Mandy Hornaday: Hey, Nicole, welcome back to Growth Activated. I'm so excited to have you here today.
Nicole Leffer: Thank you so much for having me. I always love a conversation with you Mandy, so super excited for our conversation today.
Mandy Hornaday: Me too. And I have to say, you are my first repeat guest on Growth Activated. And I cannot believe it's been a year since we last spoke and totally appropriate for you to be the first one that comes on again.
Nicole Leffer: That's so nice. Yeah, no, a year in human time, like 25 years in AI time, I think. So.
Mandy Hornaday: Yes, yes, absolutely. And I, it's wild that when I sort of reflect back to our conversations and even just, we were talked a lot about workflow automation that really was so helpful for me. So I, before we even, I have so much I want to talk to you about today, but before we even get into some of my more tactical questions, I would love for you to just sort of level set with us from a CMO perspective. I know you work with so many different B2B marketing teams. What would you say an accurate current state of AI and marketing is? And I know that's a loaded question, but I just, I am so curious to see how you're thinking about it. Cause if you just look at LinkedIn, the perspective is wild and it feels a bit out of touch. So I'd love to get your perspective.
Nicole Leffer: Yeah, LinkedIn is definitely out of touch. it's wildly varied. think there is no average right now. We have the companies who are literally just like, I think we should start, like maybe we should get ChatGPT or Copilot, or they have Copilot and they are like, maybe we should like figure out how to use this like because their company has it. You know, like they literally are barely tiptoeing in there where a lot of people were two or three years ago. And then on the other extreme of that is like this insane like I know a marketing leader that was like my team's not adopting AI. I can't deal with this. We're going to fall behind. She fired all of them, rehired all people who are deep in AI and like they are building like really insane, truly autonomous, agentic workflows connected to all their data systems. Like they are there. They are all just agentic orchestrators. mean, like this is the extreme we're seeing. And so like, I would say far more, you would think from LinkedIn far more like her.
Mandy Hornaday: Wow.
Nicole Leffer: Right? Like that CMO that like everybody's just orchestrating their agents. Like that's the version. I would say that is like the 0.1% of the 0.1%. Right? Like that is a very small fraction of reality of what's going on. And to be clear, very, very few companies should be doing anything even remotely close to that. Like, right? Like I'm not talking about like she did layoffs of like to cut jobs, like for AI to do the job. She just did layoffs to replace the people with people who are taking AI seriously. So was one-to-one on the people. I'm not seeing layoffs. That's not the extreme, except for skills-related. Just switch out the skills. doing this really advanced autonomous agentic stuff, that's very few and far between. The thing that's far more common is somewhere in the middle of those two extremes. Maybe they've started. They have ChatGPT or Claude or Copilot or Gemini, that's their core tool. They're putting together some continuous workflows. Everybody is using it on a daily basis, but it's really more through the chat interface. That's still kind of the average. But I would say, honestly, I see just as many companies that are barely starting to dip their toe in. All they have access to is Copilot. That's all they are allowed to use. and they've never really used it. like.
Mandy Hornaday: Interesting.
Nicole Leffer: It is almost shocking how wide the variety across the board honestly is.
Mandy Hornaday: Yeah. Yeah. And Nicole, just level set with us, I guess, to sort of prime the rest of the conversation. What is your ICP when you think about like, which types of companies are you working with the most just so we have a frame of reference in terms of?
Nicole Leffer: Yeah, yeah, of course. So I am working with B2B companies anywhere from series A startups, although I would say not that many that are series A, like usually they're at least like B or C by the time they're working with me, all the way through like Fortune 50 household names. So usually there's at least 10 or 15 people on the marketing team. Sometimes there's 150 plus people on the marketing team. I'm not working with like solopreneurs, except for like they will take my course sometimes, but not like that I'm talking to all day every day. I'm not talking to like really tiny, but a lot of them are earlier stage, but literally all the way up through public traded, sometimes fortune 50, fortune 100, fortune 500 companies. And all, all B2B largely tech. So, you know, the tech ones are probably going to be the further ahead.
Mandy Hornaday: Wow.
