When AI Rewrites the CMO Playbook

Penguin-Hat Drew: The new CMO playbook is still being written, and unfortunately, somebody stole the answer key.

Doing proprietary research that is of interest to the media and your customers has a lot of value. If you wanna be cited, you've gotta create something worth citing.

Use AI aggressively to remove sludge, and think much harder before outsourcing what makes your brand distinctive.

Maybe a truism here: The most expensive AI tool may be the free one your team built themselves.

Drew: Hello, Renegade Marketers! It's time for a very special Drew on Drew episode. I mean, what could be more special, in which I get to interview myself. Why? Because I've had a lot of fascinating conversations with CMOs lately. I know I have them all the time. But I wanted to connect a few dots, and because this is episode 540, and as you may remember, if you're a longtime listener, every 10th episode I do a Drew on Drew or something else special. Here's the headline: The CMO playbook is being rewritten in real time, and that may not be news, and this is definitely not gonna be news, but AI is changing how we build marketing organizations. We talk a lot about org design, how we create content, for better or for worse, how buyers find us, obviously LLMs, how we think about PR, think about it more, and even how we decide what technology to build versus buy, and this is a relatively new conversation that we're having, uh, in, in Huddleland. At the same time, some decidedly old-school things, like creativity, yay, relationships, yay, original research, and trust seem... and I, yay is for those two, seem to be getting more valuable, not less. So we're kind of racing ahead, and we're sort of looking backwards. It's an amazing time to be a marketer, and I'm gonna try to make sense of what I'm hearing from CMOs, what we're experimenting with at CMO Huddles, what seems to be working, and where I still have a heck of a lot of questions. So let's let the interrogation begin. Okay, Drew Smartypants, the CMO playbook seems to be getting rewritten in real time. So what are the biggest changes you're seeing right now?

Penguin-Hat Drew: Nice. Nothing like starting with a softball question. And this is not simply an AI story. It's, like, really, we're fundamentally reconsidering how marketing gets done, how many people do you need, and, and so forth. So across the leader huddles, and those are for the members of our leader program, and the one-on-one conversations, which I've been having a lot of, CMOs are wrestling with org design, AI adoption, both the training aspect, what tools to buy, where to use it, measurement, and that's on a broad scale. How do we measure what we're doing with AI? Content, yeah, how quickly can we get a lot of slop done? Search, um, both. SEO still seems to matter, but AEO, or if you wanna call it GEO, those things obviously are rising. Brand, talent, and technology, all of this simultaneously. And so you can talk about we're, we're— at the Super Huddle, which is coming up October 22nd, 23rd, we're gonna be, uh, unveiling the findings of the FOMO, that's the Future of Marketing Org task force. I wish I could say that we settled on one particular org design. It's not gonna work that way. It's like everything else in marketing. It's, uh, bespoke. Uh, you have to find the right one. But we do have some frameworks for you to consider. All right, so the shift that I'm watching the most carefully is from AI experimentation to AI opera- opera— I can't even say the word. Operationalization. Okay, so we're gonna operationalize AI.

Drew: First of all, what do you mean by opera- re- the word that you cannot say? The word that shall be said, but you can't say it. What do you mean by that?

Penguin-Hat Drew: Here's the deal. We used, uh— A year ago, and we talked a lot about this on this show, uh, we used to say, "Just go out and play, experiment, drive." These are— And this was a reasonably decent thing. But when we fielded our, the AI maturity study, um, which we partnered on, uh, with, um, the folks at Benchmarkit, we discovered that on a scale of zero to 100, the median was 53, meaning half the people were sort of way down here, and half, um, were 53 or more. And very few, I think less than 8%, were actually in, 70 and above. Now, we graded this pretty harshly. Our scoring was harsh o-on purpose because we think this is such early days. But the distinction between those that are experimenting and those who are really turning it into an operational function is important to understand and to be thinking about. So just, we know, for example, using ChatGPT isn't transformation. Yeah, it does so many cool things, and I love it, and I spend a lot of time with it, and it saves me hours. That's not the same as what we're talking about. We're talking about processes and governance and not having 18,000 agents, uh, running amok. It's about, gosh, really some systems thinking across a wide range of the organization. And just to sort of bring some clarity to this, if you think about the early days of marketing technology and MarTech, um, there was a lot of experimentation, adoption, and security, uh, breaks. And I think the same thing has happened or is happening with AI. Um, these agents that you're building in-house are vulnerable. And so the oper-operational, I cannot say that word. That's gonna be a theme in this thing. You're gonna keep asking me about it, I know it. Anyway, these are the kinds of questions that we'll be wrestling with at the Super Huddle. Again, nobody has a finished playbook, and I think that's what makes this part of the fun. So if I were to summarize this question, the thing I would say: The new CMO playbook is still being written, and unfortunately, somebody stole the answer key.

