I've been in marketing for over 20 years. I've watched the internet arrive, watched social media arrive, watched smartphones change the entire distribution of attention, and now I'm watching AI arrive. Each of those shifts felt seismic. Each one was. And each one produced the same type of breathless prediction: that the fundamentals of marketing were about to change forever.
They didn't. The delivery mechanism changed. The speed changed. The scale changed. But the thing that determines whether marketing actually works — whether someone feels understood enough to take action — has not changed at all. And that's the lesson AI can accelerate everything around, but cannot replace.
What AI Does Really Well
Let me be direct about this, because I'm not one of the people standing in the way of the technology. AI is genuinely extraordinary at a specific set of marketing tasks, and any business not using it is leaving time and money on the table.
Scale and speed. AI can produce fifty variations of an ad headline in the time it takes a human writer to draft three. It can generate email sequences, social copy, product descriptions, and landing page drafts in minutes. For execution at volume, it's the most powerful tool marketing has ever had.
Pattern recognition. AI finds what's working across thousands of data points before any human would spot the signal. It can identify which message resonates with which segment, which subject lines drive opens, which creative formats perform across different placements. This is genuinely useful intelligence.
Iteration and testing. The old way of A/B testing meant waiting weeks for statistical significance on two variants. AI-driven tools can run multivariate tests continuously and adapt in near real time. The feedback loop is compressed in a way that used to take entire analytics teams.
All of this is real. I use these capabilities in my own work and with every client I take on. I'm not here to argue against AI. I'm here to argue for understanding exactly where it runs out of road.
AI can optimize the message. It cannot originate the insight that makes the message matter.
The Thing It Still Can't Do
Empathy is not a soft word. In marketing, empathy is a strategic capability. It is the ability to stand in the specific moment of a specific customer's life, feel the anxiety they feel, understand the desire they have, and communicate in a way that says: I see you, I understand this, and here is what you need.
AI can simulate empathy at the surface level. It can write copy that sounds warm. It can produce personas based on demographic data. It can generate messaging frameworks that hit the right emotional triggers. But it is working from patterns in existing data — which means it is, by definition, derivative. It can remix what has already worked. It cannot originate the fresh insight that comes from genuinely understanding a customer's reality.
Strategic positioning works the same way. The question of how to position a brand is not a data problem. It is a judgment call about what a company stands for, how the competitive landscape is likely to shift, what cultural currents are moving in a direction that creates an opening, and whether the organization has the conviction to hold a position under pressure. That judgment comes from experience, from reading rooms, from pattern-matching across decades of watching what actually holds and what collapses.
Timing is the third piece. There is a reason the same message lands differently depending on when it goes out. Launching a premium product during a moment of broad economic anxiety requires a different frame than launching during a boom. A message about community resonates differently after a shared cultural moment than it does in ordinary time. AI can identify historical patterns in timing — but it cannot read the present moment the way a seasoned marketer can, because the present moment is always slightly outside the training data.
The Real Danger
Here's where I want to be clear about the risk, because I see it happening already. When brands fully automate their voice — when every piece of content is AI-generated, every campaign is AI-briefed, every creative decision is delegated to a model — they lose the specific human signal that made their brand identifiable in the first place.
Brands that fully automate their voice don't sound bad. They sound generic. And generic is the most dangerous place a brand can be, because it's invisible. You won't get negative press. You'll just slowly stop mattering.
The best brands have a voice. That voice comes from a specific point of view, which comes from a specific set of beliefs held by real people who built the thing. When you replace that with AI output, you don't get an upgraded version of that voice — you get an averaged version. You get the statistical center of what a brand in your category tends to say. And that is the opposite of positioning.
I've watched this happen with content. Teams switch to AI-generated blogs, social posts, and email campaigns. The volume goes up. The quality seems fine. But over time, the engagement drops. The audience feels the shift even if they can't articulate it. There's no human in there anymore. And people don't connect with algorithms — they connect with other people.
How I Think About It
The frame I use — and that I bring into every client engagement — is this: AI owns execution, humans own strategy and voice.
The creative brief still has to come from someone who understands the audience. Not just demographic data about the audience — the actual feeling of what it's like to be a person in their situation, wanting what they want, afraid of what they're afraid of. That requires a human perspective, informed by research, experience, and the willingness to make judgment calls.
Once that perspective is clear and articulated, AI becomes incredibly powerful. It can take that brief and generate a hundred executions. It can test and iterate. It can find the version that performs. But the brief — the strategy behind the brief, the positioning it reflects, the emotional truth it's built on — that still has to come from a human being who did the thinking.
The practical application is straightforward. Before you use AI to create any marketing asset, someone needs to answer three questions from a place of genuine understanding: Who exactly are we talking to right now? What do they actually need to hear? Why should they believe us? If you can answer those three questions with specificity and conviction, AI will amplify your message effectively. If you skip those questions and hand the work to AI, you'll produce a lot of content that says nothing.
What Over 20 Years Taught Me
Every major shift I've lived through in marketing produced a period of panic, followed by a period of adoption, followed by a leveling out where the fundamentals reasserted themselves. The brands that survived each shift were not necessarily the ones who adopted the new technology fastest. They were the ones who maintained clarity about who they were and who they served, and then used the new technology to deliver that message more effectively.
AI is the fastest delivery mechanism marketing has ever seen. The scale it enables is real. The efficiency it creates is real. The speed is real. But a fast delivery system still needs something worth delivering. That something — the insight, the empathy, the positioning, the timing — is still a human job. It has always been a human job. And over 20 years of watching trends arrive and depart tells me it will remain a human job for longer than the current conversation acknowledges.
Use the tool. Use it aggressively. But stay in the driver's seat. The car is faster than it's ever been. That's not a reason to take your hands off the wheel — it's a reason to know where you're going before you press the accelerator.
