Ep #137: Why AI has been a major turning point for us CMOs

Transcript:

00:00:00 Casey: In this episode, we’re gonna dive into what I’ve been playing with Claude Code and Anthropic’s latest model, Opus 4.6, and why I think that this is a major turning point for us as CMOs. Let’s go.

00:00:18 Casey: Marketers of the world, why do we work hard to solve small problems? Why do we reinvent ourselves and our clients over and over? And why are we giving away marketing strategy for free? With advancements in AI, we’re all seeing the marketing department shrink from the bottom up. And companies need you to serve them as their fractional chief marketing officer.

00:00:41 Casey: It’s time to solve bigger problems and bring home a bigger paycheck. It’s time to create the lifestyle we deserve and to make a greater impact. This is the Fractional CMO Show, and I’m Casey Stanton. Join me as we explore this growing industry and learn to solve bigger problems as marketing leaders. The Fractional CMO Show is sponsored by CMOX, the number one company to teach you how to attract, convert, and serve high paying fractional CMO clients on your terms. 

00:01:14 Casey: Hey, welcome back. It’s Casey here, and I want to tell you about my experience with Opus 4.6. And what I’m seeing right now really feels like a demarcation between before and after. Like you, I’ve played with the AI tools here and there, I’ve paid for Perplexity and Claude and ChatGPT, and find myself defaulting to ChatGPT. Oftentimes, to answer questions or to help me with small tasks or just the kind of general stuff that you do. It saved me time. It’s good. I enjoy it, but it hasn’t made a real fundamental difference for me in a lot of ways because I’m not an implementer, right? I’m a CMO. I’m a chief marketing officer. I don’t write emails for my clients, so therefore I don’t need to leverage ChatGPT to write emails for me. I also like writing too, so it doesn’t help that much. 

00:02:11 Casey: But I really feel as though everything changed on February 5th, 2026. That’s when the Opus 4.6 model came out, and a buddy of mine bumped me to check it out. And I did. And what I found has completely floored me. It feels like Opus 4.6 is the possibility that we were told. It is what we heard AI could do. And there’s plenty of problems with it. But I found it to be very interesting. So I want to get into that. 

00:02:47 Casey: But first, I want to give you just my hesitation with AI. You know, I think above all, my biggest hesitation with AI is that it makes the user mentally weak. And I think that that is a significant issue for me to just kind of keep reminding you on. We want to be Renaissance men and women. We want to kind of go back and learn how things work. We want to be interested in it. The problem with this latest Opus 4.6 model is when it runs stuff for you, it runs it so expertly well that you don’t even have to really know what it’s doing and you just get an output and it feels like success. And that just feels sloppy.

00:03:27 Casey: My fear is that you will overcorrect by using AI too much, you will give up critical thought, and you will become intellectually weak. You will lose your edge. Just like you’ve lost your edge in long division now that you can use a calculator at any time. Using AI means that you’re not really critically thinking. And this is happening to me, and I see it, and it kind of frustrates me. Like the quality of my prompts into AI sometimes are absolutely atrocious. You know, they lack proper spelling. It’s just kind of like I’m falling asleep while I’m typing them. That’s like how it reads. I mean, they’re terrible. Yet the AI can surmise enough of the intention of what I’m trying to do and it can do it to a level of success for me. And that feels interesting. 

00:04:21 Casey: So becoming weak-minded is the biggest risk with AI. To you, professionally, I think there’s other significant risks. We’re not gonna talk about environmental risks, but I think job loss is a major one to be considerate of. I think what I’ve experienced with Opus 4.6 tells me that I probably don’t have to hire developers in the way that I’ve hired them before, kind of like ever again. I would have like a small project which would be… here’s a project that I’ve kind of loved to do for a long time, which is if I send an email to someone to, let’s say, promote a webinar, and on that email I say, “Click here to register.” It always feels silly to me to take them to a page where I ask them for their email address. Because shouldn’t I know their email address? 

