Welcome back to the Marketing Remix. In this new series, The AI Edit, we're here to demystify AI and marketing, separating fact from fiction to help today's marketing leaders make smarter, more strategic decisions.
From real-world applications to bold predictions, this series dives into the tools, trends, and challenges shaping the future of our industry. Whether you're a CMO navigating tech investments or a strategist exploring new efficiencies, this series is built for you.
Reid Carr: AI is everywhere, but how much of it is actually helping marketers do their jobs better? In this first episode of The AI Edit, we're separating the signal from the noise to help you understand how AI is really impacting marketing in 2025.
To kick us off, I'm joined by two people who help shape where Red Door’s going and how we help clients evolve—our VP of Data and Innovation, Ron Hadler, and our President, John Faris. Together, we'll tackle the biggest question marketers are asking right now: what’s actually working in AI, and what’s just hype?
Thanks for joining us, John, Ron. Good to have you here.
Ron Hadler: Excited to be here.
John Faris: Excited for sure, for sure.
What's Real and What's Buzz Right Now?
Reid Carr: Right on. So let's start this off real quick. What is real and what's just buzz when it comes to AI and marketing right now? Ron, what do you think?
Ron Hadler: Yeah, I mean, there is actually quite a bit that is real. When you start talking about the impact it’s having—things like coding, analytics, personalization—it is happening, and happening at scale. So those things are real, and they're boots on the ground. You see a lot of these very large enterprises really reducing the number of developers they're bringing on, freezing hiring. So that's a direct thing.
John Faris: Yeah.
Reid Carr: What about you, John? What are you seeing?
John Faris: Yeah, I'm a little hesitant to say anything in AI is overhyped just because I don't think it'll age that well at this moment. AI's as dumb as it's ever going to get, and it's as underutilized in business as it will ever be. So eventually, the reality is going to catch up to the hype on pretty much everything AI-related.
If I had to choose one thing, I’d say that expecting AI to provide a finished product on creative assets—it’s getting better at generating text, for example, which it used to be terrible at. But if you ask it to use the same logo every time that you've uploaded, it'll start to riff on that logo and make changes. That can be really damaging from a brand safety and consistency perspective. So I'd say it's just not quite there yet, but I think it will get there.
Reid Carr: Yeah, it is interesting. I think whenever you think of something being overhyped or underhyped, if it’s still kind of blowing your mind day in and day out, it’s hard to say that anything is overhyped. I mean, there’s always something where you’re like, “Oh my goodness, that was awesome,” and then some things where you're like, “It should do this—why is it not doing this?” Right? You would think it would.
What areas of AI are underutilized?
Reid Carr: So with that in mind, what would you say are some of the specific things—John already talked a little bit about what’s not working as well as expected—but what’s something underutilized right now?
John Faris: For me, I’d say GPT deep research. It's so powerful to basically get two hours' worth of Googling and another five hours of aggregating and summarizing those things into something coherent—that OpenAI’s deep research can do in a matter of a minute or two, sometimes 10 minutes if you ask it a really complex question. But that’s really underutilized right now by a lot of folks.
Reid Carr: What are you seeing, Ron?
Ron Hadler: Yeah, I love the research—that’s an awesome answer. I think folks are still very, very hesitant to put a lot of data into that, right? Personal details and things like that. I've got both business and even personal examples where it’s done miraculous things.
But I think getting data into ChatGPT, just even through the interface—uploading a spreadsheet and really pulling insights from that—I think there are a lot of folks who aren’t doing that well. Even subtle things like cleaning the data, stitching it together, which can be thousands and thousands of records—doing that manually is very difficult. Allowing something like ChatGPT or another LLM to do that is pretty amazing, and I don’t think we’re utilizing it well enough.
As for putting personal information in there—I've had a situation where doctors were stumped, and I put in personal health info. It was more accurate than the emergency doctor and other physicians. While that’s problematic—you don’t want to give it too much info—I at least got an answer to calm myself down.
