GenAI Marketing MacGyvers

Drew: Hello, Renegade marketers. Every 10 episodes or so, I get to do something I love, which is to grab two microphones, one, two, and ask myself a bunch of questions and try not to dodge them. Today we're talking generative AI, not the someday soon kind. I'm talking about what CMOs are actually doing with it. What's working, what's flopping, and what you might want to try before your CEO asks you why ChatGPT is better at briefing analysts than your own team. We'll cover AI shortcuts that work, how to build useful GPTs without a PhD, and the emerging role of the vibe coder and how to avoid building tools no one uses. Alright, ready to get scrappy? Here we go. All right, Drew, so you've talked to a lot of CMOs. What's the most common way CMOs are using generative AI right now?

Drew: Well, I'm so glad you asked that question. It's, you know, content creation is still the gateway drug, but it's evolving from just "generate a blog post" to summarizing and recycling existing thought leadership. This means like updating all your content on your website because that's really important to AEO, creating campaign-specific landing pages with verticalized messages. Again, yes, it's content, but it's very specific for use cases, translating executive ideas into usable content. Perhaps you just ask them to dictate some thoughts, get the major themes, then you use AI to help amplify that. Those are pretty good, right? Yeah, right.

Drew: Those are good. What are some surprising shortcuts CMOs have implemented successfully?

Drew: Okay, so we've got several. So we've got AI-powered SDRs. Not for all use cases, but they're running email cadences based on intent data. One CMO cut touches to meeting in half, you know, meaning the number of touches that they needed to have in order to get a meeting by half by letting AI handle pre-threshold nurturing, obviously primarily via email. They also, other CMOs have brand-trained GPTs that mimic the internal voice and yes, avoid, even avoid the dreaded em dashes on command. And then AI generating competitive digest by scraping public information and piping it into Slack. Now, I do want to mention this: it'll give you a wide range of competitors and it'll give you different answers every time. So I do think it's important that you are a little more prescriptive and push it to have the right kind of competitive information to help you.

Drew: All right, so you got a lot of shortcuts. Cool. But why? Why do they matter? I mean, didn't the old way work pretty well?

Drew: First of all, they don't require massive investment or necessarily technical talent. So that's one. They're kind of easy. Two, they give the CMO leverage, especially when bandwidth is tight or headcount is frozen. And that's been a big topic this month: are we going to be able to hire more people, or is AI going to create, uh, make everybody you have more efficient? And right now, the consensus is we're going to be more efficient, so maybe a future hire won't happen. And then the other thing about using these tools and getting some nice use cases is it builds internal credibility. AI doesn't have to be a moonshot to win, doesn't have to be this massive project necessarily. And I think that's an important perspective that the CMOs can bring to the table to help drive innovation across the company.

Drew: Got another sort of big bucket for us to work in and maybe we could talk about best examples of custom GPTs from the huddles.

Drew: Oh, this one was really cool and it was a first. Someone has created an analyst relations GPT, and they had to do it because they lost a product manager, so they needed a way to quickly answer survey questions. So the analyst would share surveys and you need a lot of questions and you wanted to make sure that you could compare it to the internal narrative. So this analyst relations GPT is drafting responses, which I think is very cool. And then another CMO is working on a much more complex one, which is a revenue GPT that's actually connected to Snowflake and all their data. And what this enables them to do is to answer pipeline questions like, what's the velocity this quarter? Now there are some caveats on answers, the accuracy of the answers, so it's important that you still double-check your data. Another area is marketing GPTs are creating, or brand GPTs with a brand voice, the content pillars and partner messaging so that it acts like a custom GPT for the company that everybody can use. Another one that I threw into the mix and I strongly encourage CMOs to do is start to create executive GPTs to simulate what your CEO, what your CFO and your CRO might ask you about marketing. And just as importantly, when you present a plan, assume that the CEO and the CFO are going to run it through the GPT. So you might as well do it yourself in advance and find out what the GPT is going to say.