Nicole Leffer: But I do have a few other industries that seem to kind of, I don't know why specifically they find me like the same type of industry. like, I don't know what I'm doing in my marketing that is specifically attracting kind of specific industries, but like other B2B enterprise level sales, that type of thing. Also besides not just in the SaaS space, but other B2B enterprise companies as well, or enterprise sales companies.
Mandy Hornaday: Yeah. Okay. Amazing. Amazing. So, okay. So I, I'm sort of stuck on this 1% of the 1% person that you shared with me because I think from my perspective, it's almost like the pressure the pressure is increasing because what we know and what we're hearing and learning is possible is so drastically different than where we're at today. And so it's like that the gap feels like it's growing between what we're sort of learning conceptually and what we're actually where we're actually at in practice. And so I'm so curious from your perspective, like, how are you when you think about bridging that gap? four CMOs today that are like, okay, maybe I'm one of those, not to call them average, because I heard you loud and clear that there's not really an average, but maybe they're a team, they're actively using AI in the prompting way. Their whole team is sort of adopting it, but they are nowhere near autonomous workflows, agents, sort of the future maybe of what we'd see. How are the CMOs you're working with thinking about bridging that gap. How are they approaching it? How are you coaching them to approach it? What does that look like?
Nicole Leffer: to go more with how I'm coaching them to approach it only because I want to be very careful. I don't accidentally violate any kind of NDAs or anything like that. So I'm not telling you what other people are like specifically. That one CMO I told you about one, she's not a client, but two, she also told me I could tell people her story. So there has been permission about that. But so one of the things that one of the big things I am really starting to talk to marketing leaders, I talking to you about is intentionality. What is your end goal here? What is it you're actually trying to do? Because just because certain things might be possible or theoretically the future, is that even the future you are trying to build towards? And before you run so, so insanely fast in a specific direction, decide if it's even the direction you and your company want to be running.
Nicole Leffer: And that's not to say that a lot of this stuff isn't things that you are going to adopt no matter what direction you're running, but what's your end goal? I really hope that people here are not building with an end goal of, want to eliminate all the humans. But I see the people on the cutting edge of this that I'm like, that is the inevitable result of the things you are building. And is that even your intention? Are you actually trying to cut humans out of this. And I think we are at this point where we need to be building our AI strategies and all of this stuff with an intentionality of what we are ultimately trying to accomplish. Because until you actually understand what you are ultimately trying to build, you can't really build it. might, you know, like think about if you just started trying to build a structure, like we're going back to the real world, forget AI for a second. Like if you're trying to build a structure, and you just start digging like a foundation and then you start like putting wood up. don't know what, you don't like, are you going to build a high rise? Are you going to build a like one bedroom shack? Like what are you building here? What do you want? And not every company has the exact same end goals, needs, perspective. Like what is it you want? Are you wanting something that's going to work really seamlessly across your entire org? Are you trying to build just for your marketing team? Are you trying to help your team be as efficient as possible? Are you trying to help your team keep the same size team you have and do the work of 10 or 20 people with three? What exactly are you wanting to get out of this technology? And if you don't slow down, stop, and ask yourself that, you could be building, one, completely the wrong things on completely the wrong foundation. And I think we're going to start seeing a lot of crumbling of marketing in general, because people have no idea what they're building, but they're building it anyway. And I'm like, what is your vision? What is your goal? What are you even trying to do? And this is the point in time where before you go too far, you need to think about that. And I think everybody is running so insanely fast that they're stopping to go, what am I even running towards? And do you really want to be the one that is running towards eliminating the humans? Like, you even, is that even the best interest of you, your team, your company, just cause you maybe could be building something that would eventually eliminate head count? Is that actually the best long-term strategy? I actually don't think it is. I think the best long-term strategy is to be building an alignment with keeping your head count where it is and massively increasing the amount of product, not just productivity, like on an individual level, but like what your company can do and expanding your dreams and your goals and all of that stuff. Those are the companies that are going to win. But I don't see a lot of people that are on the forefront actually building for that. They're just trying to replace what everybody currently does. And so I just stop, slow down, think before you actually run headfirst at everything right now.
