comes to AI and some other things is the Marketo AI conversational area. So you saw a quick demo of that within that video. But that’s where you’re going to be able to prompt some of these things and build some of these things, do little things like little audits, just the possibilities are endless. It’s really exciting to play around with this stuff. Haulable agents is going to be cool. So this is where you’re going to have workflows in your flow and you can call an agent. So let’s say you have a fills out form trigger. That person comes in, they get their program status updated, and then you have a flow that calls an agent to clean up any data that came in with that form fill that wasn’t good. So you’re going to be able to call agents for some of that data cleansing, normalization, and overall other little workflows. And then the one a lot of people are most excited about is the MCP server. So I got MCP connected again yesterday and I was playing around and I asked it to audit my smart campaigns through Claude. And it was really cool to see Claude pull a report of my smart campaign. So a lot of people are super excited about that and to be able to use your own AI platform like Claude, ChatTPT, and Gemini. They’re also going to be coming out with CRM Sync V2, faster syncs, real time events, improved scalability. If anybody has seen a backlog, they know they’re going to love to see this come out. So a lot of cool things came out of Summit. And then some enhancements to the email editor and dynamic chat. So the biggest theme overall is, again, that enabling teams to have more connected, intelligent workflows through their day. So a lot of exciting stuff.

And if we go on to the next slide, Karina will talk about use cases. Yeah. So let’s talk about use cases because common thread, I also believe AI will not replace human. In fact, interestingly enough, at Summit, Microsoft presented a gigantic organization. And one of the key points that they noticed is as they implemented AI within their business infrastructure, they had to hire more people to maintain it. So this just brings me back to an interesting quote that I heard a year or two ago at another conference where the speaker literally said that AI won’t replace humans, but the humans who know how to use it will replace other humans, which is why here we’re going to focus a lot on how we can leverage AI and the infrastructure. So AI, we have it readily available in Marketo. If we go back to the original to the other slide is when I point out something in terms of how we can use it. So use cases vary. I recommend if you have a use case, start with the problem within your organization. If you start with what’s trendy, what’s popular, what you’re reading, you’re not going to be able to report back the ROI to your organization, nor be able to amplify the importance and use cases that’s being created for the investments of these AIs. In this example, we talk about operational efficiency as it comes for our marketing or technical teams. This is really where I think we’re going to see the most valleys within it. One of these examples that I have here is just the ability for teams to be able to create smart lists, but not your traditional smart list of I want to tackle or go for a target market within these parameters. It’s really allowing AI to be able to understand and leverage certain filters within Marketo and understand which ones are the ones not widely used, which are the filters where we can tap into a different market that we just haven’t sent communication to, because our team is consistently using the same exact parameters across every single campaign. But this is an opportunity of how AI can not only do some analysis and review what’s happening within our Marketo instance. Again, this example here is a Marketo MCP. It has the ability to actually go in and build a smart list from scratch and then explain what’s his reasoning and why it’s being used. So again, it opens a much larger capability within your team, especially if you have newer teams or you have marketing members who are in your Marketo instance, they don’t have to understand what the 200 plus filters mean. They also don’t have to worry about understanding they’re going for the right account target market, they can literally leverage AI to do this reasoning and understanding and analysis for them to be able to pivot and then focus on what’s more important, which is their strategy.

In the next example, we also have some of the out of the box capabilities. For those who have started using email designers, if you guys haven’t, I highly recommend it is so amazing. Because you have a lot of components that you would use externally, such as like other copywriting tools, you have it embedded automatically here within the email designer. So you have the ability to create images, you also have the ability to create your email copies based on the voice and based on the call to action, and even the persona that you’re trying to tailor to. So email generation AI has become like a big higher topic because it instantly takes away hours of creating copies or having to create marketing briefs and it creates it automatically, either in this example with email designer or the previous example that Beth shared, in terms of using the AI assistant to share a marketing brief document that just created the entire program and the emails themselves. And then in the next example, we’ll run through what these other setups as well can look like for leveraging AI.

And the next slide.

