From AI pilots to performance: 3 things leaders need to unlock

By Jim Lecinski, Clinical Professor of Marketing

At the Kellogg Marketing Leadership Summit in May, AI in marketing was, naturally, a central topic. The CMOs I spoke with are experimenting with AI across nearly every part of marketing, from customer insight and segmentation to creative development, media optimization and measurement.
It reminded me of the early days of digital transformation: high energy, real urgency and recognition that the next stage will require more than enthusiasm and pilots.

The harder question now is how to move from experimentation to impact. With AI tools in place, promising use cases identified and pilots underway, how can marketing AI become repeatable, measurable and capable of driving real business advantage?

That is the question I addressed in my keynote talk at the summit, sharing findings from my work on The AI Marketing Canvas. In studying companies already applying AI in marketing, my coauthor Raj Venkatesan and I have found that leaders need three unlocks to scale AI successfully.

1. Clarity on use cases

The challenge is not a shortage of possible AI applications. In fact, most marketing teams now face the opposite problem: too many possible use cases, too many tools and too many disconnected experiments. The better question is strategic: Is our AI agenda primarily building efficiency, or is it building advantage and creating value?

"AI can help marketing teams do more of what they could not do before. It can expand the organization’s capacity to learn, create, personalize and serve customers."

Efficiency matters. AI can help teams move faster, reduce friction and improve productivity. But the larger opportunity is advantage: using AI to understand customers better, learn faster, personalize more effectively, test more ideas and create experiences that were not previously possible at scale.

For example, IKEA now provides a free AI interior design app called IKEA Kreativ to help shoppers who want to design their perfect room with IKEA furnishings. The hope is this novel experience will attract more shoppers, who will buy more products, more often.

If AI is described mainly as a way to do the same work faster and cheaper, employees may reasonably wonder what it means for their own value. A stronger message is that AI can help marketing teams do more of what they could not do before. It can expand the organization’s capacity to learn, create, personalize and serve customers.

2. An AI roadmap

Experimentation is necessary, but experimentation alone is not a strategy. Teams need a repeatable way to identify opportunities, prioritize the highest-value use cases, test responsibly, measure outcomes and decide what should scale.

Without a roadmap, AI activity can become fragmented. One team tests a creative tool. Another pilots a customer analytics model. Another experiments with content generation. Each effort may be useful, but the organization does not necessarily build shared learning or momentum. 

A roadmap creates discipline. It helps marketing leaders move from scattered pilots to a portfolio of AI initiatives tied to business goals. In The AI Marketing Canvas, we propose a five-stage roadmap for marketers to follow.

3. An implementation plan

AI transformation is not simply a technology rollout. It is a leadership, capability, communication and change-management challenge.
That starts with how leaders frame the work. CMOs should be explicit about which objectives they are pursuing, because that choice shapes investment, measurement and organizational commitment. 

Implementation also requires rethinking how work gets done. What should humans own? With what should AI assist? What can AI automate? What requires human judgment, taste, ethics or accountability? As teams begin working alongside AI tools and AI agents, CMOs will need clearer standards, governance, accountability and continuous capability building.

To unlock progress with AI in marketing from pilots to performance, CMOs need clear use cases, a practical roadmap and an implementation plan. Together, these can transform today’s experimentation into tomorrow’s impact.

For more AI stories and ideas from Kellogg faculty, be sure to subscribe to Leadership + AI on LinkedIn.

The views and opinions expressed in this post are those of the author, and do not necessarily reflect the position of the Kellogg School of Management or Northwestern University.