From AI Users to AI Builders: The Next Phase of AI in Marketing

Most marketing teams have already crossed the first finish line.

They write prompts. They run drafts through an LLM. They've found their personal productivity gains and they feel good about it.

But the next race has already started. And the uncomfortable part? It's not about tools anymore. It's about people.

Dan Slagen would know. He runs marketing and AI transformation at Zapier, and he's been living inside this shift for the past year.

"No tool, no platform, no LLM is coming to save you," he says. Read that twice. The guy who sells automation for a living is telling you the software won't save you.

So what will?

When the bar jumped

When Slagen arrived at Zapier, the marketing team already had solid instincts from the early waves of AI adoption. They were comfortable. Maybe too comfortable.

Then late 2025 happened. Coding agents landed. Cursor. Claude Code. Codex. Suddenly non-technical marketers were expected to do things that, six months earlier, only engineers did.

The bar to "be good at AI" didn't rise. It teleported to a different floor.

Think about what that feels like. You just got the hang of prompting. You're feeling sharp. Then somebody moves the goalpost to a stadium you haven't even bought a ticket to.

That's the moment Slagen walked into. And he decided to handle it in a way most leaders wouldn't.

Make it real, even if it hurts

Here's what most leaders get wrong about change: they try to make it gentle.

Slagen did the opposite. He brought in an external AI power user to demo, in front of the whole team, what was actually possible right now.

The reaction? Stunned. Confused. Some people were upset.

"I'm not a fan of leading by fear," he said. "But every once in a while you need people to understand how big this moment is."

Wow. That's a strong line. And it's honest.

Because here's the thing: false comfort kills teams. If your people don't feel the ground shaking, they won't move their feet.

Three weeks. No excuses.

After the shock came the support.

Every single person on the marketing team got three weeks to learn a coding agent of their choice. Any tool. Any pace. Slagen designated five internal power users as coaches, available around the clock.

Then he said something I can't stop thinking about:

"I'll never know how many times you met with the coaches, what questions you asked, or how bad your first build was. My only ask: in three weeks, you need to be ready."

No surveillance. No shame. Just a deadline and a hand on your back.

That's trust used as a management tool. And it works because the deal is symmetric: I give you autonomy, you give me a result.

Seventy videos

Then came the build-a-thon. Two weeks. Every marketer had to ship something.

Results ranged from internal dashboards to fully functional products. Each person submitted a short video walkthrough.

Slagen watched all of them. Roughly seventy.

He wrote a detailed, personalized response to every single one.

Think about that for a second. Seventy videos. Seventy handwritten replies. From a CMO who runs AI transformation at one of the most recognized automation companies on the planet.

Strong. That's how you make people feel seen. And people who feel seen will run through walls for you.

The team ended up sorting themselves into three skill tiers. Not because Slagen ranked them, but because they self-placed. Then he could give each tier its own next challenge, so that everyone, regardless of starting level, had something they could win.

The insight is simple but rare: people don't need the same challenge. They need a challenge they can actually meet.

What this means for you

Slagen will lay all of this out on the MAICON 2026 main stage, in a session built for marketing leaders who are already past the "we use AI" phase and are now staring at the harder questions.

How do I bring my team along? How do I build a culture that doesn't just use AI but creates with it? How do I lead when there's no finished playbook?

Here's what I take away from Zapier's story:

The companies that win the next phase of AI won't be the ones with the best tools. Tools are table stakes. They'll be the ones whose leaders were brave enough to scare their teams, generous enough to coach them, and patient enough to watch seventy videos.

The shift from using AI to building with AI is a people problem dressed up as a technology problem.

Solve the people part. The technology part already works.

---

Slagen presents at MAICON 2026, in a lineup focused on real-world AI strategy, implementation, and leadership.