So the next piece is using AI for decisioning and optimizations of your campaigns themselves. So again, it’s beneficial to your technical marketing operations team, but it’s also beneficial overall to the organization. And in this example, we’re talking about the analyzation, what can take days, if not months to set up a report, or even digging for QA and trying to understand like, if there’s a problem, why is there an issue with a certain record account or group of people, you can leverage AI to do that decisioning, understanding the analysis, and then to take the appropriate next steps. This example, I’m just using one example to understand again, like what this action looks like when you have an identified experience, whether you’re looking at the funnel velocity, you want to understand the engagement, but your AI has the opportunity when connected, again, internally from Mercado, or if you have an external via API Webhook MCP, the agent has the ability to analyze when there’s low engagement within Mercado, it can then trigger a campaign to initiate any type of processes, you can use your executable campaigns in this instance as well. But ultimately, your AI has the ability to launch emails or whatever type of engagement strategy you’d like for it to leverage. So again, good guardrails, user requests and executable campaigns for this. But if you’re more free, and you want your AI to make these decisions and be that final step at the end of QA, you can definitely do that as well. But these are use cases that you guys can look and implement again, based on your business problems that you’re noticing, that you can easily fix and provide significant, very strong use cases for repeatable processes. So now from this, let’s transition now to all your checklist and what you’ll need to ensure you have this set up properly.

Yeah, so we wanted to make sure you guys had a good takeaway. So if we go to the next slide, we put together a checklist.

And overall, I want people to understand that similar to like, if you’re implementing a tool, you need you want to have a strategy in place. So we put on this checklist, the overall strategy, so making sure that you have defined AI use cases and business goals aligned to this, that you have clear operational ownership and governance, like we’ve been talking about, we want to make sure that you have rules in place for how AI is going to do building for you. And the alignment across the teams and your organization is going to be super important for that. So just like implementing a tool or a platform, you want to make sure that strategy is in place, your teams are aligned. And then it goes into everything we’ve been talking about around structure and organization, where you have your naming conventions, your folder structures, your program templates, operational foundations, like the lifecycle and the channels, and then overall governance and the ability to scale and make things repeatable. So we wanted you to have a takeaway. Feel free to take a little screenshot of it if you want. This is your checklist to make sure that you’re AI ready, especially with these new tools that are coming out through Adobe. So we hope you find this helpful for sure. And then we can dive right into the Q&A if we have time.

And the next slide.

Before we get going, I know that this has showed up a couple of times. There will be a recording of this along with slides available on this webinar page later today once everything has been uploaded. So if you’re looking for that content, just come back to this site and you’ll be able to access it whenever you need.

All right. So some of the questions that we’ve been seeing have been just suggestions along the lines of where can we get additional information about these new pieces of content, such as the email builder or any other pieces of access, such as the beta coming out for MCP and things like that. So I know that there have been a couple of responses directly to that here in the chat, but does neither of you guys want to have any responses to it? My thing is to keep a close eye on the Marketo user groups. There’s a lot of people doing demos of things and look for some of the champions on LinkedIn. We have stuff for posting.

I’m going to be doing some more webinars in the future at some point. So the Marketo user groups and the community is always the first thing I will say, but also keep an eye out for the champ stuff on LinkedIn and stuff. Yeah. I also I specifically about the email designer. Marketo has a lot of learning paths. I think they’re becoming more popular now, but they specifically have an email designer learning path, which is very good because it drives through all the components of how to set up an email designers and explains what its fragments or tools or sorry, not tools, but variables within it are. Kimberly also mentioned there is an office hours as well. I recommend you reach out to your CS. They can provide you the link for you to be able to sign up, but it’s very good hands on with the product team of being able to set up or ask your questions or understand how this thing, how the email designer works.

And then you have the experience league. Definitely leverage it. There’s a plethora of information, whether it comes to understanding what AI use cases could be, how you can leverage email designer or other solutions. Definitely important to learn to understand how you can leverage them. And these are great mediums for you to just do it on your own. Just go and explore or reach out to your CS and they’ll direct you to more specific links that you can leverage.

Yes. Thank you so much, everyone. I know that there’s so many questions about this content and I know that there’s going to be more content coming along the lines with more deep dives and more.

Champion office hours. Again, office hours are also a great place to come if you have any other questions about certain topics. Doesn’t have to be AI, but always happy to answer them. And if we can’t point you in the direction of some really good content or documentation and the like. So thank you again, everyone. I know that we’re a little bit past. So thank you for staying with us and enjoy your time. It’s going to be an exciting couple of months coming up. So thank you all.

Bye, guys. Thank you. Bye. See you guys.