Don't Wait for Permission. MOps Has to Lead the AI Era.

Most AI projects fail before they start. Not because the technology doesn't work. Because nobody asked the most important question: whose data is this, and is it actually any good?

Niels van Meerte Janse has spent 15 years at the intersection of email, martech, and data strategy, watching organizations chase omnichannel promises while organizational dysfunction kept getting in the way. He's Partnership Director at PRDCT and a jury member at the EMAS Awards. This conversation is about why Marketing Operations has to lead the AI era, not support it.

The technology flip

Five years ago, Niels' roadmap for a new client took two to three years to execute. Not because the ideas were bad. Because the infrastructure wasn't there. Data pipelines. Real-time capabilities. Technology was the bottleneck.

Then something shifted. The roadmap that used to take years started getting done in six months. And instead of celebrating, Niels found himself with an uncomfortable question: the tech was ready. The people weren't keeping up.

He calls this the technology flip. Every department is running AI pilots right now. Most of them are failing. Not because AI is bad. Because the data foundation isn't there.

The gold nobody's claiming

Marketing Operations has spent 10 to 15 years building exactly the foundation AI actually needs. Customer data. Behavioral signals. Profile enrichment. Cross-channel context. The combination of data, technology, and customer knowledge that makes AI work.

"All the things we did the last couple of years — it's gold."

The day before we recorded, Niels had been at a prospect who was proud of his new AI agent. It was connected to the CRM, finding blank fields, filling them in automatically. Data quality had "improved dramatically," he said.

When Niels asked how he was measuring data quality, the answer was completeness. Filled-in fields. Nobody had checked whether what the AI filled in was actually correct.

"Shit in, shit out. It never gets old."

Completeness isn't accuracy. And scaling that mistake is very expensive.

The KPI root

The organizational dysfunction that keeps AI from working isn't structural. It's incentive-based. If your personal target is database size or email open rate, you're not going to think cross-functionally. You'll protect your channel, claim your attribution, and make sure the numbers point your way.

Niels is honest about this: he's never seen an organization truly fix it. And the reason isn't that people don't want to. It's that the KPIs work against it.

His prescription: tie incentives to top-level business goals. Find where conflicting KPIs are creating friction between departments. They're almost always there once you look. And open that conversation with data the CMO doesn't already have.

Don't wait for permission

The boardroom isn't going to invite MOps to lead AI initiatives. Niels is clear about that.

"Don't wait for permission. Don't wait for permission because you're not going to get it. Show them."

Run experiments. Maybe without asking first. Show what's possible. Go with results, not proposals.

The risk isn't that MOps tries to lead and fails. The risk is that MOps waits. Niels saw a brand's board shut down every AI initiative for a full year after one badly run pilot in the wrong department. One wrong experiment. One year of setback. That's what happens when MOps doesn't claim the position first.

The bottom line

The best AI projects don't start with the technology. They start with the data. And the team that has been living inside that data for a decade is Marketing Operations.

Most organizations don't realize that. If MOps doesn't go in and claim that ownership, show what they know, share the insights, make the case, someone else will try. And they'll make a mess of it.

The mandate isn't coming from above. Go claim it.

Find Niels on LinkedIn: linkedin.com/in/nielsvanmeertejanse