00:05:14 Casey: So years ago, I hired a developer and he did a great job building me a Google Sheet script that allowed me to put the email address in the destination URL of the hyperlink in the email. And when someone clicks it, it would log that into the Google Sheet, and then it would redirect them to a ‘thank you’ page. And then behind the scenes, that Google Sheet would work with, let’s say Zapier to register the applicant or the RSVP into the upcoming webinar. And that was a great piece of code that I had written. And I think it costs $1,200, that’s like the number that comes to mind. 

00:05:56 Casey: And it was very effective for me for years I ran it. And today it’s a lot easier. You can just do something like… if you’re using a GoHighLevel or ActiveCampaign or Salesforce or whatever, you could just say, “If someone clicks a link, apply a tag,” and then you create an automation that says, “If this tag’s applied, then register this person for a webinar.” So that’s a lot easier, but still this idea of me going out and hiring a developer to do that, solving a very clear and distinct problem, took him five business days, cost me $1,200, and it had its faults.

00:06:33 Casey: Today, with Opus 4.6, I can get that same script written in probably 10 minutes, maybe five minutes, and then have it tested. And that’s just different, right? I haven’t had that before. I remember telling my wife that Opus 4.6 feels like I have a $300,000 developer available at my whim to do whatever I want. So long we’ve been told that becoming a coder, going to code school, you know, like that is a great role. That’s a great job. It’s high paying and it certainly was high paying. 

00:07:09 Casey: But right now we’re seeing that it really isn’t that in demand anymore because we have these little pocket coders. We have this ability to generate code as we request it. And that is just, again, fundamentally different than where we were before. I think we’re going to see a lot of job loss from these new AI innovations. And what I think we have to do is we have to protect ourselves and our families. So we need to learn how to use them effectively. I still think that a lot of AI copy lacks soul. I find it uninteresting to read, but prompted well, it does a good job. It certainly is effective. It can be effective if done right. 

00:07:53 Casey: So I want to talk about the specific application of meeting marketing technology with Claude Code. That’s the thing that’s been, dare I say, a game changer for me. Like to be able to come up with an idea, you know, that intersection of marketing and tech, and then to be able to get it done. So as an example, if you’re familiar with Cloudflare, Cloudflare does a lot of things, but one of the things it does is it provides DNS for domains. So if you have a website like yourname.com, you can set your DNS to Cloudflare, change your name service to point to Cloudflare, and then Cloudflare manages it, which is great. There’s a free tier that’s read very rarely, do I ever feel like I need to pay for Cloudflare? The free tier is adequate. 

00:08:41 Casey: And then they rolled out their Workers, so Cloudflare Workers, which are very interesting. If you don’t know about those and you’re kind of nerdy about this stuff, take a moment to read what a Cloudflare Worker is. But Cloudflare has now, I don’t know when, but as far as I’m concerned, today I learned, Cloudflare can allow me to generate an API key for the Cloudflare Worker, give that to Claude Code, and Claude Code can set up workers on my behalf. 

00:09:12 Casey: And this is… it is a dream for me. So here’s a specific example. In a hobby of mine, I decided to start noodling around and building a little bit of software. And it was to solve a specific problem of mine. And I won’t give you the details of what the software does, ‘cause that doesn’t matter and I could talk for hours about it. What matters is, I had a problem that I wanted to solve personally for a hobby. And I set out to build software to do it. And then I shared that software with a couple of people through GitHub. And it’s kind of gotten big.

00:09:46 Casey: I’ve had 1300 downloads and at the time of recording, people have used it for a cumulative count of over 2000 hours. So I think that’s pretty spectacular. Just this cool little tool that I built. But one of the things I wanted to do, when I launched the website for it, is I wanted to track downloads and use of the app. Like I wanted telemetry. I didn’t even know it was called telemetry. But I wanted usage data and I could go and find all of these tools that provide that. 

00:10:22 Casey: But I just thought, “Well, couldn’t the app send a beacon out to some sub domain of mine and can’t I collect that information?” It was one of those thoughts that felt cool. But by the time it came out of my mouth, it sounded pretty stupid. Like, I didn’t know what I was saying, because I’ve never done it before. So I went to Claude and I said, “Hey, here’s this idea that I have. Can I get some telemetry of like… can I get some usage data of people who download the app? Where can I get that from? And then people who use it, like how often they use it, how much they use it. But I want it to be 100% anonymous.”