Drew: Let's move on to the next question. So we've all been experimenting with AI for a couple of years. What separates experimentation from actual AI maturity? And I know you covered that, but let's get into detail on that.

Penguin-Hat Drew: So everybody's using— I think it's, oh, 98% of companies are using AI in one way or another. But most of that is at the, productivity enhancement phase, which is just getting started. They are not doing what, um, Paige O'Neill did at Culture Amp, where they took work and they divided it, found there were 5,000 different, individual tasks, and figured out how they could streamline and create processes to significantly reduce both the amount of time and the amount of people it took to get that work done. We know that there's lots of experimentation going on, but there are far fewer repeatable workflows that are actually tied to measurable business outcomes. And, and I clearly see this, and I mentioned this, I think, before, that we're at this point where it was ready, fire, aim, and that aiming part is so problematic. Again, if you step back and say, if we have a strategy and we say our strategy is to have our customers love us, to be the most beloved company in the category that we compete in, how you use AI will be very, very different than if you say, "We want to grow the company by 10%," which is, again, not a strategy. It's an objective. Knock yourself out. You want to grow the category, you want to grow your company by whatever percent. Yeah. You should have a growth goal. But that's not the same as a strategy, and we need to have a strategy and a clear strategize. So it's not, yeah, we're doing AI. Everybody knows that. But we need some kind of maturity progression. And the one I've been thinking about, it's like we went from chat to creating systems to exploring, and a lot of folks are building— Heck, I built eight agents, which of course used up all the tokens of all my staff in a day. So we had to turn them off because I don't— We— I still have one going, but I, I don't know how to build agents efficiently. And I didn't know that when I turned all these on, they were gonna burn through this. So this is part of the learning process. So before you build agents, make sure that they are in service of a, a business objective, uh, and a part of your strategy to be more competitive. And it's fascinating, with the AI maturity benchmark study— Uh, by the way, you can go to cmohuddles.com right now and go to our AI maturity calculator, and you can see how you did relative to the benchmark, and it also provides you some nifty recommendations, um, there. And I, I really leaned into Ray Reich, uh, very heavily, uh, at Benchmarkit to help create both the benchmark study and, um, the calculator that you'll find on cmohuddles.com. Now, I think there's... a useful distinction here between AI activity and AI maturity. You could be doing everything with AI, but that doesn't necessarily mean that makes you an AI-mature company. It probably does mean that you're burning through a heck of a lot of conversations. One other last thing on this point is on, uh, cmohuddles.com, you'll find an extensive blog post on a conversation that I had, in the same afternoon, one with Amanda Kahlow, who's the founder of 1mind and, of course, uh, a founding sponsor of the CMO Super Huddle, and my friend Michelle Killebrew, who, who runs a really interesting consulting company. And both pushed on the same management challenge, that adoption can move faster than organizational discipline. So just make sure you understand that. Adoption, we can get everybody using it, but the organizational discipline to actually make this work effectively and safely and cost-effectively and fast is beyond what most companies have set up. So one of the warnings that came out of those conversations that I, I love is: AI can make the wrong business faster. If you're doing— If you don't, if you're not pointing in the right direction, you'll get there faster. Um, that's one thing that AI can guarantee.