Mandy Hornaday: Yeah. Yeah. Well, and we've, it's funny because to me, we've had, this is like an age old problem that we as marketing teams have. Like we are activity based. We like, we just run, we don't necessarily stop to think about the strategy of what we're doing and why. and so if anything, think AI amplifies a lot of that chaos that we've, if we've never at the root built in like strong operational processes and systems and, and strategy level thinking then yeah, absolutely.
Nicole Leffer: I I see so many people building these like wild agent automation, like autonomous workflow. And I'm like, what are you even actually doing with that? Like, it looks cool. It's shiny on LinkedIn. But like, you had 850 different dashboards you vibe coded. Are you really using all of those? Like, and what are you using them for? Like, you know, and it's like, it is cool, but like, what's the point? And if you don't know the point, you're just wasting.
Mandy Hornaday: Yes.
Mandy Hornaday: Oh, absolutely. I've been like, to me, again, another problem we've always had in marketing is like, I say we, I'm generalizing here, but a lot of teams have the problem of activation and adoption. And so it's like, I see the same things. know, there's so many cool things where you can do all this competitive research on a, on a weekly basis and it delivers it right to your inbox. And I'm sitting here thinking, but who's using that? Where is that influencing your other processes and is sales actually consuming all of that that you're giving them? Or is just more noise? And if anything, are we overwhelming ourselves in a lot of ways?
Nicole Leffer: Right. And I think that's the other piece of this is like, I don't know, maybe I'm silly and idealistic, but I feel like one of the benefits of this technology should be that it gives us some like space back and like thinking time and like maybe a little bit more balance in our lives. And like, maybe it's that we don't have to work some insane number of hours a week to be a marketing leader or to be a marketer. And instead what I'm seeing is the people that are like, You know, they're giving up their nights, they're giving up their weekends, they're giving up like their entire life to be building all of this AI stuff to what ends to automate yourself out of a job. Like you're given like, it's just, I, and I don't think that so many people, and it's fun and it's exciting. And I, I like doing it too. And I love it. And it makes a lot of sense if you are like a solopreneur or somebody, you know, that's working in a very, very small team to be doing that kind of thing. But if you have a team of 80 people, you don't have to give up your whole life to be building these AI things. It's just, what do you want out of this? Do you want more work, or do you want the work to give you balance? It's kind of interesting. I'm in a CMO community, a Slack community, and somebody said, what do you guys do while the agent is running? And I commented, was like, go for a walk. Use it. And they all are going, you know. And then a few people were like, no, go for a walk. That's like a really good way to spend your time or meditate or make coffee. But I'm like, like, you know, most people's mindset is like, you got to keep context switching constantly. And I'm like, no, if you don't have to do that hard work and you're going to come back and you're going to review it, go and enjoy your life during that time that that work is being done for you. And I think like, you know, we get to decide what we want this technology to be. And we've got to set the norms and the leaders are the ones who are going to set those norms. And maybe I'm just crazy that I think maybe if we wanted to, we could build a future where we just like actually can breathe a little bit more and enjoy our lives. And then I mean, it's probably like a little hypocritical because like I'm like live and breathe AI all the time. I've forced myself away from my computer to go for a walk. But like, I just think it's something we should all be thinking about with this.
Mandy Hornaday: Totally. And what's interesting to me is are you finding within the organization? So I love how you were, you're encouraging CMOs to think about the end game and build something that they want. And I think like, when I think about myself personally, I've considered going back in house and I've thought about like what, that would look like. And for me, like, it sounds like a dream to be able to run a flat lean team of really smart people who are powered by AI. Like I used to run a team of 25 and I actually didn't love that work. I love like the actual work and working with really smart people who collaborate and align really well. But I'm curious. So it's like, well, I have a vision of what I would want to build. Are you finding that CMOs are bumping up against like what CEOs and CFOs think that they need to be building? because you could be a CMO and say, I want to keep my whole team, but we're going to 5X or 10X our outcomes. our CFOs and CEOs on board with that.