00:11:04 Casey: And Claude Code came back and said, absolutely, we could do it in these different ways. And following the prompts, I let it deploy a pretty comprehensive approach to 100% anonymous usage tracking, which then led me down the road to say, “Well, what about analytics? Like, can you just set up analytics on the website?” And it’s like, “Sure.” So I gave it some access to Google Tag Manager and Claude Code set up Google Tag Manager expertly. That got me. Again, that’s MarTech. That’s not the marketing that I feel strongest in. It’s kind of like something I dabble in, but I always feel a little unconfident when it comes to setting up Google Tag Manager the best way possible. 

00:11:51 Casey: So I asked Claude to do it and it did it, and I audited it myself and I am pleasantly thrilled by the result of it. So that’s the connection to MarTech. And then if you think of your clients that have more sophisticated needs. I’ve got a client that needs some content rewrites for their website and it’s significant. So it’s a lot of cities and they need multiple pages per city. This content rewrite to be rewritten by a human is quite expensive. So asking Claude to do it and providing it a multi-step approach, Claude Code then writes all of this for me based on the parameters that I set, much like ChatGPT could. But it can do it with… by resetting its context window regularly so it feels like it’s coming to it with a fresh perspective instead of kind of forgetting what it’s doing.

00:12:44 Casey: So in Claude Code or Claude Cowork… and I haven’t used Cowork, I just used Claude Code, but I’m sure Coworkers is excellent. You ask it to make a plan for something, and then it saves that plan, and then you ask it to execute that plan multiple times. And when you do it that way, where you formulate a plan and you kind of establish that as the cornerstone of your work together like, “In this instance, when we talk, this is the plan that you’re operating inside of.” Every time you invoke Claude again and ask it to do the work, he reviews the plan and does it. As a result, I found that the quality of that content that gets rewritten is much, much higher because its context window isn’t so full that it’s kind of forgetting what it’s doing. 

00:13:29 Casey: So there are these approaches where we can kind of straddle AI and technology to support us in marketing and our weaknesses. I’m not a great graphic designer, I’m certainly weak in that. I could go to Claude and I could ask it to maybe make some images and I could pull it some skills and plugins, which are kind of bolt-on add-ons that you can get from the community… and from Anthropic to extend capabilities. But for me, that’s not a use case that’s interesting. 

00:14:03 Casey: What is interesting for me is solving a specific problem that I have. So let me tell you about a problem that I see marketers make when they try to play with these platforms: is that they just build stuff that doesn’t matter. There’s two ways to do this. One way is to build something strictly to play and have fun. And then the other one is to build something that specifically solves a problem. And I think both are fine ways to do it. 

00:14:34 Casey: I started just playing and I played for a long time. Two and a half weeks so far. I’ve been coding three weeks. I’ve been coding with Claude Code to develop this app and I’ve just been having a ton of fun. But now I can take that experience that I got through the joy of learning and I can turn and I can look at a business problem and I can just… like a sniper, I can just attack that specific problem and solve it. And there’s a superpower there. And it doesn’t matter how I solved it. What matters is I identified the problem and I was able to prompt it in a clear enough way where a solution was proposed. And I read through it and I accepted it then I let it do its work. 

00:15:18 Casey: So my encouragement to you in hearing this is that first of all, I really want you to lock this in your brain that February 5th is a demarcation of before and after. And the AI that came before it is just weak. It’s like drunk. It just doesn’t work the same way. And the AI that came after, it is significantly different. Now, we recently at CMOX, we did a call with our sister company, CTOX, and these are chief technology officers. And these guys are incredibly smart and capable technologists. They said to me, “Well, I don’t know, Casey, it was really like early January when they got access to some of these models and they were using Claude Code, maybe back in November of 2025.” Okay, but for me, it was Opus 4.6 coming out in Claude Code. And that’s when it felt like everything changed. 