Drew: Well, that was a long-winded way of answering the question, but okay, we'll, we'll give you a, you know, five out of 10, kind of like your, uh, your, your model. Here's a question that came up and you, you wrote about it, which is, could AI actually make B2B marketing more creative?

Penguin-Hat Drew: Well, first I want to say I sure as heck hope so. I, I'm the guy who used to talk about B2B as boring to business. I've always felt that B2B marketing could just do a little bit and really stand out because everybody else is so safe and talking about their speeds and feeds and, you know, there's so few real, true brands that carry it all the way through. But what was interesting is, you know, when, when some people think about AI and creativity, they think about, well, AI can write copy, and AI can create images, and AI can create video. But in my conversation with Katrina Wang, the CMO of New Relic, she talked about it, it differently. AI is automating so many of the menial tasks, she and many members of her team have more time to be creative. And I love that part of that and that interpretation, and I do believe that would be— really create a creative rena- renaissance. I'm not sure everybody is there, or I certainly, I wish everybody was there, but Katrina talked about that with such enthusiasm that I said, "Okay, I hope you're right." And what— the point that Katrina made was, you know, at least she and a lot of marketers didn't get into marketing because they be- they dreamed of becoming spreadsheet warriors or systems operators or attribution lawyers. They, they got into it because they wanted to be creative. And, you know, they, they could've gotten into accounting or sales, but no, they, uh, that was part of the promise of being in marketing. And one could argue that the rise of MarTech kind of correlated with, uh, perhaps a decline in, in B2B creativity because you could systematize things. But what Katrina talked about, and New Relic has roughly 85 agents supporting competitive intelligence with hum- humans supervising the work, but they've been able to reconcile disparate data sources and then think about human capacity consumed by meetings where everybody argues over whose numbers are right. And so if you get the data right and we agree on the data and, and the AI helps us, then we can argue about the creative direction of where the company is going and the language we use and the positioning that we have. So Katrina's biggest point in all of this was that removing this sludge, these unproductive things, could give marketers more time to create. Well, I certainly hope she is right. She offered evidence of New Relic's own brand work, uh, with the "Welcome to the Superhuman Era." It wasn't super AI, superhuman. Um, and the idea there is AI isn't necessarily the human, it's that humans can become more capable because of AI. I appreciated that, and again, I'm rooting for her. Before I sort of wrap up this chapter, I do want to talk about positioning. Um, a lot of companies are sort of slapping on AI, and it's a, uh, uh, AI washing, if you will, of their positioning. Our friends at Firebrick, also a founding sponsor of the CMO Super Huddle, they have a process and a way, and he's, uh, uh, Bob Wright will be talking about how you can use your AI. Bob talks about naming your AI and using it as part of your overall positioning strategy. You can't just slap AI on the name. It's got to be integrated into the positioning in a meaningful way. So, anyway, you have to hear Bob say it, and I've, I think I've got him on a Tuesday Takeaway or a Wednesday Wisdom talking about it. But, uh, the other thing I wanted to connect with, I did have a conversation with Aviv, the CRO of DataRails, and you can find that post on, on cmohuddles.com. But what was so interesting to me, so Aviv had been a CMO, he became the CRO, and marketing s- uh, reported into him, as did sales. But he talked about, even though he thinks like a revenue architect, he gets to decide where he spends his money, and he deliberately protects brand spending to keep them from becoming this transactional company. And his TikTok spending has, uh, been a really great reminder that not every creative bet arrives with a, a tidy attribution about— He wrote, shared a story, and I think it's in the blog post, where a salesperson called and said, "Have you heard of DataRails?" And the, and the guy said, "Oh yeah, I've heard of you. I see your TikTok ads all the time." Which of course, the salesperson stopped and reported back. But it's that kind of anecdata that makes you believe that maybe this AI era, and by the way, Aviv is a big adopter of AI, could help us get to a point where our creative, um, work is actually, not only the channels that we use, but the, the kind of work that we put on those channels.

Drew: All right, Drew, another long-winded answer, um, but I think you made some, uh, interesting points, barely. So what happens when everybody uses AI to create content?

