Marketing Isn't Dying. It's Being Quietly Rewired.

A friend of mine runs marketing at a mid-sized B2B company in Germany. Last quarter she told me, half-laughing, that her team had quietly subscribed to three AI tools her CIO didn't know about. "The official platform is so locked down," she said, "that we just… went around it."

I laughed too. Then I read the new Forrester report on AI in marketing, and I realized her little workaround is the whole story.

The headline nobody is saying out loud

Here's the punchline: AI will not kill marketing. But it is going to rewire how marketing actually gets done — and most companies are not ready.

Thomas Husson, a principal analyst at Forrester, puts it plainly. Marketing is still going to be about understanding your customer, defining a brand strategy, and delivering on the brand promise through experience. The discipline doesn't disappear. The plumbing does.

That distinction matters. A lot of people are arguing about whether AI replaces marketers. Wrong argument. The real question is which parts of the marketing job get automated, which parts get amplified, and which parts nobody has thought through yet.

The strange split between the top and the front line

Forrester's State of AI Survey for 2025 turned up a number I had to read twice.

55% of EU B2B marketers think AI is overhyped.

But then — 81% of frontline marketing practitioners describe themselves as enthusiastic about the technology.

Think about that for a second. The people writing the checks are skeptical. The people actually doing the work are excited. How does that happen in the same organization?

Husson's explanation is sharp: vendors over-pitch productivity gains that don't show up in production. Executives get burned, get cautious. Meanwhile the people in the trenches keep finding real, small wins — a faster draft, a cleaner summary, a decent first pass at segmentation — and they want more.

There's a name for this pattern. Roy Amara, who ran the Institute for the Future, observed that we overestimate the short-term impact of new technology and underestimate the long-term impact. Amara's Law. And it fits the AI-in-marketing story almost too neatly. The noise is loud right now. The actual reinvention of how marketing works is a five-to-seven-year arc, not a quarter.

Europe is behind. But not for the reason you think.

Here's a number that should make EU leaders uncomfortable.

Production use of generative AI: 62% in EU organizations, versus 72% elsewhere.

The easy excuse is the AI Act. Brussels made it heavy, so companies froze. That's the story you hear at conferences. Husson doesn't buy it.

The real bottleneck isn't regulation. It's culture.

Most EU companies still run marketing market-by-market. Germany has its stack. France has its stack. The Nordics do their own thing. Nobody owns a global strategy, so nobody can deploy a global AI layer on top of it. Add to that: 28% of EU B2B marketing decision-makers say they can't even identify where to apply the technology. That's not a tool problem. That's a roadmap problem.

The CMO vs. CIO problem

This one surprised me.

CMOs lead AI strategy in only 8% to 10% of organizations.

Eight to ten. In most companies, the CIO or CTO drives the deployment. They care about the right things — data governance, security, scalability. But they don't always understand what marketers actually need to do their jobs.

So what happens? The official platform becomes a fortress. Secure. Auditable. Slow. And the marketing team, which still has campaigns to ship, starts buying its own tools on a corporate card. Forrester calls this "shadow AI," and it's spreading.

The risk isn't that marketers are using AI. The risk is that leadership thinks AI is under control because the official channel is locked down — while the actual AI usage is happening in places nobody is watching.

The market is churning to fix this. 2,500 new AI solutions entered the market in a single year. 1,211 legacy tools got retired. The plumbing is being rebuilt in real time.

The 50% discount, and why it matters

Here's a detail I love. Forrester applies a 50% discount to every vendor productivity claim before it goes into their financial models.

Vendor says: this tool saves you an hour. Forrester writes down: thirty minutes.

That's not cynicism. That's experience. And it's the kind of discipline more companies should borrow before they sign contracts. The motivational piece matters too — a tool can save time and still drain morale if it makes the work feel mechanical. Productivity isn't just minutes saved. It's whether the human at the keyboard still feels like a human.

What happens to the people

This is the part that keeps B2B leaders up at night.

57% of frontline marketing decision-makers in the EU expect AI adoption to reduce headcount. 68% believe it will create new roles.

Both can be true. They usually are.

Husson makes a point here that stuck with me. The digital parts of marketing — the bits that are already data in, data out — will feel the impact first. The "analogue" parts, the pieces that depend on ethics, emotion, reading a room, understanding why a human being might trust another human being, those are harder to automate. Not impossible. Just harder.

But there's a trap hiding in the easy savings. If companies stop hiring juniors to cut costs, they cut the pipeline that produces seniors. You don't get a senior strategist by waiting six years and then hiring one. You get one by hiring twenty juniors, letting most of them grind through the messy work, and watching a few emerge. Break that chain and the talent problem shows up five years late, and all at once.

The Adobe-IBM deal, and what it signals

The conversation in technology leadership has already moved past basic generative AI. The new word is "agentic" — systems that don't just answer questions but act, with guardrails.

The recent Adobe and IBM partnership is a good example of where this is going. Adobe brings the experience platform and the customer data. IBM brings watsonx for model governance, plus Red Hat OpenShift for the infrastructure layer that lets the whole thing run across regions and comply with rules like the AI Act and the NIST AI RMF.

Five layers, end to end: data, intelligence, orchestration, governance, infrastructure. The point isn't the architecture diagram. The point is that somebody has to stitch the whole stack together before an agent can do anything trustworthy in a regulated industry.

That's where the real work is now. Not in the model. In the wiring.

So where does this leave you

If you're in marketing right now, three things are true at once.

One: the discipline you chose is still the right one. People still buy from people they trust, and brands still have to earn that trust through experience. AI changes the speed, not the mission.

Two: the way you do the work is going to change underneath you, whether you lead it or not. The companies that figure out the CMO-CIO handshake, kill shadow AI, and build a real roadmap will pull ahead. The ones that don't will keep writing checks for tools their teams quietly ignore.

Three: the human stuff — the ethics, the emotion, the anthropology of why anyone cares about your product — just became more valuable, not less. The more AI handles the mechanical work, the more the analogue parts of marketing become the differentiator.

Husson's closing thought is worth holding onto. Success in B2B will demand an approach that combines technical orchestration with a deep understanding of how human beings actually make decisions. Despite all the digital acceleration, marketing is still rooted in real relationships and professional integrity.

I keep coming back to my friend with her three unauthorized AI subscriptions. She wasn't being reckless. She was trying to do her job inside a system that wouldn't let her. That gap — between the official stack and the real work — is where every marketing organization is going to be won or lost in the next few years.

Close it, and you win. Ignore it, and you'll read about it later in someone else's case study.