00:16:13 Casey: So call it when you want to call it. Say it’s Feb 5, 2026, say it’s Thanksgiving 2025. There is a significant difference. And if you’re using AI and you’re not experiencing that difference, I wager that you’re not using AI to the level that it’s available to you. So I think, you know, how do we apply this to CMOs? Well, I’ve got a project where I needed to deploy a website and I was going to hire a developer to effectively change the DNS to point from the old web server to the new one, which that’s easy. But the hard thing was there’s a ton of subdomains and the subdomains need to be mapped over to subfolders. 

00:16:56 Casey: And that was kind of beyond me. My brain kind of hurt thinking about how to do that. That’s not an HTX file. That’s not a 301 redirect. Like, where do you establish those things? So I went to Claude Code and asked the question so that I could go hire a developer. And then Claude Code said, “If you’d like, I could just build out a full plan for us right now. And then I can test it to see if it’ll work.” And I said, “Go give it a shot.” So it did. And then it said, “Everything is working. Tell me when you want to go. Just type this command in and we’ll get going.” Which was like ‘bash launch.sh’ or something. So I typed it in, I hit enter, and I sat back and watched. And 10 minutes later, the work was all done. 

00:17:42 Casey: Now there’s risk involved there. Right? I just took over the work of a developer. I took over the work of someone who manages a server. That’s scary. Maybe I shouldn’t do that. I feel a little dangerous here. I feel a little experienced. So I felt like it was an okay risk. And I had a zone backup file that I could upload to the DNS if I needed to. However, I risked it. And the reward was everything I had wanted in minutes instead of hiring a developer and getting them on and doing that whole thing.

00:18:16 Casey: So this isn’t my encouragement to you to become a technologist. Hear me clearly that the world needs you as a CMO, as a marketer. But I think we’re getting to a point where delegating tasks to humans is going to be a more laborious effort than it is to do it to AI. And for that reason, just like the start of this podcast episode said, “The marketing department will shrink from the bottom up even faster.” That’s where we’re going. 

00:18:51 Casey: So maybe you’re feeling excited about the progress of AI, which great. Maybe you’re feeling daunted or scared or behind. Or maybe you took a vacation and you haven’t really heard much about Claude Cowork or Claude Code or… I’ve got a great buddy Bill who is a technologist and he says that he’s finding that ChatGPT’s codecs is better for a lot of stuff that he does than Claude Code is. 

00:19:17 Casey: So it’s like, it doesn’t matter, right? It doesn’t matter… your platform, but like maybe you’re feeling, “Man, I’m behind, I wanna get caught up. I wanna try something.” Here’s my overwhelming encouragement. Choose a platform. I don’t care what you choose. Choose Claude, choose Chat. Just choose one of those two. And then just build a project out for yourself. Ask it to build a website for you or ask it to build some functionality in your life that would be helpful. Or ask it to set up a way so that when you give it your meeting notes, it will summarize tasks from it and add those tasks to your monday.com or your TickTick or your Trello or whatever you use for task management.

00:20:02 Casey: Just do one thing like that. Set up one connection. One thing to play with. Just try one thing. See what’ll happen. That’s it, just see what’ll happen. Go play, go have fun. When you’re done with it, you can throw it away. Be like a potter. You go sit at the potter’s wheel and you just play. You just throw some clay. You see what you can make. You can’t be attached to your first thing and say, “This has to be this monumental outcome.” You just have to play. You just have to find the joy in the learning of it and to see what comes up for you because you will have a level of innovation beyond what you can do right now. Because these things that you’ve only thought of, or maybe you’ve tasted and you’ve always wanted but you never could because the client didn’t have the budget or it took too long, or you couldn’t describe it in a way that was understandable by other people. Whatever the reason was that you couldn’t get it done before, now you can get it done and you can get it done fast and you can get it done inexpensively. 

00:21:10 Casey: For my app, I built the first version of it for my hobby. And then I did like two updates to it. So I released like two additional versions of it on GitHub, to a community based on their feedback. And only at that point was I starting to get throttled by the $20 a month Claude package. So I moved from, I guess it’s Claude Pro to Claude Max. And there’s a $100 level and a $200 level. I wanted to see how far I could get on the $100 level. So that’s where I’m at. And I put in considerable time every day—a couple hours a day, I’ve been playing nonstop for three weeks. And while my sleep has suffered a bit, I feel like I’ve grown tremendously in my ability to understand how to leverage AI. And I really think it’s changing the way I see the marketing department. 