Penguin-Hat Drew: I think there's a dark side here of democratized creation. Look, everybody had a typewriter, everybody had a computer, everybody had a pen, but not everybody was a writer. And obviously, I strongly believe the same thing applies. Part of it just comes from if you look at something and are— do you have the ability to know whether that's good or bad for your target? And let's face it, if you have no experience and you put a draft in, uh, and ask ChatGPT to write something, you're gonna go, "This is pretty darn good." I have found, uh, of late, just with my own content, that because it's trained on the hundreds and hundreds of articles that I have written, it knows my voice. And because we go through a comprehensive, crit, uh, process where we go back and forth on what this is gonna be about, what the key points are, what the outline is, who the target is for this particular piece, what the point of view is, what makes us special. And by the time we get to this, and I've input, input, uh, input it back and forth, it, one, it's taken me longer to write it than, um, which, uh, than it will take you to read it. That's one good rule of thumb. And two, I'm hard-pressed to say I could have written better than this content. And I know that will surprise some writers, and maybe I'm not as good a writer as you are. But I'll have people read these posts, and they'll say, "Oh my God, that's great. You're a good writer. Don't change a word." And, and so I think you can create great content with AI, but I think you have to be a good writer to begin with to get there. So one, um, thing that I'm struggling with right now is that there's a lot of talk that all AI content is bad, and I, obviously, I disagree. But I think there's a budding concern that if you put all of the content that you created AI touches, and there is a pushback from the larger marketplace, uh, like it's gon- might happen in GDPR, uh, the equivalent of GDPR in Europe, and they now say, "Well, how much of this was AI?"... And that's a big question. We know that ChatGPT, um, is thinking about watermarking, if not already watermarking. We know for certain that Anthropic is doing watermarks. We know that, uh, Google Docs has the ability now to, uh, that has an AI detector on it. So the question becomes, do you have to disclose that? How do you disclose it? What is considered, um, AI versus human? I think we're in a really squishy territory, but I flag this because I would say to CMOs, "Don't fire your copywriters or your— certainly, don't get rid of all your content people," because it is unclear that if you are, uh... It was— It's clear to me that if you're wholly dependent on junior people to create your copy, you're gonna have junior copy, no matter how good you've trained, uh, because those are the people judging it. So, other part of this is if you don't ha- aren't working with original things, like I always have transcripts of CMOs and conversations that I'm working with to make this content, uh, unique, you are gonna be producing slop. So there is some brand risk in creating, being heavily reliant on AI. That's a main point that I wanted to make. So the last thing I just want to wrap up on is: Use AI aggressively to remove sludge, and think much harder before outsourcing what makes your brand distinctive. So all right, those are the two things. And this final thing, I call it a Drew-ism, is that cheap content can get very expensive if it makes your brand invisible.

Drew: Wow, this is going on and on and on, Drew. Come on, let's get to something exciting. Like, is AI actually making PR more important?

Penguin-Hat Drew: Yes. Categorically yes. This is one of the most counterintuitive conclusions from recent leader huddles. AI can write the press release, it can build the media lease, list, and it can personalize the pitches. However, automating PR activity isn't the same as generating PR results. So what it kept coming down to was whether the internal PR team or the PR partners, uh, the agencies that these CMOs were using, it came down to having strong relationships with the journalists in your industry or trades, and that was paramount value. I can't emphasize it enough. Now, press releases going out on the wire do have value, and there's, uh, several CMOs in our community who are cranking out, uh, more p- uh, press releases than they ever have before. Hopefully, some of those are newsworthy because they, they do have value out there. You know, uh, one, it's content that you can put on your website. Two, hopefully it has news. We also know that earned authority is becoming scarcer, that it's harder and harder to get high-quality coverage. What we need to be thinking about, you know, PR has always been about you either have news or make news, and you rarely have as much news as you would like. That might be a major product announcement like your iPhone, um, 18, for example, which I currently covet because the camera looks pretty cool. Anyway, back to the subject of PR. You really need to do things that are newsworthy, and that could get you into the point where you're doing original research and building, uh, studies that you can do year after year after year. It's funny, it worked 10 years ago, it worked 20 years ago, and it's working now, and it's working great for a number of brands in our community. The, the point here is if you've been underestimating, uh, PR, you are probably, your share of voice is lower than your competitors who have outspent you. Um, share of voice matters 'cause that will show up, uh, ultimately in citations in the LLM. So, that's going to bring us to a very, this very old-school PR tactic, which is, uh, doing research.

