From Mass Media to Personalized Messaging

The same fishing pole can mean something different to each person who sees it.

For a grandfather who wants to teach his granddaughter how to fish, it is a memory-making device that helps deliver a chunk of hotdog on a hook into the water so the laughter and joy can ensue.

To a 28-year-old recreational fisherman, that same top-of-the-line model might be a status symbol to cast in front of his buddies at the annual weekend getaway they have been having since their fraternity days.

If you had to pick one marketing message to convince the grandpa and the recreational fisherman to buy this latest-and-greatest fishing pole, what would you say?

Kristian Hammond sees the time approaching — and, in some ways, already here — when that question won’t matter.

“We’ve always wanted to market products to an audience that we understand quite well,” said Hammond, director of Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program. “It would be great to understand the audience down to the level of individuals, but we couldn’t write all that marketing material for all those people — until today. Today, we can.”

Artificial intelligence (AI) and adjacent technology such as Large Language Models (LLMs) are increasingly allowing marketing professionals to pitch their products in a variety of ways to individual customers with language, images, and design most likely to catch that customer's attention.

For example, the grandfather might see a website banner with an image of an old man and young girl joyfully casting a line at sunset, while the recreational fisherman sees an in-app advertisement with an image of buddies cooking the day’s catch over a campfire and copy that reads, “It’ll give your friends fishing pole envy.”

In his role as MSAI director, Hammond has been keeping a keen eye on how AI is starting to change a variety of industries, from journalism and the law to education and insurance. His goal is to incorporate speakers and case studies from all these areas into the MSAI curriculum.

The purpose is to provide the program’s students a range of examples of how AI is impacting different professions — and how the students' AI expertise will be needed now and in the future.

"MSAI students get great opportunities to learn about and solve marketing problems thanks to existing partnerships we have with Northwestern’s Kellogg School of Management. They are a great source of projects and problems for us,” Hammond said. “Our students are interacting with them and with the master's students in the MBAi program to look at these problem spaces.”

Northwestern's MBAi Program is a joint-degree program offered between the Kellogg School of Management and the McCormick School of Engineering. Students from MSAI and MBAi work together on a client-facing problem for their capstone project.

What MSAI students are seeing when they look at AI’s influence in the marketing industry is a rapidly changing environment of increasing personalization. Like a fishing pole, the same pickup truck, computer, or firepit means something different to each person inside a potential target audience. Hammond said AI is paving the way for that increased personalization.

Truly individualized marketing isn’t inconceivable, he said.

“We could end up having language models shape the marketing for each individual,” he said. “The message you get is not a message for the general public; it's a message just for you.”

Challenges exist. AI hallucinations — when an AI system returns factually inaccurate information to a human user’s query — can create misleading marketing messages companies must ensure do not end up in advertisements.

But despite those dangers, Hammond believes marketing provides an attractive sandbox for AI experimentation.

“No technology is perfect. That's just the nature of things,” Hammond said. “AI can still be used and you can find the places where it works well. Marketing is a really good example of it. Yes, AI makes mistakes, but nobody dies. That makes marketing a fascinating area for generative AI systems and large language models in general.”