Nicole Leffer: that obviously is going to depend on the company, right? To be honest, what I'm seeing more than anything is everybody just feels like they're supposed to embrace AI. And nobody even has a clue what they actually mean by that. CEOs, CFOs, CMOs, I don't care what CNO you are. There's just this like run as fast as you can at embracing AI. And yes, obviously, there is going to be a get on get everybody But I think that especially right now, we are at a place where you can absolutely sell. We're going to use this to whatever x our productivity, not our productivity, our revenue generation and like what we can accomplish and all of these things. And also sell that along with that, like, we can't just do the same with less people because you still need that expertise. Right? You still need that human curation. Like you still need the person who knows what good is because like you like you still you have to change the skills of that person, but you still need the person who understands it and maybe more than ever because
Mandy Hornaday: Yes.
Nicole Leffer: Like one of the things that I am definitely seeing in the gap now, there's the people who sound productive, right? Like there's the people all over LinkedIn who sound like they're doing really good things. And a lot of those, if you actually look at what they're creating, it is absolute garbage when you look at the output. So let's just be clear and honest about that. Like just because you saw it on LinkedIn doesn't mean it's actually good. But there is also this gap, wherever you fall on this adoption scale of good and bad output, right? Like you can have output that's not actually useful. You could just be scaling very quickly garbage. And I think like what you really, as CMO needs to be communicating to their leadership is like, need this team, the humans, so that what we are scaling is good and quality marketing and effective marketing and resonant marketing, not just marketing. for the sake of marketing. It's not just random acts of garbage going out. And these tools are absolutely brilliant and capable of doing incredible things, but they need somebody steering them and reviewing them. And so few companies, one, actually are putting together the guardrails and everything in place for not just what goes in, but how to guide the AI on what good actually looks like. And that's where the human is so important. is understanding and being able to control what is a good output versus a bad output, reviewing it, editing it. Now, if your team won't do that, then you need new people that will. Again, it's not like getting rid of people, period. But you need people. That's going to be the most valuable thing on your team is people that genuinely understand good versus bad of what the output is and can curate.
Mandy Hornaday: Yeah. Yep. Yep. Well, and it's interesting because one of the things I think potentially is a huge gap or that I don't actually see a lot of marketing leaders thinking about when we think about, I guess, automating a lot of our stuff is the human element of alignment and collaboration and bringing the rest of the org along with you. How are speed and alignment working together or are we just going fast and doing stuff in marketing and leaving everyone else behind. And so I also think like the human element of, I we're change makers, we're storytellers, like internally with an organization. How are we going to fill that gap that's so important to bring everybody along with you?
Nicole Leffer: think it so depends on the size of the org for this. think that how it's working internally, org size and internal everything makes such a dramatic difference. There is no one size fits all on this. What I do see is the bigger the org, the slower they are moving. And it's largely because they are trying to do a lot of that internal alignment. And I don't think that's You know, I would say like two years ago, was like, big companies are moving too slow and it's maybe going to like catch them really off guard. And now I'm like, actually, I think that some of these smaller companies are moving too fast and that's going to catch them off guard. So I kind of have like shifted perspective on that. But yeah, mean, like you do need to make sure that what you're building isn't in a silo, just marketing. And I think you do run the risk when you are going so insanely fast that you're not talking outside of your marketing org. You're not communicating with sales and product and CS and all of these different parts of your organization. You better all understand what everybody else is working with. marketing in many ways is ahead on AI, but not if you consider, especially tech companies, products, engineering, all of that. They're deeper than we are with AI, right? So if they're building in one way with AI and you're building something completely counter with your AI and you're building out these entire systems oblivious to what they're doing, you're going to have issues, right? And if sales isn't on board with what you're building, so yes, you absolutely can't just go. And this is one of the things I'm just seeing happen. And it's especially oftentimes these leaders who seem to be the farthest ahead. They get an idea in their head. They start building like insanely. They spend an entire weekend building an entire agentic system, blah, blah, and come back. And then it's like, what does this have to do with your whole org goals? Just because you had the idea, it doesn't mean it's a good one because it has to fit in the whole puzzle.