Drew: So, what did the CMO tenure study that you did, um, with, uh, Findem, um, teach you about creating content that actually earns attention?

Penguin-Hat Drew: I'm so glad you asked that, Drew, because the CMO tenure study is something that, uh, we, uh, partnered with an organization called Findem. Findem has this massive database of, uh, everybody who's ever worked. Uh, and I don't know exactly how far it goes back, but it goes pretty far back. So— And one of the things that I really wanted to get a sense of was what's tenure look like on a very broad level? You know, and what was interesting there is the tenure studies that, uh, Spencer Stuart have, have done and Forrester have done are great, but they really only look at, say, Fortune 500 companies and public companies where that data is available. And they only look at it one way. So we wanted to look at it both on a historical level as well as current CMOs and how long they had been in seat. So on a historical level, we learned that, yes, from 2010 to, 2022, and I'll explain why, tenure pretty much dropped just steadily from about four years to, um, I think it was 2.6 years. And that's the historical view. But we had to stop in 2022 because some of those CMOs are still in role today. So we did an— another thing. We looked at 13,000 CMOs. This is amazing. 13,000 CMOs in the US at companies with 100 or more employees. Now, at least, almost half of those were B2B. And this is a new thing. This is the largest study of CMO tenure ever done, and it's proprietary research, and we're excited, uh, that, uh, a bunch of, uh, media companies picked it up, um, and we're actually seeing, uh, some traction as a result of the coverage of this study. So, what are the lessons learned? I mean, again, proprietary research. Do your own study. Do a study that is of interest to, um, your target, to the media that covers your target, and do something that you could perhaps do, again and again. And so what'll be nice about this study, we— there's a lot of things that we did in this first round that were great, and we've established our benchmarks, f-for it, and now we can do it again in a year and look at it and say, "Well, did it change? What are some of the other trends?" Um, by the way, there was very little difference between B2B and B2C CMO tenure. Probably one of the biggest findings was the differential between public companies at 4.0 and VC-backed, which I think was 2.6. And then there was a lot of difference by vertical industry and by size. Typically, the larger the company, uh, the longer the CMO, uh, was in position. So, lessons learned. Yes, doing proprietary research that is of interest to, um, both the, the media and your customers has a lot of value. It doesn't— it's not opinion. Create some evidence to help tell a story. So, if you wanna be cited, you've gotta create something worth citing. Uh, I- it sort of reminds me, uh, of the Ben Franklin quote, "I-if you would be loved, love and be lovable. If you want to be remembered, either write something worth reading or do something worth writing about." So, okay. Thanks, Ben. Man, yeah, I haven't made it into one of these for a while.

Drew: Next. All right, now let's see. Separate from the PR success, what should CMOs actually learn from the CM- the tenure study?

Penguin-Hat Drew: Well, I've covered some of that. So the main thing— Yeah, the median tenure for CMOs in position right now is 36 months. I talked about the difference between public and VC-backed, and PE are right in the middle at 3.1 years. So you know, as a CMO, if you take a job at a VC-backed company, it's not impossible, but you are on a very, very short lease. You know that, particularly the smaller the company, the more likely it will be to pivot, the more change there will be made. Uh, and, and the result is, um, a lot of instability. And so there are some things that you could do even with VC-backed companies. There's a lot of questions that you need to know about the CEO, about the category, about the resources that... you will get, what success actually looks like in getting, um, these things, before you. So you're interviewing for, uh, them to make sure that they are a good place for you. Look, you know, the, the chances are if you go to a VC-backed, you might be able to stretch that number to three years, or more importantly, you'll have the choice. And this is the key thing, because sometimes CMOs like to work for three years and then move on, or, or four, whatever it is. So tenure is not a goal. What is control of your career is the goal. And so, you know, the more you know up front, um, the more you understand the odds that you'll be there in three years, um, versus 18 months. It's hard to beat the averages if you don't know what you're averaging against. So lots of research up front. And we can— If you want to know a, a longer, more detailed list on what to do in that situation, if you have a VC-backed company you're about to go to work for, you can certainly, uh, ping me on LinkedIn for that information.