00:22:04 Casey: So here’s your homework. Install Claude Code or get on Claude Cowork or get on ChatGPT Codex and go build something. Go build something for fun, go build something that brings you joy. Go build something for the board game that you like to play, go build something for the animal shelter that you would love to support. You don’t have to give it to anybody. Build it just for you, just try it out. Find something in your life that you could just kind of play with with the AI tools and I guarantee you will walk away from that experience feeling kind of transformed. You’ll feel a sense of accomplishment, but also clarity. 

00:22:47 Casey: And then just don’t stop, continue to play and tinker. And we talked about this a ton just a couple days ago in our boardroom chat. And it’s what we all wanted to talk about. We kind of got caught up and a couple of our members had specific questions about contract renegotiation and managing onboarding clients when you’re already have a full load of clients, how do you add the next few? We’ve got one member, sorry, two members that shared that they have a million dollar target for revenue as a fractional CMO this year, which is just a beautiful thing to witness. And I think that they absolutely have a strong fighting chance to hit it. But after we got through those general kind of questions and updates in boardroom, we spent the rest of the time together, all of us just nerding out about this and showcasing what we’ve been working on and helping each other. And it’s a community of people that are fighting to win for themselves and their families, but they’re also looking at everyone else in boardroom excited to support them. 

00:23:48 Casey: One person is rebuilding a marketing department. They joined a company and that company effectively fired the marketing department and hired the CMO, and now the CMO has to build that company from the marketing department from the ground up. And they’re asking for introductions and the folks inside the boardroom are eager to share those introductions. So it’s just a cool place. It’s kind of fast paced, but it’s people doing real awesome things. Big wins. 

00:24:14 Casey: One member said today, she sent a WhatsApp and it said something to the tune of, “I just pitched a client at X amount and they just said, can we start immediately?” Like, no other questions. Just ‘can we start immediately?’ No question about price or anything. And this is after she had increased her prices. So if you want to be around this kind of people, if you want to be the kind of person who sets big lofty goals for yourself as a CMO and that does the work that’s required, but also is like of the times and is using AI in a way that is supportive, then I’d love to help. 

00:24:47 Casey: Just go book a call with my team. Go to cmox.co/call. You can book in a call with my team, with Justin. He’s a great guy. He’s been around. He’s been with me for a couple of years. He’ll ask you a couple of questions to make sure [inaudible] we can help you. And then after that, he’ll pass you on to John and Melissa. And then you can ask them any questions that you’d like about what we’ve got going on at CMOX. And they’ll see if really this is the thing that you need to get to the next level. 

00:25:12 Casey: As you hear this, we’re going into March and we’ve started to close the doors every single month on the number of new members that we can add. Had a lot of demand. I think there’s a lot of people just realizing how important it is to become a fractional CMO. So when you have your conversation, they will share if there’s room to join in the current month. If not, we have a waitlist for the next month. But I’d love to see you on that call just so that you know if this is the right move for you or not. 

00:25:38 Casey: All right. To wrap this up, I want to put a bow on this and say, I’m not an AI maximalist. I think that there’s plenty of conversations that we could have around something called primary satisfaction, so the things in life that really bring us deep joy. Like sitting by a fire or sharing meals with people. You know, like those things are very important in our life. I don’t think AI is the panacea to solve everything, but I do think that to be an effective CMO, you have to leverage AI now and you have to do it in a way that helps your client solve for probably both speed and cost. But above all, solve for speed. That’s what you want. You want clients out… are fine paying high rates if you can go fast. And with AI, it’s the time to go fast, to innovate, to get to the next level, to build that critical mass of demand. All right. With that, I’ll catch you on the next episode. Take care. 

00:26:46 Casey: Thank you for sticking around for the full episode. As you know, learners are earners, but you’ve got to take action on what you heard today. For more information and show notes, visit fractionalcmoshow.com. If you’d like me to answer your questions on an upcoming episode, you can share your question at  fractionalcmoshow.com. And last, please hit the like and subscribe button so that I know that this content is helpful to you. Alright, go get em!