AI adoption is messier and faster than anything marketing has lived through. The good news: the more the machines do, the more they need us.
Let’s start with a confession that almost every marketer I know has whispered at a conference cocktail hour: “I have no idea what I’m doing with AI, and I suspect nobody else does either.”
If that is how you feel, you are in excellent company. Right now, the marketing world looks like a kitchen where someone handed every cook a jet-powered blender and forgot the lid. There are pilot projects everywhere, tool sprawl, a fresh vendor pitch every hour, and a vague sense that the smoothie is going to end up on the ceiling.
AI adoption is messy. It is messier, frankly, than the digital marketing technology wave that came before it.
But here’s the thing to walk away with: this is a redemption story. Not the vague “it’ll all work out” kind, but a very specific, practical kind. And the hero of the story isn’t the algorithm. It’s the gloriously wild, unpredictable, experienced human in the mix.
We’ve Seen This Movie Before
Cast your mind back to the mid-2000s. Marketing was about to go digital in a serious way, and nobody quite knew what that meant. In 2008, AdAge created a new section called DigitalNext with the goal to make digital marketing more understood and accepted by marketers. I was one of the first contributors to write about digital for that new section. Over the next ten years, more or less, I watched digital marketing unfold in slow motion. I wrote articles like; “Traffic Verification – the most nettlesome issue in adtech” or “Who’s to blame for the Current Ad Tech Trust Gap?” or “Beware the Siren Song of Marketing Automation.” As marketing burst out of its niche marketing toolkit box into mainstream digital transformation, the growing pains were plain for anyone paying attention.
Search went from a curiosity to a budget line. Social went from college kids poking each other to a primary channel. Programmatic arrived and promised to buy media like a stock exchange. Data management platforms, demand-side platforms, supply-side platforms, tag managers, attribution tools and an ever-growing parade of acronyms followed.
Those of us who lived through it remember the famous marketing technology Luma scape charts, the ones that started as a tidy slide and grew into something resembling a star map of a very crowded galaxy. Every year there were more logos. Every year, CMOs were told that the next platform would finally bring order to the chaos.
It didn’t, exactly. What actually happened was slower and more human. Teams learned. Some tools stuck, many didn’t. Job titles were invented (remember when “growth hacker” was a thing?). Organizations rebuilt themselves around data, sometimes gracefully, often clumsily. By the time the dust settled, roughly a decade and a half later, digital wasn’t a department anymore. It was just marketing.
Fifteen years. That was the time span to figure it out. We stumbled, we overspent, we chased vanity metrics, and eventually we found our footing.
Now Press Fast-Forward
AI isn’t giving us fifteen years. Since generative AI burst into the mainstream in late 2022, the adoption curve has been compressing at a pace that makes the digital era look leisurely. If digital transformation was a fifteen-year marathon, AI transformation looks like it will play out in less than half that time. We’re already a few years in, and the tools available today would have sounded like science fiction to the marketers who were arguing about banner ad viewability not that long ago.
And the consequences may be bigger. Digital marketing technology changed how we bought and measured media. AI is changing how we think, write, plan, analyze, create and decide. It touches the strategist and the copywriter, the analyst and the account lead, the agency and the brand. It doesn’t just sit in the media plan. It sits in the room.
So yes, it feels messy. Compressing fifteen years of organizational learning into six or seven guarantees some whiplash. Pilots that go nowhere. Tools that overlap. Policies written on Tuesday and outdated by Friday. Leaders who are publicly excited and privately terrified.
Here is the reassuring part: messy is what transformation looks like from the inside. It looked like this last time, too. We just had more time to forget how awkward it was.
Why Marketing Is About to Get More Fun
I know, I know. “Fun” is not the first word most people associate with AI right now. “Anxiety,” maybe. “Budget review,” certainly. But stay with me, because I genuinely believe marketing is about to become more enjoyable than it has been in a long time.
I was documenting the digital transformation through all its messiness during the first ten years – roughly 2008 – 2019. I know the tech pain, the human stress and pressing need to demonstrate marketing ROI that justified the investment.
For the first 10 years, there was not much “fun.” That is until about 2020 and I saw the other side of the hill when marketing was able to take bold leaps and go where no marketer has ever gone before. It was thrilling in equal measure to all the angst everyone felt in getting there.
It is that history that gives me a uniquely optimistic perspective to see how AI’s transformation of marketing will play out. I know that after the turmoil, confusion, false starts, aborted pilots and a lot of questions, marketers are in for an era of where marketing heroes can create great value for their companies.
