← Marketing Trends
Marketing Trends
What Is an AI Agent? The Plain English Explainer for Marketers
00:24:29
Stephanie is joined by guest cohost Lacey Peace to break down what AI agents actually are in plain English, why they are fundamentally different from automation, and what this shift means for marketing jobs and the skills that will matter most going forward.
Loading summary
Transcript37 lines
- [00:01]
A
Have you ever wished that your entire video production workflow lived in one place? Well, you're in luck because that is ltx, the creative suite that was built by Light tricks, all for AI video production. It takes you from idea to final 4K video all in one workspace. And under the hood, you've got LTX2, the next generation creative engine, powering native 4K synchronized audio and cinematic quality three fast. So if you make content professionally, this isn't just another tool. This is your new creative home. Go explore it at LTX Studio. A CMO from a multi billion dollar software company. He told me that their 2026 marketing plan already is obsolete in January. I've never heard one month in all the work that we did. It doesn't matter anymore. Like we had to scrap the whole playbook. AI agents.
- [00:52]
B
AI agents.
- [00:53]
A
AI agent. It's the biggest technical thing ever in my lifetime. I mean, it is so profound. AI agents, they are like software workers that can complete tasks without step by step instructions working in the background for you. Definitely jobs will be lost. Hey everyone. Welcome back to another episode of Marketing Trends. I'm your Host, Stephanie Postols, CEO and founder of Mission.org and today I'm excited. I'm eager because I brought on my VP of content, Lacy Peace, to join me as a guest, as a co host to go through everything around AI agents. So this has been a big topic on the show in the past couple episodes and I thought today would be a great day to go through exactly what they are, how to be thinking about them. And Lacy's been diving into it too. So Lacy, welcome.
- [01:44]
B
Thank you. Thank you, Stephanie. Always a pleasure. And I'm so excited to do this today. It's something that has been coming up across the board on all of our different shows, whether it's tech experience, marketing. So being able to have like one place to direct people to like, hey, what is an AI agent? What does it do? How does it function? How should I be thinking about it? I think it's gonna be really valuable.
- [02:05]
A
Yeah. And especially how to be thinking about it for your marketing org. So it was interesting the other day I was talking to a CMO from a multi billion dollar software company. I think it's like six or six billion dollars maybe in revenue. So huge software company, very cutting edge. And he told me that their 2026 marketing plan already is obsolete in January. Like it was already wasn't going to work anymore, which I don't think I've ever heard that I've heard like quarters Being delayed and maybe midway through adjusting. But I've never heard one month in all the work that we did. It doesn't matter anymore. Like we had to scrap the whole playbook.
- [02:42]
B
Yeah, well, it's not even just obs, like so I like that he used the word obsolete because a lot of times you maybe you just don't hit your goal, right? So it's like at the end of the quarter and like, oh, shoot, this goal is still relevant for later on in the year. We just didn't quite hit our target. But literally like the goals that they had set no longer matter. That's wild.
- [03:01]
A
Yep. And the, the speed of transformation too, I think is the interesting piece because it was in a two week period that it quickly changed. And so a lot of times when we see tech coming in and different transformations and all that, you can kind of see it coming. It could take months, years. But this was a two week time period that threw out their entire marketing strategy for the whole year. And that's why we're on for today, to figure out like, okay, what changed in two weeks? That is now making every marketing team, every CMO question what they're going to be doing this year. And it's all around the topic of AI agents, agentic workforces, AI Org charts, all of that.
- [03:37]
B
So it's funny because we've been talking about this for a year now across all of our shows, right? We've heard the word like agentic AI to death that we don't even want to like say that phrase anymore. But Steph, what happened in the last two weeks that really like triggered something in all of us that were like, oh, this is real now.
- [03:54]
A
Yeah, I mean it is interesting because like you said, we've been hearing about it for a long time, but I think that there hasn't been an ability to really see or understand what does that mean, especially to a non technical audience. So yes, people have been talking about it. Yes, there's been a lot of technological improvements, but I think especially when you think about it from a marketing team perspective or even from a CEO perspective, it hasn't been put into a story or into a workflow in a way that hit where it's like, oh, that's like, if you can do that, that changes every assumption, every framework I've ever thought before. So I think it was a lot of the visuals that started coming out. Openclaw came out, which originally was claudebot and then they changed their name to openclaw and people got to start seeing like this is what can happen very rapidly to scale up a whole workforce of AI agents. And I think those visuals is what made everyone start thinking like, not only is this new technology, but now I can also see how it can change the entire way I think about my company, my org, how we're working as a team. And so yeah, openclaw was really the big game changer that came out. Even though there's been many things so far. I mean, I know you've played around with cloud. Claude Cowork, was that also in the past two weeks or has that last two weeks? Yeah, okay, that too.