Mandy Hornaday: Mm. Right. And out of curiosity, one of the things you said about product and tech and how advanced they are, are you finding that, I guess, how insular are we from a marketing perspective building these things? Like, are you finding that within a lot of organizations, they're leaning on their tech teams to help build and bring them along? Or are you finding that like the orgs that are winning marketing is sort of owning the end-to-end build and creation of AI automations and workflows.
Nicole Leffer: so I'm not seeing a lot of companies that like their product, like, know, like their engineers building their product or helping build their marketing AI stuff. Like I've seen a couple of like those are doing, they're doing some cool stuff because they are using those resources in that way. but I am not seeing a whole lot of that. I'm seeing a lot more needing to lean on like the IT type tech support, you know, like that piece of it. The more advanced you get also like you're on a huge risk as a CMO, just like, unless you come from like an engineering background and you understand and cybersecurity and data security and all of this, and like you understand that tech. I think this is one of the things like people are getting really confused about with AI is that thinking that marketers are suddenly supposed to be engineer cybersecurity data security experts and like. This is really risky because they aren't right. Like very few marketers actually understand. then like, we're going to have some real catastrophic stuff start happening because people are building so fast without considering that. like the smart orgs are working very closely with their like tech security, IT, like all of that. They are not just like the, unless the marketer that they are have doing it really understands that stuff like deeply not just like they watched a YouTube video last week, but like actually has an education. They are not setting their teams up to use Claude Code and their command line terminal and like doing this like highly technical engineering stuff and building their own MCPs to connect all their databases. You would think based off of LinkedIn, that's what everybody is doing. But like, that's a bad idea, right? Like your team...
Mandy Hornaday: You would, yes.
Nicole Leffer: probably does not have the understanding and knowledge of how to securely build the data connections between your AI tools and your other tools in a way that you are not putting all of your internal data at serious risk. The vast, vast majority of marketers, that is not where your education is, nor is it what you do. And it's just craziness that we are expecting that like, marketing people should be doing that. people who are good, it's just a very different mindset way of thinking person. Like that's a different skillset. And everybody on your team, just have to like, maybe we need, I'm just thinking of this as we're talking. I've never said this to anybody before. I think maybe we got really spoiled with the fact that it feels like these AIs know everything about everything and we forgot that humans don't.
Mandy Hornaday: I love it.
Nicole Leffer: Right? Because we're used to, you can ask ChatGPT or Claude or whatever about like basically any topic and it's an expert for the most part. But humans are not. So you can't just assign humans to do the things that are like completely expert level things outside of their own experience. And I think that is a mistake I'm seeing a lot of marketing leaders making. Like you just, that's not what we're made to do.
Mandy Hornaday: Yeah. Well, it's been interesting. I did like my own little Claude Code hackathon weekend a couple of weekends ago. But I only do it for my own personal stuff because I don't do it for any clients. I just know that I'm okay with putting, I guess I'm okay with putting myself at risk and my own stuff. But I'm not, I just know that there are things I do not know. And it does worries me too. even like someone was asking like, do you, I guess even just having it like rewrite or update your live files and things like that. I, I don't know enough. know I don't know enough to know whether or not that's even safe or secure or so I like, haven't taken the step to allow it to do things like that. But
Nicole Leffer: saying like nobody should be using Claude Code, right? Like, and like when I'm saying Claude Code, I'm meaning like literally the tool called Claude Code that you use inside your computer terminal, like that is like very much a developer engineer tool. There is some cool stuff you can do with that, right? Like there's some really cool stuff. I see some wild things happening and there are definitely marketers who are fully ready to be using that and can be using that. And maybe I would even say should be using it. But I'm just saying it's not the norm. when I hear these marketers say, I'm going to get my whole team set up on this, like, is your whole team engineers? Do they all have this understanding of all of that stuff that comes along with it? Because I don't know. And then they're building, they're doing this autonomous stuff. And then they're also building out connections to their tools and giving it internet access. And they're just stacking together things. just because it's technically possible. like, because just like you said, you don't know what you don't know. I'm like, what are you, you know, what are you doing to make sure what guardrails do you have in place that you have built into your agent to make sure that it doesn't take your Salesforce database and like get conned on into giving it to some random person on the internet. And they just like, what do mean I need to like make sure that doesn't happen? You know, so, it's like, even if you the CMO knows that. Are you the VP of marketing does that like are you sure your team knows that because you're you know It's just one of these things where like not everybody needs to be doing the most advanced cutting-edge stuff possible Just because there is advanced cutting-edge stuff possible
Mandy Hornaday: Yeah. Yep. So what, when you think about, I guess where you would encourage CMOs to start today where they've got, let's say they have a team that is just using like the prompting function and they do want to get to the state of building autonomous workflows and things. where would you like, what are some practical tips and places to start or a framework to follow if you will, that they could be taking to make movement?
