A Very Unbiased, Level-Headed Take on AI in Marketing

Let me just start by saying that if I hear (or more accurately, see on LinkedIn), “look at what (insert any of the plethora of AI tools) can do, it will revolutionize your digital marketing program” one more time, I will singled handedly invent a time machine with the sole mission of going back in time to remove the following words from the english dictionary; “will”, “revolutionize”, “your”, “digital”, “marketing”, “program”. 

Side note: I may also go back to Punxsutawney in 1887 just to ensure we never start to use a groundhog to predict the weather, let alone one that's less accurate than flipping a coin.

Now, before I sound like a bitter beaver (which would probably do a better job at predicting the weather than Phil), let me be clear: AI, especially with the recent developments in Agentic AI that makes it significantly more accessible for the everyday user, is a very powerful and cool tool. The use cases are endless, and with proper caution and preparation, can be a tool that levels up various systems in any business. I am not arguing with the notion that it provides value; that's undeniable, but the idea that these tools will “revolutionize” your marketing program is… wrong.

Now, I am sure that at least 1 of the 7 people who read this (5 if you don’t include immediate family) will say to themselves, “Hey, I tried it, and AI did revolutionize my marketing program.” To that, I would say, 

“I think that's an indication of how far behind you were, not how far ahead of the curve you are now”.  

Much of what the most recent round of AI tools do (search term mining, ad copy creation, campaign creation, performance analysis and insight generation, to name a few) are either things that could have already been automated through non-agentic AI tools (that have been around since the 1950s, thank you Alan Turing), or through custom tools built by developers using a heuristic model similar to what is used to train AI models. To be fair, AI, especially Agentic AI, has made it “easier” to make these tools and set up the automations, but it didn’t create the process as a whole.

And that’s really the important distinction.

AI didn’t invent new capabilities; it just made existing ones easier to access, faster to execute, and harder to screw up.

Which sounds great… until you realize what that actually leads to.

If everyone suddenly has access to the same tools, the same processes, and the same shortcuts, then the question isn’t “what can AI do?”, it’s, 

“what happens when everyone is using it the same way?”

What Does This Mean?

The introduction of AI to mainstream marketing hasn’t raised the ceiling of what’s possible; it’s raised the floor. The worst campaigns, strategies, and marketing teams aren’t as bad as they used to be.

But the gap between the middle and the top? That still exists. And if anything, it’s becoming more important.

The problem is that these tools lack the things that separate the bottom 50% from the top 20%. And if you’re not careful, over-relying on them can quietly move you from a top-tier marketer to a very average one.

Let me give you a tangible example:

“AI can be used to create ad copy.”

Sounds great. Is true.

But if everyone is using AI to create ad copy, and we’re all using similar models trained on similar data, we’re going to end up in a world where everything starts to sound the same. Even if you layer in your brand voice and value props, you’re still pulling from the same underlying patterns.

What AI does here isn’t create differentiation; it compresses it. It brings everyone, regardless of starting point, closer to the middle.

Want another one?

“AI can build and optimize campaigns so you don’t have to.”

Again, sounds great. Is true.

But let’s say a campaign is targeting Audience A, with a $250 CPA target, and is set to pause if it doesn’t hit that goal.

Now let’s say performance dips, not because the audience is wrong or the strategy is flawed, but because of something external:

• A competitor just got aggressive in the auction

• Landing page speed dropped

• Sales team stopped following up as quickly

• Or something else that isn’t cleanly visible in-platform

AI sees the result (performance down), not the cause. And because it operates on patterns and thresholds, not context, it makes the “correct” decision based on incomplete information.

The campaign gets paused. The problem doesn’t get solved. 

On the surface, that might sound fine: “if it’s underperforming, turn it off.”

It makes sense. It’s easy.

But if that’s how you approach it, you’re not just making your life easier, you’re making your competitors’ lives easier too, because now you’re not just pausing a bad campaign, you’re potentially cutting off an entire segment of the market that could perform if the real problems were addressed.

Instead of fixing the root cause (landing page issues, auction pressure, sales follow-up, messaging mismatch, etc.), you’ve removed your ability to reach that audience altogether.

So what happens?

• You bias your strategy toward what’s currently working, not what could work (losing future revenue)

• You narrow your reach and over-invest in the same pockets of demand

• You miss entire segments, not because they’re unprofitable, but because they were evaluated in the wrong context

All because the decision was made using a binary framework (hit target / miss target) in a world that isn’t binary. When decisions are made without context, you don’t just lose efficiency, you lose opportunity.

What Can You Do?

You can build custom GPTs trained on proprietary data that no one else has. That’s real. And it’s powerful.

But it also requires:

• Significant data volume

• Clean, structured inputs

• Time to train and refine

• Ongoing maintenance

For most businesses, that’s just not realistic.

So the alternative is…

Use human brains, especially smart ones that have a wealth of experience.

Despite what Gottfried Wilhelm Leibniz suggests (that human reasoning could eventually be reduced to mechanical calculation), we’re not there yet.

Humans are still the apex predators of nuanced thinking.

We understand that not everything is binary.

We can evaluate tradeoffs.

We can weigh context, experience, and imperfect information.

We make decisions in gradients, not absolutes.

It’s the reason strategy doesn’t sit with interns, and why the most important decisions are still made by people who have seen enough to know that the “right” answer usually depends.

That’s the part AI hasn’t replaced.

Closing Thoughts on AI in Marketing

So, in summary:

AI is cool. It’s useful. It will make your life easier.

But it’s not going to magically separate you from everyone else.

In most cases, it will do the opposite.

And if you’re thinking about replacing your marketing team (or any team) with AI… I’d think twice, unless your goal is to look exactly like everyone else in the middle.