Drew: Okay. Those are cool ideas. Great. Thank you for that. But what mistakes are CMOs making when building these GPTs?

Drew: Ah, another softball question. Thank you for that one. Uh, one, they're skipping the strategy step. It's like, who is this for and what do they need it to do? You really need a sort of, like, you need an outcome that you want, and it's really important that you think about the who and the what. Number two is launching GPTs before mapping the workflows and priority use cases. So in other words, say, hey, we got this brand thing out there, and then you start to think about it. Wait, we don't have as much competitive data in there as we should, or we don't have as much pricing information as we should because we're going to give this to sales, we're going to give this to our SDRs. So you really have to think about where this is going to go before you create it. Then finally is assuming that everyone will figure out how to use it because we're all using these things, which is really a problem. You need to train folks, bring in professional trainers, or at least get some tutorials so that folks can get certified. And don't make the mistake that one of our huddler CEO famously asked it a pretty silly question and then said, okay, we're not going to do that because the GPT said no. Don't, you know, might as well figure out these use cases and be prepared for the silliness that can happen out there.

Drew: Okay, so we've got a GPT, but what's the best way to roll it out?

Drew: Another great question. Wow, that's a little struggling here, but one, start with a well-defined audience like sales or executives or your SDRs, or maybe it's your product people. With that defined audience, you can make sure that, again, going back to getting the content, the right content in there. And when you're building these GPTs, build it with a clear problem in mind, like saving time, increasing consistency, and obviously improving speed. Ideally, it would also improve quality. Okay, and the last thing is test it with a small group. Get the feedback, improve it before you roll it out to the entire company.

Drew: Now we're going to get to sort of a MacGyver moment. The rise of the marketing MacGyver. Let's, can we just answer the question? What the heck is a vibe coder?

Drew: You betcha. Vibe coding is essentially a scrappy, self-taught team member who's connecting tools and building, say, automations or building applications. And it's something that you can go on ChatGPT or Claude and just start doing and say, "Hey, I want to vibe code," and you can partner with it to figure it out next. They're not necessarily a developer, although they're starting to get those skills. And ideally for your marketing MacGyver, there's someone who's starting to learn some tools like Zapier or N8N or Make or Clay because it takes more than just ChatGPT to build some interesting applications. One CMO said it, "It's like someone on our team just got obsessed and now they're essential." So hopefully someone on your team will step up. But if they don't, I want you to think about where's that vibe coder going to come from? You're going to need one on your team.

Drew: All right, can we get more specific and talk about workflows that are being MacGyvered right now?

Drew: Well, in truth, it's a wide area, but one is like scrappy competitive activity, which you can summarize via a GPT and then share it via Slack. Another is, say, generating video scripts for sales enablement. It might be auto-tagging and sorting incoming content requests based on urgency and audience. So maybe it's reading your emails and sorting those things out. We're working on a major web workflow project with Level that has all sorts of different steps with the hope that we can present to the CMO community just to talk about what it takes to actually build some of these tools. And then lastly in the MacGyver is RevOps-style AI tools that track pipeline and content performance, kind of like we talked about a second ago.

Drew: Okay, so how do you enable these MacGyvers?

Drew: Okay, so first you've got to give them permission to tinker. We had Udi Ledor; he talked about psychological safety and the willingness to make mistakes as a way of driving experimentation in the marketing department and around it. Well, the same thing is here too. You have to give them a safe space where they have permission to tinker and to fail. It's okay and to share those failures. Ideally, they have a few small wins that save some time and reduce some steps, and those need to be celebrated as do the failures. And then consider creating a formal AIOps role inside of marketing. Perhaps they report up to marketing ops. But I think we're going to get to this point. Just last week we saw a job posting at a big company—it might have been Okta—for an AI specialist to go into marketing ops.

Drew: Okay, next up, let's talk about some of the most common generative AI traps and how you can avoid them. So what should CMOs not do when implementing Gen AI?