John Faris: And Ron, how do you prevent that data from going out into OpenAI’s training model?
Ron Hadler: Two things: Temporary chats are one of the most important features. In ChatGPT, top right, temporary chat—it’ll go away as soon as you click off. The second is to use something like Teams—those accounts don’t share anything with the broader model. And third would be to use the API. That, of course, requires coding and an application, but it prevents your chats from becoming part of the model. Anybody using the free version—everything you say is part of the model. You are the product.
Reid Carr: Yeah, right. Anything free—you are the product. I think that’s the way it typically goes.
What are the core skills marketers need to learn?
Reid Carr: Now, you say all that as if it’s kind of easy-peasy, but it does sound like there’s still quite a bit of runway for folks to try stuff, explore. But for some of the bigger things like putting data together—like you just said—what are some of the core skills people need to learn right now?
Thinking about our audience—marketers—what can they learn right now that seems approachable to get into the things you're talking about?
Ron Hadler: So I think you need to attack workflows and how you want to go after them. As a marketer, you already have a set of skills. You are what you don’t automate. So if you look at the things you don’t like to do, that’s where you want AI to come in and really help.
Skill sets? It’s chatting. As long as you can chat and not expect a one-shot answer, you’ll be successful. A lot of folks fall down because they expect a single perfect answer. It’s a conversation—rarely do you talk to a human and get the full answer in one question.
Reid Carr: Yeah. One thing I’ve noticed when asking those questions is how often I’m not precise—I realize that someone’s inferring things from my tone or body language in real life. These models don’t pick up on that, so I’ve gotten better about clarity.
Ron Hadler: I mean, it teaches you that you don’t know English very well. If you don’t know what a word actually means, it’ll help define it. It acts on the true definition. It knows English better than most English speakers.
Reid Carr: Yeah. Oh, that’s fascinating.
How should leaders invest in AI?
Reid Carr: Well, okay, so pivoting a little bit from all of that—obviously this is a time investment to sort out—but what about the actual investments? From a monetary standpoint—for CMOs, decision-makers, leadership—how should they judge making AI investments? What’s their approach?
John Faris: With technology in general, you have the question of should I build or should I buy? And building in this environment, at the pace that things are moving and the amount of money that companies like Meta, Google, OpenAI are putting into their AI products—I think building’s pretty risky unless you have a lot of money and resources to throw at it. You’re just not going to be able to build fast enough to keep up with them.
There’s so much benefit you can get out of a $25 to $200 subscription—what I’d call subsidized rent from Google and Meta, particularly in their advertising suites. There's quite a bit you can do there.
I don’t know if you heard what Mark Zuckerberg said recently, but they want to basically do it all. Same token—advertising and promotion are one sliver of marketing. It's an expensive sliver, but there’s so much more work to do in brand strategy, segmentation, targeting, positioning, brand diagnosis—all those foundational elements of marketing. You want to be able to use AI for those too, not just advertising and creative. That’s where the more classic buy decision comes in—whether it's OpenAI, Google's AI suite, or Microsoft. That’s the easiest entry point right now. If you want to build, do it in parallel—but don’t wait to start.
Reid Carr: Yeah. Ron, what do you think? We talked a little bit about security too. That has to play a pretty big role.
Ron Hadler: Yeah, security is a huge consideration. We already talked about what happens when you put too much info into a free, open model—that info becomes part of the model.
When it comes to security, there are lots of models out there—DeepSeek being one, made in China with close relations to the Chinese government. Not a great place to start entering information, regardless of what their terms of service say.
While security is important, I’d say part of the investment CMOs should really hone in on is: is there a direct impact? Don’t get lost in the new shiny object. Make sure you're solving a problem and seeing immediate results. Not tens of thousands of dollars of investment. These things move so fast—you can do a lot with even a $20 subscription, like John said.