Nicole Leffer: So I would be spending time, actually, before you start building, putting on paper a few things. One, when you're looking at your workflows, one, what is the current workflow? But two, how does it change when you use AI? Because it may or may not. Where are the places of human bureaucracy that you can strip out? That's part one. You also need to be identifying how do you want a human to be involved in this? What are the points you want the human to do? Where are you comfortable with the AI making decisions, things like that? The other thing, and this is where I just don't see anywhere near enough people paying attention to it, is what does good look like? Do you know? Put it on paper. What is good? What is bad? What is the output of this? Because you can't start building a until you actually understand it. like the starting point to me is like one understanding what you're gonna be building and two really starting to look at what does good look like and getting that because you're gonna need that direction for the AI. You need to also understand like what context does the AI need to be able to do this? Like this workflow to create that good output that you have designed. And I would honestly like, maybe start with what is the good output? Like what does the good output look like? And then go, what does the AI need in order to be able to create that? What context, what information, what guidance from a human, like all of those pieces, what all is necessary to do that? Those are the starting points because until you can curate all of those things together, you can't build effectively. So whether that is even in a chat, in ChatGPT or in an autonomous agent. That is the thing you need to get. And I think that so many are skipping kind of how do you evaluate? Like how do you make sure it's good? And ultimately like that's what you need to be building around. The other thing I would like strongly encourage everyone to do if they have not started building, I think you have one serious advantage over everybody who is already super far along involved. You're very down that path. You have the opportunity.
Mandy Hornaday: Yeah.
Nicole Leffer: to start building tool agnostically. So a lot of people who were the first ones started moving so fast that they've built their structures and they've built everything out in this specific ecosystem. So they've built it all for a specific AI agent, a specific tool, because they just wanted to get in and they started doing that. I think the ones who are going to be at the real advantage are going to have super, super portable. workflow. you're putting together, it's more the structure you're building for the structure, the you're building for the context, the structure, the curation, the human involvement, the what does good look like and all of that. And like actually having your team builds that we see, we have seen for the past few years, we don't know who the front runner is going to be. Right? I know that there is like a huge almost
Mandy Hornaday: Mm-hmm.
Nicole Leffer: cult following around if one specific company right this minute. But you know what? A year ago that cult following was around a different company. And a year from now it might be around another one. So like, we don't actually know. We don't know what in a week from now, we don't know by the time we get off this call what somebody is going to launch or come out with that is going to upend what we thought we knew is happening so fast. So I would not be building for any specific tool. would not be attached to any specific tool. I would be looking at what are the components that go into this and how do I build this in a way that I am not dependent on any single AI. What happens if the AI you were building for triples their price compared to everybody else? What happens when the AI you were building for becomes an option that is going to cost you more than humans would? We don't know. We just don't know. And so you don't want to put yourself in a situation where you are so pigeonholed into any specific tool. And I think a lot of people are under the impression it's certain things right now. Because they're hearing everybody talk about a specific tool, they think that's the only tool that can do it. It's not actually true. For the most part, it's like right now I'm seeing and hearing Claude Code everywhere. And especially in marketing leaders, it's like we have to have Claude Code. And don't get me wrong, Claude Code is an incredible tool. So I'm not trying to be anything negative on Claude Code at all. It is very cool, very, very powerful. Codex can also do that, right? Codex in the OpenAI system can also do that. So I'm trying to think about how do I build this in a way that I could use it on Claude Code or Codex or something else that will come along. That way, I'm not stuck in any specific ecosystem. And I'm really in a situation that I'm going to be better for the long term. And that's where I would be advising, because I am starting to see companies jumping tool to tool. I would really be thinking of it in that lens.
