Problem-Framing Is the New Literacy: Q+A with Northwestern University's Jim Lecinski

How should modern marketers be thinking about artificial intelligence and machine learning? And how should marketers be developing a strategy and a plan to implement AI into their marketing toolkit?

These are the two questions The AI Marketing Canvas: A Five-Stage Road Map to Implementing Artificial Intelligence in Marketing posed to marketers in 2021. At the time, ChatGPT didn't exist and many CMOs weren't yet familiar with AI. Those who were, were largely experimenting at the edges.

"My goal was to help marketers that AI wasn't science fiction. It was a practical tool that could drive real growth," explains author Jim Lecinski, a Clinical Professor of Marketing at Northwestern University's Kellogg School of Management.

Four years is a long time in tech. It may as well be eternity in AI time. Lecinski and his co-author, Raj Venkatesan of the University of Virginia's Darden School of Business, are releasing the second edition of the book in January. Just as AI has evolved, so has the focus of the book, which is more about making AI work in the real world.

With two months to go, we spoke with Lecinski about how to go from "zero to superhero," the most foundational AI skills, how to incorporate AI into the syllabus, and much more.

The AI Marketing Canvas includes a 2x2 framework to identify the highest-value AI use cases for a business. How did you develop that?

Pattern recognition! We kept seeing the same pattern inside companies: dozens of ideas, lots of excitement, and no clear way to prioritize where to apply AI.

Yet, applications fell into a few clear grouping. We built a 2x2 that clarifies the four use cases. One axis is the benefit or value you want AI to unlock, either productivity or value creation. The other axis is the beneficiary who would directly realize those productivity or value creation gains: your marketing team or your customers.

We pressure-tested this 2x2 with companies ranging from startups to the Fortune 50, refining it until it consistently surfaced the same thing: a short list of high-impact, realistically achievable AI opportunities. It’s simple by design, because simple frameworks get used.

What are the five steps to go from “zero to superhero” with AI for marketing? How do they translate to the classroom?

The five steps form the AI Marketing Canvas maturity model: Foundation, Experimentation, Expansion, Transformation, and Monetization.

This is the best practice roadmap marketers need to follow to be successful with AI:

• Foundation: You must start by collecting the first-party training data and customer signals you need.

• Experimentation: Next, you run small, well-scoped tests with your key partners like Google and Meta using their built-in AI.

• Expansion: After that, you scale what works starting to develop your own in-house AI competency.

• Transformation: Eventually, you redesign and rewire your marketing processes and experiences.

• Monetization: Finally, for some companies, you license what you’ve built for yourself as a new revenue stream.

  
  

The steps, followed in order, are the best path to go from zero to superhero with AI for marketing.

When you think of companies doing exciting things with AI, who comes to mind?

I always look for firms doing two things: improving today’s performance and building tomorrow’s capabilities. Coca-Cola has, in many ways, been leading the way on this. Starbucks has also implemented a strong AI program using first-party data from their app. And John Deere is turning decades of agronomy data into entirely new revenue streams. These companies understand that AI isn’t a feature or fad; it’s a strategy.

What do you think is the biggest misconception professors have about AI?

Many assume AI is a shortcut tool, not a thinking tool. They see it as a way for students to avoid the work rather than a way to deepen the work.

That misunderstanding leads to defensive teaching, or even banning AI in some extreme instances, instead of adaptive teaching. The reality is that AI raises the bar for critical thinking. 

What do you think is the most foundational AI skill?

The most foundational AI skill is the ability to frame a problem. If you can articulate what you’re trying to achieve, what data matters, and how you’ll know if the solution worked, you can use AI effectively. Without that, even the most powerful model becomes a guessing engine. Problem-framing is the new literacy.

Which marketing function benefits most from AI? Where does AI still need improvement?

Right now, performance marketing and personalization see the fastest returns because the data is rich and the feedback loops are tight. The next frontier is creativity: ideation, experimentation, concept testing, and insight generation.

Where AI needs improvement is grounding. It’s still too easy for models to hallucinate, misinterpret context, or confidently return the wrong answer. So regardless of where marketers apply AI they still very much need a strong review processes.

What’s the most effective way you work AI into your syllabus?

I weave it through everything rather than isolating it in a single week. In my “AI for Marketers” course, students use AI to analyze data, generate hypotheses, segment audiences, create synthetic personas, design and test new product concepts, generate go-to-market creative assets, and assess results.

But in all cases, students are evaluated on the thinking behind their choices, not solely the output generated by a model. The rule is simple: AI can do the heavy lifting, but you own the judgment.

Many professors worry that AI will erode students’ critical thinking. How do you ensure that doesn’t happen?

Professors should design assignments that require interpretation, comparison, justification, or decision-making. AI can give students raw material, but only they can turn it into insight. When the grading rubric rewards reasoning over output, students use AI as an accelerator, not a substitute. The key is shifting the focus of assignments from “write this” to “decide this.”

How do you find students’ AI skills before they start your class? How do their skills compare with their AI literacy?

I give a brief diagnostic survey before the quarter starts. Students indicate how familiar and comfortable they are using AI to help solve a marketing problem, gather insights, and make recommendations.

What I’ve learned is that their hands-on skills are usually better than their conceptual literacy. They know how to get AI to produce text or images, but not always how to apply it in a strategic and systematic way across the entire marketing process. That gap is where the real teaching happens.

What is the biggest danger of AI illiteracy?

The biggest danger is false confidence. People who don’t understand AI’s limitations tend to trust it too much, and people who don’t understand its strengths tend to use it too little. Both create risk. AI literacy sits in the middle: knowing when to rely on the machine and when to override it.

How do you use AI on a daily basis, both professionally and personally?

Professionally, I use AI as a research partner, idea generator, writing assistant, and simulation engine/sparring partner. It helps me prepare cases, explore scenarios, and test how a marketing decision might play out.

Personally, I use it for everything from meal planning to travel recommendations to summarizing dense documents. I often describe AI as my “thinking companion.” It doesn’t replace my judgment, but it absolutely extends what I’m able to do in a day.