How do you approach AI governance?
Reid Carr: Yeah. Shiny object syndrome is common organizationally and individually. People see something on LinkedIn, they want to try it. That brings up governance. John, from a leadership standpoint, how do you approach governance within your organization?
John Faris: Having an AI policy is really critical—basically a requirement now. If you’re a marketing agency like us, legal teams from clients are asking about it. It ends up in contracts, so we have to deliver on it.
In summary: make sure people are using sanctioned tools that have been vetted. Don’t use a local model of DeepSeek on your home computer when you're acting on the company’s behalf.
Implement the right security settings—ideally forced into paid, company-tied accounts (not your Gmail). On an executional level, make sure people provide expert inputs—data and prompt-wise. They need to review AI results critically and layer their own expertise. If AI is used to produce final outputs, be transparent with clients about that. There's movement toward being transparent with the public too—like content provenance.
Reid Carr: Yeah, I’m surprised how many times clients’ legal teams haven’t explored this at all.
Reid Carr: I think what I want to emphasize for our listeners is: if you're in leadership and haven’t seen an AI policy, it probably doesn’t exist. Marketing—maybe more than any other function—should force the issue with legal. Legal teams should extend that governance to vendors too.
There are cascading effects. Even if you don’t directly release information, a vendor might. And that’s a risk. Your customer data is critically important.
John Faris: Or an overzealous employee—they're trying to do the right thing by experimenting with AI but just don’t know how to do it safely. Provide the policy, train folks on it, and give them the tools to use AI responsibly.
When should a team trust their insight over AI?
Reid Carr: So you’ve got overzealous folks who want to get into AI and use the tools—but there's also human intuition. When should a marketing team trust their insight over AI?
Ron Hadler: When it comes to human emotion—that’s hard to replace. If you’re trying to evoke emotion or need an emotional decision, that’s where humans should step in. AI can offer options, but having a human in the loop is key.
That’s even where some hype around AI autonomy comes in. We’re not quite there yet. Like John said, anything we say about the future may make us look foolish, but right now—autonomous work is difficult.
A recent example: a coding platform called Cursor had an autonomous bot that created fake policies. Those fake policies caused customers to cancel subscriptions. That’s what happens when you remove the human from the loop.
Are people too trusting in these tools?
Reid Carr: So would you say people are trusting these tools too fast, too soon?
Ron Hadler: I think we’re getting a little dumbed down. Junior associates leveled up fast—they produced more professional answers. But it’s like how the previous generation got addicted to doomscrolling on social—we’re seeing folks addicted to easy answers from AI. They’re not using critical thinking. So that’s where you need to invest—in strategy and real thinking.
Reid Carr: Like how we don’t remember phone numbers anymore—we’re offloading things to machines and using those parts of our brain less.
John Faris: In marketing, attention is everything. You get it through out-of-the-box creative thinking. And AI—trained on everything that’s been done—isn’t truly original. We might get there someday, but not through LLMs. That real creative spark just isn’t there yet.
Reid Carr: It can sound creative, but it’s still derivative.
John Faris: Exactly.
Reid Carr: Some argue all stories are derivative—but you really feel it when it’s forced.
Where is marketing headed?
Reid Carr: So with that, where’s marketing headed next? Not just AI, but with all these tools available—where are we going?
I’ll weigh in: these tools let us do things we couldn’t before—either because of time or budget constraints. So we’re not just doing cheaper work—we’re doing better work.
John Faris: One example: segmentation. It used to be too expensive for mid-size companies. With synthetic data, you can do it faster and cheaper. It democratizes access.
Where's marketing headed? More broadly, where's knowledge work headed? That’s existential.
We believe brand building and performance marketing are greater together. AI takes that to the next level. We’ve kind of reached peak performance marketing. Brand building is critical—not just for humans, but for AI too.