Drew: Okay, speaking of being cited, what have we actually learned from CMO Huddles' AEO experiments?

Penguin-Hat Drew: Okay, uh, this is a, a fun one. So first of all, we had almost zero, um, AEO, uh, six months ago, and, and part of the problem was, uh, we host, uh, our website, primary website is on a platform called Wild Apricot. Unfortunately, while it's a great community management platform, it's not very good for blogging. In fact, it's really bad for that. Uh, it's something I hope someday they will fix. So we made the strategic decision to add Webflow on our front end, and that brought with it a lot of opportunity for us to add schema and redesign our approach. So I'm gonna say it's about two months now, maybe it's a little bit more than that, but we have repurposed a lot of the content that we had before, updated it, added schema, added Q&As, added summaries 55 words or less at the beginning to make it easier. We've added our .txt file, which, uh, no, I think that's wrong, .llms file. Anyway, one of those. And we've made— We went from a one star to a three star in about two months on a scale of five. W- we've really leaned on Webflow with, um, their deep-dive, um, analysis, uh, that they do, which is at webflow.com/aeoma. Uh, highly recommend it. That was our benchmark. We've made a lot of progress. Obviously, we now have 150 posts. We used to have zero. We're seeing our citations increase dramatically. Our site tra- traffic from LLMs has increased dramatically, and we still have a long way to go. It turns out there are a lot of pages that haven't technically been updated. We're working on that. You can't do all of this stuff at once, and maybe you could. We couldn't. We're a small organization. But what we are finding and what we're really proud about is the content that we're putting up on this website is very helpful for CMOs. It's all original content as far as, um, based on interviews that I have done or others have done or the research that we're doing. We've also had a lot of help from our friends at Pepper Content. Pepper is also a sponsor of the CMO Super Huddle. They have a, a really cool tool called Atlas that allows you to assess and figure out the topic areas that, uh, where you're ranking, where you're not, and, and what kind of content you should create. And by the way, we see that in the work that Pepper is doing, um, as pure upside from where we are now. We're still, um, behind, um, one of the, uh, of our competitors, um, but we have made a lot of progress. And so, I think the, the larger lesson for us is AEO shouldn't be treated as a mysterious black art. There are things that you can do on a technical basis right away that will make a huge difference. After that, then, then it does come down to your, your content strategy and, and obviously your PR strategy and your Reddit strategy and all those other good things.

Drew: Let's, we're, uh, maybe we're, it's wrapping this up quickly, but everybody seems to be building AI tools, and I'm just wondering if you can comment on when should a CMO build versus buy?