- [05:09]
B
That hasn't been even been that much. Yeah, it's all happening like right now. I think everyone's releasing a kind of similar products though I have been seeing quite a difference in what their capabilities are.
- [05:19]
A
Yeah. So you've been diving more into Claude Cowork, where I've been looking more into openclaw, which is interesting, you and I both going and diving into the rabbit holes and then coming back up and being like, what can you do? What can you do? Whoa. Okay.
- [05:32]
B
Like we need what broke, what worked. Yeah. Before we get into any that, I think we need to set the table of what even is an AI agent. We've heard it again so many times on all of our shows, but I haven't gotten a concise, clear description of what it is. So Stephanie, could you walk us through that?
- [05:50]
A
Yeah, so I'm going to do it in the non technical way because that's what works for me. And I'm not going to get into all the details, but I will say AI agents, they are like software workers that can complete tasks without step by step instructions working in the background for you. So if you think about ChatGPT, that's like asking a freelancer a question, whereas an AI agent is like hiring someone and then giving them an outcome or a goal and then it's working to complete that goal. So you don't have to tell it step by step of how to do it. And here's the framework and all this. You're just saying let's take it from like a marketing perspective. Instead of setting up all these automations and systems and telling it do this, then this and then you've got Zapier and yes, no, you don't do any of that. Instead you can say something like launch a Webinar campaign targeting CMOs and optimize for registrations and then you let it get to work and you make sure of course that it has access to all the platforms and the APIs. So it can go in and maybe it's going to go into YouTube, go into the ad platform. It already has those accesses and the, and the API keys to start pulling that data. But it's outcome oriented, it's not based off you telling it. Here's the five steps of how it needs to be done. It's doing it for you and then on top of that, after it completes it, it'll show you and it'll recursively learn each time so it can get better every time. Where you can look at it and be like, yep, this is exactly what I want. And now make it even better because I want the CAC to go down, for example, once again, outcome goal oriented. And it's learning and getting better every single time. And it's not just relying on your human brain to tell it exactly what to do. So it's a very different way of thinking where, you know, you and I are used to telling someone, telling a system like this is exactly how you have to do it. And so it's really getting out of that practice and more saying like, this is the outcome. And I'm going to orchestrate this by watching over the AI agents and making sure they're doing it in a way that's getting to that outcome and then making it better and better each time.
- [07:48]
B
I mean, it's like having a really good. Well, no, it's great. I mean it's like having a really good employee. Right? Because if you're able to delegate out to an employee, hey, there's this outcome I want to achieve. You don't need to think about the step by step process of how it is actually done. They might come to you and say, hey, I need access to this thing or I need to be, you know, be able to log into whatever. But it's not actually relying on you to help drive that outcome. Or like, like again, a really good employee is versus maybe like a virtual assistant who can follow an SOP really strongly. That's what previous automations were like, you know, do this, then do that, then do this. But when you watch it do the actual thinking process, it is thinking more like a human. And there's stuff that it throws in there that I would have never prompted it. That's super helpful. Or I can pause it and say, hey, actually don't do that, do it this way.
- [08:37]
A
Yep, yeah. One thing that I've noticed over the past couple weeks of doing, you know, different interviews is some of these folks I'm talking to are mixing up automations and AI agents. And so I think also making sure, like making sure to clarify the difference between those two because I think what a lot of companies have set up, set up so far is automations for things. They're like, oh yeah, I'm already doing AI automations and that's not what this is. So that's what I feel like I'm having to tell people. Like, yes, that's awesome. But this is not what this is like. Automation is rule based and agents are goal based. And so there's two very different things.
- [09:13]
B
And it is a mindset shift that I even have to audit myself. Like, while I was working with these tools the last couple weeks, I consistently, consistently found myself in that mode of like trying to give it step by steps of how I would do something and not just starting with the outcome and letting it sort of find its way on its own. And then maybe me putting some guardrails or like supplementing it with a little bit more information. So it's, it's, it's like the ultimate delegation, right? It's just like what you and I used to talk about with the who, not how book that we read. To be able to think about how to use these AI agents effectively, you need to be an expert delegator, but not necessarily an expert at recording yourself do an SOP step by step.
