A few weeks ago, I was at my kids’ track meet, standing on the sideline making small talk with another parent. She’s a nurse, and knows nothing about B2B marketing. And somehow we ended up in a 20-minute conversation about how AI has made everything sound exactly the same.
Her 12-year-old daughter had called her out for using ChatGPT. Not because she saw her use it, but because the writing sounded like everyone else’s writing.
A 12-year-old could detect it. Which means your buyers definitely can.
I’ve been thinking about this a lot lately, and a recent conversation I had for my podcast, Pipeline Brew, crystallized something I hadn’t quite been able to put into words. We’re not just dealing with a content quality problem. We’re dealing with a source problem, and the distinction matters.
AI is a rearview mirror
Here’s the thing about AI-generated content: it is, by definition, backward-looking. Every output is a function of what has already been written, published, and indexed. Which means the more your marketing relies on AI as a creative engine rather than a productivity tool, the more it converges toward the average of everything that came before it.
That’s a structural problem, not a prompt engineering problem. You can’t write your way out of it with a better system prompt.
Original thinking, genuine insight, and ideas that actually move people forward require a perspective that AI doesn’t have: one formed by what’s happening right now, in your market, with your buyers. That perspective lives in your team. It lives in conversations with customers. It comes from someone who actually does this work every day and has an opinion about it.
The marketers who figure out how to use AI for speed while protecting the parts of their work that require real human judgment are going to be very hard to compete with. The ones who don’t are going to produce faster, cheaper versions of content that was already indistinguishable.
Personalization at scale is just generic at scale
Everyone’s talking about personalized outreach. I’d argue most of what’s being called personalization right now is just variable substitution at volume.
Swapping in a company name, an industry, a persona is not personalization. It’s mail merge with better branding. The person on the other end has received that message 20 times today. They know what it is. They don’t feel seen, they feel targeted, and those are not the same thing.
Real personalization comes from understanding. You have to know what actually keeps your buyer up at night, what they care about that they’d never put in a job description, what language they actually use to describe their problems (hint: it’s almost never the language in your messaging framework). You don’t get that from an ICP spreadsheet. You get it from talking to people.
I’m still surprised by how many marketing teams treat customer conversations as a quarterly event, if they happen at all. Block an hour a week and listen to call recordings. Ask your sales reps if you can join a customer call. Do the discovery before the campaign. It’s not complicated, and the grounding it gives you makes everything downstream better.
The missing middle is a real problem, and it’s coming faster than most people think
There’s a talent dimension to all of this that I think is being underestimated.
If you ask most companies right now whether they’re hiring for entry-level marketing roles, the answer is increasingly no. AI can do that. Which is true, in a narrow sense. But those entry-level roles were never just about the output. They were how the industry built its next generation of strategists, creative thinkers, and future leaders.
The people who are running marketing teams today learned by doing work that was, frankly, beneath them. They wrote briefs. They QA’d campaigns. They sat in on calls and listened. They made mistakes on low-stakes projects before they got handed high-stakes ones. That’s how judgment gets developed.
If we stop creating the conditions for that to happen, we should not be surprised in five years when there’s a gap in mid-level talent that nobody knows how to fill. The pipeline of future leaders doesn’t build itself.
This isn’t just an industry-level concern. It’s a company-level one. If your team is shrinking and AI is filling the gaps, you have to ask: who’s developing the judgment? Who’s making the creative calls? Who’s going to grow into the senior roles you’ll need in a few years?
The answer can’t be “we’ll figure it out later.” That’s exactly when it becomes a crisis.
What this actually means for how you work
None of this is an argument against AI. I use it. My team uses it. It is genuinely useful, and the efficiency gains are real. The question is where you draw the line between what AI accelerates and what you protect as distinctly human.
My working answer: use AI to produce faster, but use human judgment to decide what’s worth producing in the first place. Use it to draft, not to think. Use it to scale the execution of a good idea, not to generate the idea itself.
And keep talking to your customers. Not because it checks a box, but because it’s the only way to stay calibrated to what’s actually true for the people you’re trying to reach. No model trained on historical data can tell you what your best customer is worried about this quarter.
That’s still yours to find out.
If you want to dig into this further, I covered a lot of this ground with Caroline Clark, Managing Director at Radish Agency, on a recent episode of Pipeline Brew.