Here’s what I keep hearing: “We’re using AI for our marketing now. We don’t really need a marketing person.”
I get it. I really do. You can ask ChatGPT to write a positioning statement, generate a content calendar, build ad copy, and even draft a go-to-market plan — all before your morning coffee gets cold. It’s impressive. It’s seductive. And it’s dangerous if you mistake the starting line for the finish line.
Let me be clear: I use AI every single day. It’s an extraordinary tool. But a tool is not a strategist. A hammer doesn’t know which wall to knock down.
The Starting Problem vs. The Finishing Problem
AI is exceptional at getting you started. Need 10 blog topic ideas? Done. A first draft of a LinkedIn post? Seconds. A competitive analysis framework? Easy.
But marketing isn’t a starting problem. It’s a finishing problem. It’s a sustaining problem. It’s a “we launched three campaigns and none of them moved the needle, now what?” problem.
I worked with a fintech company that had used AI to generate their entire messaging framework. On paper, it looked polished. The value propositions were clean, the personas were detailed, the competitive positioning matrix was comprehensive. Textbook stuff. The problem? None of it reflected how their actual customers talked about their actual pain points. The AI had produced a beautiful document that described a market that didn’t quite exist. It took us three weeks of real conversations — with prospects, with the sales team, with churned customers — to find the real story.
AI gave them a starting point. A human found the truth.
When the Biggest Brands Get It Wrong
If AI could replace marketing judgment, you’d expect the world’s biggest brands — with the biggest budgets and the smartest teams — to get it right. They haven’t.
Coca-Cola, a company responsible for some of the most iconic advertising in history, went all-in on AI-generated ads for their 2024 holiday campaign. The reception was brutal. Audiences called it soulless and low-effort. Alex Hirsch, the creator of Gravity Falls, joked that Coca-Cola’s red comes from the blood of out-of-work artists. They tried again in 2025. More backlash.
McDonald’s rolled out AI-powered voice ordering across 100 locations. The AI kept adding random McNuggets to orders and stubbornly refused to remove bacon from McFlurries. The experiment was scrapped.
Google’s Gemini AI generated historically inaccurate images that sparked outrage. A London event company used AI-generated visuals to market a children’s “Willy Wonka Chocolate Experience” — gorgeous, fantastical imagery that bore zero resemblance to the sad, half-empty warehouse attendees actually walked into. The police were called before the day was over.
These aren’t small companies with no resources. These are organizations with massive marketing teams, enormous budgets, and access to the best AI tools on the planet. They still got it wrong because AI doesn’t have taste. It doesn’t have instinct. It doesn’t understand what Rory Sutherland calls “the logic of the irrational” — the messy, human reasons people actually make decisions.
The Five Pitfalls Nobody Talks About
Harvard Business School professor Julian De Freitas published research in 2025 identifying five psychological pitfalls of AI in marketing: people blame AI first when things go wrong, one AI failure contaminates trust in others, consumers judge companies harshly when they overstate AI capabilities, humanized AI gets held to a higher standard, and people are outraged by what they perceive as programmed manipulation.
Here’s what this means practically: when your AI-generated content misfires — and it will — the reputational damage is amplified, not reduced, because AI was involved. The margin for error is actually smaller, not larger.
Meanwhile, a Deloitte survey found that nearly 70% of consumers are concerned that AI-generated content is being used to deceive them. Your audience isn’t just indifferent to AI marketing. They’re actively suspicious of it.
What AI Can’t Do (And What Actually Builds Pipeline)
AI cannot sit in a room with your founder and extract the one sentence that defines why your company exists. AI can listen to three sales call recordings and spot the pattern that your messaging completely misses the buyer’s real objection. But, AI cannot look at your competitor’s latest campaign and say, “They’re vulnerable here — and here’s why.”
AI cannot tell you that your brand’s tone feels off. It cannot tell you when to zig while everyone else zags. It cannot mentor your junior marketing hire through their first campaign launch, or tell your CEO — respectfully but firmly — that their pet project won’t work.
These aren’t nice-to-haves. These are the things that actually build pipeline, differentiate your brand, and compound over time. This is also why depth with one tool beats surface-level fluency across fifteen.
So Where Does AI Fit?
Right where every good tool fits: in the hands of someone who knows what to do with it.
AI is your research assistant, your first-draft machine, your brainstorming partner, your data cruncher. It’s phenomenal at acceleration. But acceleration without direction is just a faster way to get lost.
The question isn’t whether AI is powerful. It is. The question is: who’s steering?
A fractional CMO — or a full-time one — brings the judgment, the pattern recognition, the taste, and the gut instinct that comes from years of watching what actually works versus what looks good on a slide. AI can produce the slide in seconds. A human knows whether the slide should exist at all.
If you’re curious what that human judgment looks like in practice — I’ve written about what the fractional CMO role actually involves.
The Bottom Line
Use AI. I do, enthusiastically. But don’t confuse a tool with a strategy. Don’t confuse output with judgment. And don’t confuse getting started with getting it right.
Your marketing needs a brain. AI is a brilliant calculator. But it’s not the brain.