Drew: One, don't treat it like a toy. Random prompting leads to random results. Two, don't expect AI to do strategy. This is really important. Garbage in, garbage out. It can help you think through strategy. It can create structures for strategy, but remember, it's working on past data and it's just an amalgam of a lot of thinking. It is not the same as original thinking, so be careful in the strategy area. Use it as a thought partner. If you're going to launch some tools, don't launch them all the way across the organization. Don't do it without training. Tools without context won't get used. You know this because you brought in a lot of MarTech that didn't get used, and you know why? Because you didn't have the skilled people to do it. It's the same way with generative AI. It seems simple, like everybody knows how to use Google, but in truth, there were people who were better at using Google than others. So that's the thing. We've got to train people to do it.

Drew: Okay, now we're avoiding traps. What's the best advice for CMOs just starting out? Maybe they're new in their role or they finally got permission, because some companies were slower because of various concerns, to use some of these tools.

Drew: Okay, so number one, start small, win fast. In other words, content is the low-hanging fruit, as I mentioned. Build a GPT. It's not hard. We've got a Drew GPT I use all the time. It's so helpful. Frankly, it's really easy if you have someone in your organization who has written a book because you can upload the PDF of that. Now, if you're worried about its support, I'm going to make this point. If you're worried about Claude or ChatGPT using your content and borrowing it, you can turn "learn from my content" off so your content will stay private. Number two in this area is partner cross-functionally. Your ops and analytics teams probably have ideas on ways that you could use this more efficiently, and what you're really thinking about is the challenge with ChatGPT initially and a lot of folks is that it was a solution in search of a problem. So think of the problems, and like workflow is a great one where there are repetitive tasks over and over again, and then start to think of some outcomes that would be better. Can you have faster meetings or faster meeting reports? Can you have better content? Can you update the content on your website faster? And can you create content that actually gets more engagement rather than less through an analysis of the content?

Drew: Okay, this is really good, but where the heck is this headed?

Drew: I might use the answer, "What do I know? I'm a penguin," but I have some thoughts. Okay, I think you're going to start to see—you know, we're going to move from siloed hacks to integrated systems where the AI and your CRM and your web and your analytics are all going to come together. I'm seeing this at a lot of companies, like HG Insights, for example. Now that they have TrustRadius and they have MadKudu, they're thinking about how do they build this revenue intelligence platform that's fueled by AI. You're also going to have, potentially, content automation. You're going to move from there to full buyer journey orchestration. Think about it. In an ideal world, you would be able to anticipate what this prospect's information might need, might be, and you could test a bunch of different ways and you could test iterations and you could create a more efficient buyer journey, a better buyer journey. This isn't just about you getting the sale; this is about helping your buyer get the information that they're looking for. And we're going to move from, "Boy, this is fun experimentation," to where it's having measurable pipeline impact.

Drew: All right, Drew, where is all of this going? Let me answer that for you. Generative AI isn't some distant dream. It's already your SDR, your content strategist, your analyst relations partner, and maybe soon your CFO whisperer. But only if you take action.

You want to go deeper? We will be talking Gen AI and so much more at the CMO Super Huddle, November 6th and 7th in Palo Alto. You'll hear from the likes of Noah Brier about building stuff, Liza Adams, who's made digital twins famous, Jon Lombardo on synthetic research, Samantha Stark on Gen AI video, Tahnee Perry on organization design, and Webflow's Guy Yalif on AEO. Plus we've got Armen Najarian of Sift who will show us how they're using Gen AI agents to turn customer champions into pipeline. Oh, and we just landed Denise Persson, CMO of—wait for it, wait for it—Snowflake. She's going to be talking about how she partnered with her CRO to drive the tremendous growth that is Snowflake. So there you have it. Not a dream, it's already happening. We're excited to keep that conversation going, but until next time, keep those renegade marketing caps on and strong.

For more interviews with innovative marketers, visit renegademarketing.com/podcast and hit the subscribe button.

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!