Penguin-Hat Drew: So I want to acknowledge, I love building these things. I was able to build a tool that went onto our website. I've built a number of apps, as I mentioned earlier in the show. I think it's really cool that anybody, including me, can actually do this. I also recognize the danger of it. You can build a prototype instantly, and then you build an illusion that you've eliminated the need for a particular software. Building isn't owning. I wanna make sure that you recognize that when you build a tool that is of se- serious importance to the business, let's say a, a CRM or a, um, a, a website, a CMS, or, um, gee, a marketing automation tool, you are now in that business. You have to integrate it, you have to secure it, you have to maintain it, you have to update it. And probably most scary of all, you are dependent on the, um, LLM that you built it upon. And, I mean, you don't know if you build it on whatever the current one, the, the next iteration of it or four iterations down will still support it. And you don't know if the person who helped you build it, unless that's you, will be here in six months in that capacity. And what I don't see is the discipline that software developers have when they track their code, and so somebody can go in and actually look at it. That doesn't exist in a lot of the builds. So I think I'm getting to a rule, which is: Build when the capability is genuinely differentiating. Like, you can't find it out there, and it will be a competitive advantage. But be more skeptical about rebuilding commodity infrastructure. I think, um, just because you can build it doesn't mean you should. And probably the most important thing is with the old world of SaaS, we kind of knew, we knew what our costs were. With AI, when you build something, it isn't until you build it that you know how, how much it costs you to build it, and then you don't know how much it's gonna cost to maintain it and actually continue to use it. So that's, uh, an important consideration. And until you understand the unit economics of AI, again, I'm not saying don't build stuff. I think there's amazing things that you can build. Just pause. Hit— take a beat before you approve something pretty broad that might be customer-facing, that might involve, um, something that you— is, is really important to the business long term. So I, I have another Super Huddle sponsor that I can give a shout-out to, our friends at Conversion AI. Um, they are a native marketing automation platform, so you can dump Marketo, you can dump Pardot, which you probably should have done a long time ago. Sorry, Salesforce people. Their argument is around replacing legacy marketing infrastructure, and it raises a really provocative question. If the underlying platform is becoming AI-native, should your team really be spending its time creating pieces of infrastructure themselves? So they've created this, uh, tool, and what I think is so cool about the Conversion product, having talked to Tammy Chan, one of their customers, is you don't need to s- invest hundreds of hours of training in order to use the tool. And, you know, I think, uh, Tammy was saying they had a relatively new hire straight out of school who figured out how to use it and was running campaigns in a couple of weeks. Well, gosh, that's a lot different than what it takes to, uh, to do, uh, with a Marketo or Pardot. So think about AI-native tools to replace your existing tools if they're not coming along. Don't build your own marketing automation system. I, I think that's, uh... You don't wanna be in that business. The point is, don't always buy, don't always build. It's knowing what is strategically worth owning and what you can just let these other fine people, like the folks at Conversion, build on your behalf and maintain on your behalf.

Drew: So let's see. We are getting to the end, but we're not there yet. Are we even measuring the cost of AI correctly?...

Penguin-Hat Drew: I don't know how to answer this question. I think this is the least mature, uh, area of AI management. We can sort of manage our costs, but the token usage, the integration, the security, the human re- reviews, the maintenance, there's a lot of costs that seem to be hidden. There's also just the duplication of, of agent building that's happening. So, um, Amanda, my conversation with Amanda and Michelle, which I mentioned earlier, surfaced, like, this, uh, duplication problem that organizations can end up with thousand-- I think, uh, Amanda mentioned that she's talking to somebody who built 18,000 agents and was really proud of it. I think they had a hackathon. The cost of maintaining those and figuring out which ones and integrating all of those, it's absurd un- unless you really actually have a plan. Just because you can build it doesn't mean you should. There's an opportunity cost. What did the team not do whilst spending six months maintaining, uh, something, uh, another company already sells? So, I think CFOs are getting a lot more interested in AI, uh, how the usage scales, how the costs are. I hear a lot of CMOs talking about the CFO saying, "Okay, this is the max amount of tokens. Go." So I think this build versus buy thing is gonna become a lot more important in the decisions of CMOs. They just have to be in control of it. So— Maybe a truism here: The most expensive AI tool may be the free one your team built themselves.

Drew: All right, we're moving along. With so much changing so quickly, how should a CMO decide what to experiment with next?