My confidence is buoyed by three characteristics of AI that, when well-applied, lend themselves to giving marketers the spotlight to shine brighter than ever before.
First, it will be more predictive. For most of the digital era, marketing was a rear-view-mirror business. We launched, we waited, we read the dashboards, and then we explained what had already happened. AI flips that orientation. Models that can anticipate demand, flag audience shifts and simulate creative performance before a dollar is spent let marketers spend more time steering and less time doing archaeology on last quarter’s reports.
Second, it will be far less tedious. Think about how much of a marketer’s week has been consumed by the grind: reformatting the same asset for fourteen placements, reconciling spreadsheets that refuse to agree, writing the fortieth variation of a subject line, building the deck about the deck. AI is extraordinarily good at that work. When the drudgery goes, what’s left is the part most of us got into marketing for in the first place: ideas, stories, strategy and the occasional brilliant hunch.
Third, it will be much better at driving real outcomes. This is the big one. Programmatic media promised precision, but much of what it optimized toward were proxies: impressions, clicks, viewability scores, cost per something-or-other. The machinery got very efficient at hitting numbers that didn’t always connect to whether anyone bought anything, trusted anyone more or came back. AI gives us a genuine shot at connecting effort to outcome, linking creative to context and detecting customer intent signals more privately than ever with visible business results in ways the auction-driven pipes of the last era never could.
More predictive, less tedious and pointed at outcomes that matter. That’s not a scary future. That’s a marketing department people will actually want to work in.
The Secret: The More Automated the System, the More Experienced Human Overseers are Needed
If AI is going to do so much of the work, where do people fit in? We hear tales of just two people running multi-million-dollar businesses – effortlessly. This is when our human feelings of inadequacies kick in. At best, we have a hard time conceptualizing this and at worse, we don’t believe this type of business is sustainable and of the quality we would want.
This is where the story turns, and where I think most of the public conversation gets it backward.
The common assumption is that automation reduces the need for expertise. The more the machine does, the less the human needs to know. It sounds logical. It is also, in practice, wrong.
Consider a simple example. Ask an AI tool to write a grant application. In a minute or two, you’ll get something that looks impressive: polished paragraphs, a confident needs statement, a tidy budget justification, maybe even a logic model. It reads well. It has all the right section headings.
Now here’s the catch. If you have never written a grant application, you have no way of knowing whether it’s a good one.
You won’t know that the funder in question cares far more about community partnerships than the draft lets on. You won’t notice that the outcomes are vague in exactly the way reviewers punish. You won’t catch that the budget narrative contradicts the timeline, or that the whole thing reads like it was written for a different foundation. The machine produced something that looks like a grant application. Only experience can tell you whether it works as one.
AI can produce the shape of good work in seconds. Only experience can tell you whether it’s actually good.
The same is true in marketing. AI can generate a media plan, a campaign concept, a customer journey map or a competitive analysis. A junior marketer might look at the output and think, “Great, done.” A seasoned marketer looks at the same output and thinks, “That audience assumption is shaky, that channel mix ignores what happened last holiday season, and that headline is going to land very badly in this market.”
Aviation figured this out long ago. Modern airliners can practically fly themselves, yet nobody suggests we put a first-week trainee in the captain’s seat. Precisely because the autopilot handles so much, the pilot’s value is concentrated in the moments that matter most: when conditions change, when something looks slightly off, when judgment has to override the system. More automation didn’t make pilots less important. It made experienced pilots essential.
That’s the paradox. Experience is not being made obsolete. It’s being promoted. You are moving from the person who does every step to the person who knows which steps matter, what good looks like and when the machine has quietly wandered off the road.
Function Meets Form: The Real Unlock
If there is one idea I’d tattoo on the forearm of every marketing leader (temporary tattoo, let’s not get carried away), it would be this: the unlock lies in understanding the intersection between AI function and human intellectual form.
AI function is what the machine does brilliantly: speed, scale, pattern recognition, synthesis, endless variation and tireless execution. It can read ten thousand customer reviews before you finish your coffee. It never gets bored. It doesn’t need a weekend.
Human intellectual form is the shape we give to that raw capability: judgment, taste, context, ethics, empathy, and the hard-won sense of what “good” looks like in a particular situation for a particular audience at a particular moment. Form is knowing which of the ten thousand insights matters. It’s recognizing the joke that will charm one market and offend another. It’s the instinct that says, “This is technically correct and completely wrong.”