Mandy Hornaday: And what is the practical, is it just have core documentation that's housed in your platform agnostic storage, like cloud storage? Is that how we do it, that then you could take that documentation and dump it into any of the different platforms? Or is there something else that could kind of future-proof you or make sure that you remain platform agnostic?
Nicole Leffer: think a lot of it is, yeah, it's where you're storing the documentation, how you understand what is happening, how you are creating AI systems that document their memory in a way that it is not specific to that tool, that is not living only in something that that tool's ecosystem. Now, when these tools are working on our own computers and in our own files and stuff, that actually gets easier, right? Like that actually, it's just about setting stuff up in a way. Like, I would be thinking about when you could do things six different ways. Because a lot of this stuff you could do multiple different ways, right? How do you do it in the way that is portable versus the way that is tool specific? How, what, you know, and I will say, like, Anthropic has been really incredible as they've built their product that they do open source a lot of their major innovation. So they make it an open standard, which is fantastic, right? So like, When you build an MCP, you're not building, I'm again, let your security tech people build your MCP, just don't do it yourself. But like, unless you are really know your stuff, but like win an MCP, which is like basically an API connection between an AI and your different tools or systems. An MCP was something Anthropic came out with, but they made it an open standard. So it has become the standard that you can use with any AI to connect with a tool. Right? So like if you're building your connections as an MCP, can portable, that's portable to different tools. It doesn't have to only be used in Claude or only be used in OpenAI. You could use that in multiple tools.
Mandy Hornaday: Great.
Nicole Leffer: When you build out skills, when Anthropic came out with skills, it started just in Anthropic. It became very, very popular. And so they made it an open standard. Now you can use skills in Codex. You can use it in your business ChatGPT. Like you can use this across the board. So if you're building the stuff in a way that it is in a portable format, that's the biggest thing is like the formatting of how you're putting that documentation. It's not so much about is it stored on the cloud? Is it here? Is it there? You're having your tool write a memory in a way that it is going to be a portable file. So you could just plug it into any other tool. just, I mean, what happens if like something happens and like you're, you would actually built your entire marketing program into something that you can't really function without your AI. Cause I'm seeing a lot of people building that. What happens the day that that tool is down? Do you just like nobody's capable of working?
Mandy Hornaday: Yeah.
Mandy Hornaday: It happened. I know. Yeah.
Nicole Leffer: Or do you just go, let me go into our backup tool. Right. So, or I mean, hey, I actually am in favor with like, hey, if the AI goes down, just take the day off. It's like a snow day.
Mandy Hornaday: Okay. So I am curious. That's so helpful. I need to, I'll put some thought into that because I hadn't really thought about that before, but I am curious and I know this, your answer will probably change or maybe changes by the day or the week or what, based on what gets launched. But where do you stand in terms of, I guess it seems like Claude Code and probably Codex, I don't have as much experience with that. But what I'm trying to wrap my head around is like when to build versus buy and what's the value of the other, the AI tools that are coming out that seem like they do something that you could do in Claude Code, but probably make it a lot easier or whatever. But I'm curious, like, are you thinking that marketing teams can really just be leveraging and get the majority of the value out of the core LLMs that they're using or do you think there is a time and a place for these tools and companies that are popping up?
