• by Sarah Mahoney , Staff Writer @mahoney_sarah, Yesterday

Ikea continues to be an AI innovation leader

For every CMO insisting their company is “all in” on AI, there’s a marketing team asking what that actually means. A few pilots? A ChatGPT subscription? A formal training plan?

So far, too many companies are still flailing. “Almost every company is playing around with AI to some degree, but most are using it in an ad-hoc way,” says Greg Kihlström, host of "The Agile Brand" podcast and an instructor for the Association of National Advertisers’ AI programs. “That offers little benefit to the organization, and doesn’t teach employees any best practices.”

Even communicating about AI internally is often a struggle. Some leaders avoid it out of fear it will spark job loss anxiety, or because they see it as a competitive secret. Others are just embarrassed by how little they’ve done. And too often, Kihlström says, teams cling to the persistent “pride of workmanship” belief that AI can’t do their job as well as they can—whether they’re creatives, media buyers, or CMOs.

Jim Lecinski, clinical associate professor of marketing at Northwestern University's Kellogg School of Management, says that mindset is holding marketing leaders back. “Too many CMOs are still thinking of AI in terms of efficiency—‘We always had someone take notes and turn them into follow- up memos, now AI will do that,’” says Lecinski, author of "The AI Marketing Canvas: A Five Stage Roadmap to Implementing Artificial Intelligence in Marketing." “That’s the proverbial faster horse, when what’s needed is a real transformation. If you had a Harrier jet, would you use it to shave five minutes off your grocery run? It’s a freaking Harrier jet.”

But those who are thinking biggest are most apt to discover transformative breakthroughs: His favorite examples include the way Ikea has used generative AI to give every shopper access to a powerful interior design tool, and John Deere, which has developed such effective proprietary AI tools that it is licensing them to others, creating a new revenue stream.

The first step? Quit faking it and learn what AI is. Many CMOs got where they are not by understanding computing, but by championing big, often Super Bowl-sized ideas. “Too many are saying, ‘I don’t need to understand this—I’ve got backend plumbing guys for that,’” Lecinski says. “But they need to know the difference between deterministic and probabilistic systems, machine learning versus generative AI, agentic models versus passive tools.”

That includes understanding the risks, from accuracy issues and legal landmines to the very real impact on team trust. Marketers can’t lead what they don’t understand, and without a core fundamental understanding, they won’t be able to develop use cases that have the most growth potential.

From toy to tool: What successful companies are doing

Those that are moving fastest toward AI fluency seem to be following a clear five-step progression, says Lecinski, who has interviewed dozens of companies about their AI adoption process. Most start small—but they start intentionally. “It’s still not too late to be early,” Lecinski says.

Train on your own data. “Use your zero- and first-party data—things these systems can’t just scrape from the web,” Lecinski says.

Lean on your vendors. Adobe, TikTok, Google, HubSpot—almost every major vendor now offers AI-powered tools, often at no extra cost.

Appoint an AI champion. Organizations further along in AI adoption typically have someone tracking every pilot and experiment. Lecinski says this person should be part project manager, part translator—fluent in both marketing and tech—and well-connected across teams.

That’s how companies are building internal AI competency, beyond what they glean from vendor and agency partnerships.

Document ROI. To secure funding and scale, AI needs to show impact—not just sales gained and time saved, but profit growth. “That enables marketers to go to a CFO with a spreadsheet and say, ‘Now we have proof. But to scale it, we’re going to need to buy some more tools.’”

Aim higher. Some companies, like John Deere, are even licensing proprietary AI tools developed in-house as new revenue streams.

The organizational bottleneck

Even when tools are available, most teams aren’t ready to use them at scale. “CMOs are focused on efficiency, not the transformative potential of personalization,” Kihlström says. “And operationally, it’s tough. To use AI across the customer journey, silos have to collapse. That’s daunting.”

AI could allow a company to create a unified email, SMS, and website experience tailored to one user in real time. But most marketing organizations aren’t built for that. Teams still operate in channels, chasing separate KPIs. “Email teams track open rates. Brand teams look at awareness. Nobody’s aligned around a single goal like lifetime customer value,” he says. “That would be a huge step forward.”

Kihlström believes federated models can help bridge the gap. Each team can retain some autonomy, but a steering group should coordinate company-wide AI integration. “You need a central function—even if it’s just a working group—to orchestrate the entire customer journey.”

The tools are here. The tech is ready. The real transformation? That’s still a leadership problem.