Function without form gives you what the internet has started calling AI slop: endless, plausible, forgettable output. Form without function gives you a brilliant strategist buried under busywork, too exhausted to be brilliant. Put them together, and you get something better than either could produce alone: fast, scaled work with a point of view and a conscience.
The organizations that win this era won’t be the ones with the most AI tools. They’ll be the ones that deliberately design how machine function and human form fit together. That means deciding where judgment sits in every workflow, who reviews what and why, and how expertise gets passed down to the next generation of marketers who will learn the craft alongside the machines rather than before them.
So how do you design that fit? It comes down to one simple, slightly unfashionable principle: trust humans – first, foremost and forever. “Trusting” AI should be relegated to the kind of trust you have for your Roomba. You know it will do the job it was designed for but you also know it will get stuck, run out of juice or mistakenly try to suck up your cat.
Here’s what that looks like in practice.
- Trust humans to redesign the workflow. When AI changes how a function operates, the new workflow should be built by the people who actually do the work, not handed down by a vendor, a consultant or a well-meaning task force that has never shipped a campaign. The people in the trenches know where the real friction lives, which steps are sacred and which are ceremony. Let them draw the map.
- Trust humans before you buy. Every week brings a new platform with a gorgeous demo and a promise to transform everything. Resist the pitch. Only adopt new AI technology when you can get a referral or recommendation from a real person you trust who has actually used it. If no one can vouch for it yet, wait. Being second with a tool that works beats being first with one that doesn’t.
- Trust humans to be the subject matter experts. AI can sound authoritative on almost anything, which is exactly why it needs supervision. When it comes to deep knowledge of your customers, your category and your craft, the experienced human knows better. Treat the machine as a very fast, very well-read assistant, not as the expert in the room.
- Trust humans to sense when things are getting too complicated. AI makes it easy to add one more tool, one more integration, one more automated step, until the system becomes harder to manage than the work it replaced. People feel that creeping complexity long before any dashboard reports it. Listen to them. The goal is simplification, not overcomplication.
- Trust human feelings despite it all. Tools will come and go, as they did in the digital era. What endures is the judgment, taste and experience of the people using them. When the machine and the human disagree, start by assuming the human is right, and make the machine prove otherwise.
So Now What Do We Do on Monday Morning?
Encouragement is lovely, but you have meetings. Here are a few practical ways to put the human element to work.
Start by pairing your most experienced people with your newest AI workflows, not keeping them at arm’s length. The instinct is often to hand AI experiments to the youngest, most tech-forward team members. Keep them involved, absolutely, but make sure someone who knows what good looks like is reviewing the output. Experience is your quality control system.
Next, write down what “good” means. Much of what experienced marketers know lives in their heads. Turn that tacit knowledge into checklists, examples and review criteria. It makes your AI outputs better, and it accelerates how quickly junior staff learn the craft.
Then measure outcomes, not outputs. AI will happily produce more of everything. More content is not the goal. Tie your AI initiatives to the business results you actually care about, and resist the temptation to celebrate volume.
Finally, protect the time AI gives back. When the tedious work disappears, don’t fill the gap with more tedious work. Use it for thinking, for customer conversations, for creative risk-taking. That’s where the fun lives, and frankly, that’s where the value lives too.
The Wild Card Is Us
Here’s what I find most joyful about all of this. For years, marketing technology tried to make marketing more “programmatic:” tidy segments, neat funnels, clean conversion paths. And yet, for years, humans refused to cooperate. We changed our minds. We bought things for strange reasons. We fell in love with brands that broke the rules and ignored the ones that followed them perfectly.
AI is extraordinary at prediction but the people making the marketing “work” remain gloriously unpredictable. The flash of a consumer insight in the shower. The campaign idea sketched on a napkin. The gut feeling that says, “Our customers won’t like this,” even when the model says they will. Those aren’t bugs in the system. They’re the whole point.
Yes, the AI adoption curve is messier than the digital one, and it’s moving at double speed with higher stakes. But we’ve navigated a transformation before, and we came out the other side wiser. This time, we have an even better advantage: we already know that the tools are only as good as the people guiding them.
So go ahead and let the machines handle the tedium. Let them predict, sort, draft and optimize. Then bring the thing no algorithm can generate on its own: your experience, your judgment, your taste and your wonderfully wild human spark.
The future of marketing isn’t human or machine. It’s human, amplified. And honestly? That sounds like a lot of fun.