How do you get surfaced in ChatGPT when someone asks about your category? Strong brand presence across the web. That’s what the model pulls from. So brand building’s importance grows, and we’ll become super efficient at performance marketing.
Reid Carr: Yeah. Peak performance marketing. Oh dear.
Ron Hadler: What’s interesting is how fast things can be done now. But it also means marketers may have more clients and more fragmentation. So their attention is split.
AI can help—but we might also see AI employees by the end of the year. That’s got some folks rattled.
There are two paths: You’re a marketer overseeing three AI agents—you’re doing what four people used to do. Or, a more draconian path: an AI boss watching over your shoulder, correcting mistakes.
I'd rather be the boss of agents than have one as my boss.
Reid Carr: Yeah. I hope there’s a middle ground. But that’s fascinating—thinking of roles and tasks, and what gets "gentrified," which is a word I’m hearing more often.
What actions should marketers take now to avoid regret later?
Reid Carr: Before we wrap, let’s ask: what will marketers regret not doing now to prepare for an AI-powered future?
Ron Hadler: First-party data. We’ve been saying this for three years. If you’re not leveraging it, you’re going to regret it. And not just having it—make sure it’s clean. If your data’s messy, AI’s going to give you dumb robots.
Reid Carr: Yeah. Data hygiene. Even in personal ChatGPT use—you can see what it knows about you. Sometimes it gets weird—maybe you used it for someone else. Like your Netflix account—“Why are you recommending this? Oh, someone else used my profile.”
John, what do you think?
John Faris: Document workflows. Have people document their own. Provide training. Give access to tools.
If you’re not doing that, you’re falling behind. You don’t know every nuance of a role—they do. And they’ll know how best to infuse AI.
Reid Carr: With workflows, it goes back to what Ron said: where should the human be in the loop?
When people document workflows, they may fear they’re building something to replace themselves. But the goal is: where should you be in the process?
Offload what you don’t want to do—or what AI can do better.
How should marketers document workflows?
Reid Carr: Are there other things marketers should think about when documenting workflows?
John Faris: Exactly what you said—imagine what you couldn’t do before. Level up. Be the boss of AI.
Ron Hadler: We’re moving up the food chain. AI handles repetitive tasks—we handle strategic ones. And one of the smartest things we did at Red Door? Focus internally first—on our workflows—before delivering AI for clients. That made us faster, more efficient, more aligned. That’s a valuable lesson.
Reid Carr: Yeah. Shiny objects are tempting. But like we say in marketing—“Nothing kills a bad product faster than good advertising.” If your shiny AI tool doesn’t deliver—or worse, mishandles data—that’s a big problem.
What is your one line of wisdom for navigating AI?
Reid Carr: Let’s wrap. Final thoughts—one-liner words of wisdom for marketers navigating AI?
Ron Hadler: AI’s not going to replace marketers. But marketers who use AI will replace those who don’t. It’s a PhD in your pocket—or a threat. Those with the PhD are going to zoom past everyone else.
John Faris: It’s more than just an intern in your pocket now—it’s also a PhD. The best time to invest in AI was 18 months ago. The second-best time is now. Get on it.
Reid Carr: This was fun, guys. Today we covered what’s real in AI, the red flags, and how leaders can avoid shiny tool traps. We emphasized human creativity, clarity, and long-term thinking.
Thanks for joining us, Ron and John. Always a pleasure. Looking forward to more conversations ahead.
John Faris: Thanks, Reid.
Ron Hadler: Thanks, Reid.
Reid Carr: If you found this helpful, be sure to subscribe, leave a review, and share with your team. Visit reddoor.biz for more insights on AI strategy and the future of marketing.
Join us for our next episode. We’re diving into where AI is making major waves: search. From organic visibility to paid performance, everything is shifting. We’ll break down what’s changing and how marketers can stay ahead.
Check out show notes from this episode and more at reddoor.biz/learn. And as always, subscribe to The Marketing Remix and leave us a review on Apple Podcasts.