- [09:55]
A
Yep. Yeah. I think one thing I've had to untrain myself as well when it comes to this space is not treating, let's just say Open Claw or any of these agentic tools, not treating it like ChatGPT, because I think like I've been hearing, you know, for so long, like you gotta have good prompts and make sure to give it a Persona. And this. Yes, like, you definitely had to do that with ChatGPT. But it was a, it's a hard pattern at least for me to break of like not going into these other tools and treating it like ChatGPT. And then I kind of would find myself getting stuck of like, well, what can I ask it for? And so it definitely takes just a different mindset where I'm like, okay, I don't want to just use it like that because it has so much more functionality. So then you have to kind of like go out there and explore and figure out, okay, what, how can I use this, what can I ask it to do? And get out of like the old way of thinking of what we've been used to with ChatGPT.
- [10:45]
B
No, that's a, that's a super Strong point. But I also think it underlines a broader business problem. Right. Is is in order for me to come up with the outcomes I need from this tool, I actually need to sit and think about what are my business outcomes that I'm trying to drive at and make sure that the. Whether it's a human employee that I'm trying to train or an agentic employee that I'm working with, both of those, you know, people, beings, things need to be aligned on what we're actually working towards. And it's been so much easier as a manager to, to kind of get caught up in the how and not the why. Feel like you're really successful because you're just like this expert executor. But that's not going to be valued nearly as much anymore because you don't need to know the how. What's going to be really differential, what's going to really differentiate people's careers and businesses is coming up with that why and being able to strategically apply it.
- [11:38]
A
Yep. Yeah. Which I think is a great point. And I know I've had quite a few marketers ask me of like, is this going to take my job? Or how do I know if it's going to. And in one way I'm like, yeah, probably. Unless you can start looking at what you're doing and if everything you're doing is based off execution and tasks and doing things manually or just moving things around, moving something to a spreadsheet and then moving it into, let's say, HubSpot and whatever it might be like, yes, but if you can zoom out and figure out what kind of system can I architect around what I want to build and start thinking more strategic. And like you said, tying it to business outcomes like that will never go away. And that's where you can add, you know, things that no, no technology could ever do. Which comes, which comes down to like, human taste. Like everyone's is different. You. And you probably won't know what my taste is until I say it. And then it surprises me too. I'm like, can't rebuild it. That's only me. Intuition, judgment, strategy. I mean, I think there's a lot of like, really good human elements that get to come online now in a way they couldn't before. Because a lot of these redundant tasks are going to be, you know, taken care of by these agents.
- [12:48]
B
Do you like to use that word, intuition in there? And I know that in like the groups we run with in Austin, that's a common word, but how would you Define that. For someone who's a business leader, that's like intuition. Like, you mean my gut? Like I should just be trusting my gut.
- [13:05]
A
I mean, I've been saying this for a couple of years. It only hits sometimes with certain guests who come on and other ones are like, no data. I'm like, okay, so I mean, yes, when I think of intuition with decision making, I do think it is that gut feeling. But oftentimes it's also based off like innate wisdom that only you have. So the wisdom that I have, no computer could really as of now do the exact same like decision making as me because it doesn't know everything I've been through in my entire life unless I had a first party camera on. That also tapped into my nervous system and what was happening in there and out there all around me.
- [13:42]
B
Like, and your past generations and like all the trauma.
- [13:44]
A
That was scary.
- [13:45]
B
Yeah, yeah, yeah, yeah.
- [13:47]
A
Who knows? Like, so you really can't have the wisdom that I have. Like every human has their own innate wisdom based off the life experiences that you, you know, have gotten that from. And so that's what I think is really beautiful about where we're heading right now is you get to use that more of. And sometimes that does look like, I don't know why I think that's good. Like the data actually doesn't say it. Like it's a different way of thinking and making judgment calls. That that's why people are going to value specific people for like, okay, your, your judgment's good. How you're thinking about position positioning, this is different. Your taste. Like, I don't know how to explain it, but it's really good. And you're able to see the different systems of what it takes to put all this together. And so it's a fun place for people who want to play in that area. And it can be scary for, you know, the ones who are like. But I like doing the, you know, AB testing and looking at it myself and creating the different campaigns and the different visuals, like all of that's going to be gone. And so if that's the area that you're hanging out in, like, yes, that probably will go away. Just like, I mean, if you look at history of like different job roles, I mean, demand gen at one point has changed drastically because it used to be, you know, email, like expert or email optimization expert was one job. So yeah, before these marketing automation platforms came online, like HubSpot and others, demand gen, you know, was very fragmented into role based into roles like campaign manager Email specialist, a list manager, an analytics person. So everything was very manual and it was very heavy, execution focused and also very siloed. And then new roles did emerge, but it was more like CRM architect and revenue operations. So it was much more like zoomed out, more visionary. You now all of a sudden have these automations and tools that are working for you so you can instead focus on the strategic decisions that are going to drive the business forward instead of like, you know, optimizing an email list.