Penguin-Hat Drew: So, you know what? I, I want to step back by saying I'm a big believer in experimentation and having tests going all the time, and at least 10% of your budget going into experimentation. That doesn't necessarily mean that experiment has to be AI-based. If you start with a business problem, not the shiny AI object and what you can do, you have a really good chance of experimenting in the right way. The question isn't where can we use AI, it's where do we have friction? Where do we have cost? Where do we have delay? Where do we have unrealized opportunity? How could we serve our customers better? I'll give a great example. Amanda Kahlow talks about this with, with their, uh, Nigel, which is an engineering superhuman, uh, that you can get from 1mind. And so one of the problems that, that all companies have is you wish you could have an engineer on every sales call, but you can't possibly afford that. You also wish that you had engineers or, or salespeople that could work 24/7, but you can't because people have to sleep. So those are two examples of, well, if a superhuman or a bot could actually fulfill the role and be on every sales call and do it cost-effectively, imagine how much closer you can get to the sale as a result of that. So that's an experiment I would love to see more CMOs take. I think if we look at the low-risk work that isn't customer-facing, if we can aggressively automate that, great. As the stakes and the brand risk rise, increase your human involvement. So one of the things that— and it's not an experiment. Like, the scientific method is establish a hypothesis and a baseline, where are you now, and then we're gonna experiment to kind of prove it. So if you don't have a baseline with any experiment that you're doing, you're not gonna end up with a story. You're not gonna know what happened. Did you move the needle? Well, I don't know. Where'd the needle start? Try to tie experiments to activity— to outcomes, not activities, and be ruthless about stopping things that don't work. It's fine. It's good. Uh, y- if you don't, you're gonna end up with 15,000 agents and no idea who is Karen feeding them.

Drew: If the CMO playbook is really being rewritten, as you said at the beginning of the hour, what parts of the old playbook are worth keeping?

Penguin-Hat Drew: Well, like every three months I look at my book. Oh, yeah, there it is, "Renegade Marketing: The 12 Steps to Building Unbeatable Brands," and I wonder how much is relevant and how much isn't relevant in the AI era, because this came out just before ChatGPT, uh, took the world by storm in 2020. Uh, that was 2022. This is, uh... the book came out in 2021. Anyway, what I love about it is I go back to the book and some of the basic things that we're talking about, brand, about positioning, about differentiation, those things don't change at all, and AI can just help us double down on those things, and more so than not. Things like creativity, judgment, relationships, reputation, original thinking, customer understanding, all that has not changed. Um, I would argue that employees are still your first target audience. That has not changed in AI. And in fact, I think there's a lot of grumbling at companies because of the way AI has been sort of forced upon folks i-in a way that doesn't seem to, "Well, what's in it for me?" Um, it hasn't been explained very well. Just do it. And so, I think if you think about your strategic place really, really well and spend time on it and allow yourself to be in a strategic position and not be changing every 30 days your foundation, those things like are, if not more true today th- in the AI era. I mean, the irony of PR being important to AI just cracks me up. PR has always been a great way of, of building reputation and, and getting people to talk about you and, and, and getting, you know, God forbid I use that term, awareness. Uh, helping people, um, find your brand. Yes, let's let the machines eliminate the sludge, um, but don't let them eliminate the humanity. This is probably the biggest part of why I'm so excited about CMO, uh, the Super Huddle on October 22nd, 23rd. We're bringing the CMOs together not because any of us has a new playbook completely figured out. We're bringing them together because this community gets to help write it, because we know that when you bring, when we huddle, um, the whole community gets warmer, 'cause, yeah, it's warmer in the huddle. Anyway, we know this. The winners won't be the marketers who resist AI, and I don't think they'll be the marketers who handed everything over to AI either. There's a narrow pathway here, the ones who figure out where the machines make humans better and where humans still make marketing matter. It's really been fun. I, I enjoyed all these questions, uh, and, uh, how you made me sweat.

Drew: That's it for this Drew on Drew episode. If you're a B2B CMO trying to figure out this new playbook while simultaneously being expected to execute, come join us at CMO Huddles. As we like to say, it's warmer in the huddle.

Show Credits

Renegade Marketers Unite is written and directed by Drew Neisser. Hey, that's me! This show is produced by Melissa Caffrey, Laura Parkyn, and Ishar Cuevas. The music is by the amazing Burns Twins and the intro Voice Over is Linda Cornelius. To find the transcripts of all episodes, suggest future guests, or learn more about B2B branding, CMO Huddles, or my CMO coaching service, check out renegade.com. I'm your host, Drew Neisser. And until next time, keep those Renegade thinking caps on and strong!