Nicole Leffer: I think it very much depends on the company and the team and your skill sets and your risk tolerance and all kinds of other things. I think that it really depends what specific use case you're talking about. The vast majority of things you can absolutely do with the core tools. The vast majority of things that are being built out. You can do with OpenAI, can do with Anthropic. Everybody's building it on top of that. It's just the wrapper, right? It's the harness it's in. So I would say like most things, your tools are gonna be like a really good option. That said, there are like certain things like I would say with like video, image. you know, workflows that maybe you do want to go beyond those. it's very few. Some data related, you know, I understand the argument for a Clay or something like that. It makes sense. Not like I'm saying go buy Clay, but I'm saying that is a little bit different, right? But the vast majority of just general purpose workflow, I don't see the point. I don't see the point of buy versus build. for everyday stuff because when you, you're just not going to get the same. And like the biggest thing out of that, when a company is building, like, let's just say, I'm just going to invent a fictional like dime a dozen tool, but like, let's say it's a AI tool that makes blog posts, right? Like obviously we're way more advanced than that, right? But like, let's just say it's a purpose built tool for building blog posts. When that company goes and does that and they're They're building to do it for lots of different people. So they are going to pick the cheapest way to do it, not necessarily the best way to do it. They're going to pick the cheapest model that they can to get to a quality their customers will tolerate. That's essentially, like, know, that's just going to be the economics of it, which means they're very rarely going to be using, the best option model, where if you're doing it inside your LLM, in a subscription-based tool like a ChatGPT or a Claude or whatever, you can pick the best model for that use case. It's going to give you the quality that you want to. And you can guide and control everything about it. You're not going to the unit economics of scale. You're going to the unit economics of your one blog post. They're going to the unit economics of potentially thousands and thousands and thousands of blog posts. Those are two very, very different things that you're both calculating for.
Mandy Hornaday: Hmm.
Nicole Leffer: And I think that's one of my own biggest arguments of why I just don't think it makes sense to go with the purpose-built tools because their economics are motivated by keeping their costs down, which means not using class models in general, not saying nobody does, nobody come after me for what they, but like just as a general rule of thumb for most of these companies. so unless it's genuinely something that is like,
Mandy Hornaday: Right.
Mandy Hornaday: Fascinating.
Nicole Leffer: very close to impossible to do. Now, I will say if you're building out like full blown and agentic workflows and know, Claude Code and stuff like that could get expensive real fast. Codex like that could get expensive real fast depending on what people are doing. So not saying there's never a reason that you would ever pick another tool, but like I would really look at the whole picture of it and kind of evaluate that piece as well.
Mandy Hornaday: Yeah, yeah, absolutely. Okay, good to know. Well, this has been such a fun conversation. I feel like I could just talk to you for hours. Is there anything, Nicole, that you would leave with us? Maybe something that you would tell every CMO to do immediately this week or just jump in or what would you, what's your last tip of advice that you'd leave with us?
Nicole Leffer: OK, breathe. That's it. Just take a deep breath. Take care of yourself in all of this. Take care of your nervous system. Calm down. Relax. This is moving really, really fast. It's going to be moving really, really fast for a while. So breathe. That would be the first thing. And then if you want something more practical with AI, sit down and figure out what is it you're actually trying to build towards. so that you can focus on what's actually most important towards that vision. You cannot build something you don't know what you are building. take a take stop, take the time to go, what am I ultimately trying to do? And then you can start focusing in on what is the most important steps that I should be taking to make that happen instead of getting caught up in the noise of what everybody else is doing just because they can.
Mandy Hornaday: Awesome. Awesome. I love it. Well, thank you so much, Nicole. Appreciate the time. I know people can find you on LinkedIn. I love all of your insights there. It sounds like you also have an online course people could take.
Nicole Leffer: Yes, I have a foundations of generative AI for B2B marketing course that is extremely popular. You can take it as an individual, also have a lot of teams that take it as a group. And it's really giving you those foundational skills to be able to use tools like ChatGPT.
Mandy Hornaday: And then of course they can also reach out to you to come into their org as well as a consultant, right?
Nicole Leffer: Yes, absolutely. I do trainings in-house for companies. work with with CMOs on their adoption strategy, consulting, advisory, all that kind of stuff. So yeah, and I might also have some more courses coming out soon as well.
Mandy Hornaday: okay. Can't wait. Awesome. Well, thanks, Nicole. Appreciate the time.
Nicole Leffer: You're welcome. Thank you so much and have a really great rest of your week. Bye.
Mandy Hornaday: You too.