- [15:43]
B
Yeah, no, I mean it makes sense but I, I want, I asked, I asked this a lot on our, my, my show experts of experience about like is there still value in understanding how to actually do those tasks though? Right? Like, is there still, if we get into a future stuff where there is a bunch of these AI agents do executing on things, do I even need to know how to do email optimization? Is that still going to be valuable?
- [16:09]
A
Yeah, I mean it's like people being like, you don't know calculus anymore. Yeah, I'm fine with it. I don't need to go and start figuring it out now, like why? Tell me why I need to. And I know many people might not agree with that take. But I think about that for email optimization, if, if you have like 10,000 decisions being made, buy a computer behind the scenes to figure out what's working best. If it's working, then do you need to know exactly all the steps that it went through and all the things that it looked at to figure out what was working and what wasn't? Especially knowing that you might be in an environment that's changing very quickly. Like, do you constantly need to understand. I don't know because I know some people don't like the black box type of angle that you know, AI has brought about of like what's actually happening. I need to know how the sausage is getting made. But when I, I know which I have it too. I have it too. But if you look at at least the past decade or so, like the best technologies have gone behind the scenes and you like, no one knows, unless you're really into it, how blockchain fully works. Like people like us. Yes, if you're like in that space. But do we need to know or I don't know, like do we need to know all the details of how things are working now? Maybe others would argue with me on this, but I just don't think so. I think focusing more on if something's working for you and you kind of know how it's doing things in the Background, but not all the details of it. I think it's okay to not know. What do you think?
- [17:31]
B
I mean again, if I reframe my thinking of this isn't a tool. This is like an employee that I trust. I'm not going to look at what my employee that I trust is doing every minute of their day and audit that and understand that I'm going to trust that they're doing it effectively. So that's, I think, I think we need to think about it like that there are things like in order for me to properly provide feedback to that employee, I need to understand loosely what the job is and why they might make a decision that they are making. But I don't think I need to know like how to actually log in and look at the email optimization tool. Right. So I think it's still going to be vitally important for leaders to know what questions to ask. And, and that comes from experience. But that doesn't mean you need to have the actual like hands on experience that maybe a current expert would have in a certain domain. You mentioned like, I know we talked about intuition and whatnot, but you mentioned taste and things like that. All that stuff gets defined by usually the marketing team. Right. Like what's the appearance or brand storytelling. So spending a lot more time on defining those things in a way that's going to differentiate your company and make them stand out and less time on the actual how to like run ads or something. I think it's going to lead to like a lot stronger businesses.
- [18:59]
A
Well, I would wonder though if you think about, you know, your brand and focusing on that like if you're now appealing to an agent buyer like they might, you know. Yeah. So that I was just talking about this on the interview the other day of like does brand or with the Twilio cml like does brand even matter anymore? Maybe, but it just probably has to be thought about differently knowing that you're just going to be telling your agent, hey, go buy this thing, look it up, blah blah, and it's like gonna not only just go to your website, it's gonna look at G2 and all these other places and then be like, so it's a different way of buying now. So then how does taste come into play? It's probably more from a structural thing, a team thing, a strategic decision making frame. Maybe, maybe.
- [19:50]
B
I think it depends on it's like steps. Right. Because I don't think we're yet at the point of my AI agent talking to your agent. Whenever I've Got something going on that's wrong. I still have to call FedEx because the box didn't show up. There's still an element of me liking a brand and wanting to work with them. I don't know that that becomes completely. Because then. Then you're basically saying we don't need marketers at all. Right? Like, yeah, I don't think we need to go there. Well, a lot of my philosophy on this is a little. Is more philosophical. I don't know it's going to be very concrete for marketers. I mean, I just had a whole conversation with this guy, Matt McCott, about this, of how important it is as a brand to define your profit. And that can be actual revenue, but it also often is something else. Like Patagonia has. Like, for us to be successful, it is revenue, and we support these certain things. So it's like, will that not matter anymore? If I'm having my AI agent go buy things for me, I think I still would prefer to know that it's purchasing from those types of brands. I think, like, the question that's coming at me right now that I think is most relevant for this episode and this conversation is around, like, okay, but is my job going to be here? Like, we kind of danced around it of, if you're an executor, then probably not. So what does that actually mean? Does that mean you need to upgrade your skills? Does that mean, like, you need to think differently? Is it even possible to have the thousand different marketing, you know, executor roles? Not all thousand of them are going to be able to take strategic roles. So there are going to be a bunch of people that don't have jobs. So we talked a lot about how AI agents are driving towards outcomes and they can execute on a bunch of stuff for us now. So, like, the actual role of a marketer, that is one of execution, likely will go away. So, Steph, is AI coming for marketers jobs?
- [21:45]
A
I mean, I hate to be the one to say it because no one else seems to be saying it, but yes, definitely, jobs will be lost. It is pretty apparent with the efficiencies that I'm seeing these AI agents bringing to a team, there's just no way that you cut out the execution roles and then you bring all of them up to a strategic level. Like, it just doesn't make sense. So, yeah, definitely will be. I mean, we're already seeing it with a lot of companies right now implementing just some of this and, you know, having big rounds of layoffs and so it's definitely only going to be happening more, especially for the execution layer of. Yeah. People who are in those roles.
- [22:26]
B
Which brings me to my next question. What do I do if I don't want to be the one that gets cut?
- [22:31]
A
Yeah. So, I mean, I would say one, it's never a better time to do your own thing. Like, you can start and launch things so quickly. And so if you're, you know, whatever marketing role you're in, you probably know the problems that your customers are dealing with, problems that you personally are dealing with. I mean, you can just look around and it is so easy to launch things now and build. So that's a very fun place to be in. But if you're like, that feels a little too far. I don't even still understand this. This was like a good 101 episode, but I still don't know, like, what to do. I would say just go and watch YouTube videos on this, Type in Open Claw, type in Cloud, Cowork, and just watch videos so you can understand the capabilities that these tools have. And then it'll start getting your creative juices. Going to be like, oh, now I. I didn't even have to install. I didn't have to get a Mac Mini and install Open Claw. I'm just watching videos and it's already expanding my mindset and the frameworks that I'm thinking when it comes to running a team or building a business or doing my marketing campaigns. And so you can just start with that. We will link up some of our favorite YouTube videos. We'll link up some of our own episodes where we've talked about this before, but I think that is step one, watch YouTube videos, and then step two is like, get in there, get your hands dirty. Just start playing around, messing it up. Like, just see what can be done on your own. Because that also is very expansive when it comes to, like, getting in there and getting your hands dirty.
- [23:56]
B
And I would add to that. Don't expect this to be like, oh, I spend an hour diving in and seeing what's happening. Like, spend a day, spend two days, break it, try again. Come up with another idea that you want to try. Really try to push it to its limits. Push your own creative limits. This isn't just a. Like, even though the tools themselves are very efficient, once they're made, the actual process of making them and understanding them is still a learning process. Just like you need to learn any other skill, you need to spend time to learn this one again. Don't expect it to be, oh, in a couple hours I'll understand this because you won't. And there will constantly be new things to learn.
- [24:34]
A
So we went through in this episode what an AI agent is, how it's different from automation, how the playbooks are changing in one month for companies everywhere, and then what it means for team structures, the work that you're doing, and how to really think about things in a very different way. So in the next episode we're going to go through the five steps you can take. So if you like this episode, hit subscribe on marketing trends so you know when our next episode drops where we're going to cover how AI agents can work for you, your team and your entire org.
- [25:02]
B
Awesome. Can't wait. Steph.
- [25:04]
A
Thanks Lacie. This is fun. You know how creative projects usually means juggling six different tools and dozens of different version histories of the project? Well, this is exactly why Light Tricks built ltx, the all in one creative suite for AI driven video. It handles every phase from conception all the way to final delivery. And it's all powered by LTX2, their latest and greatest model. It creates 4K synced audio and video with studio level detail. So no exporting, no waiting, just pure creation. Go try out LTX and see what it actually means to create at the speed of thought. Visit LTX Studio.