Skip to main content

Can AI Build a Better Marketer?

Yes, marketers can use AI to work better, smarter and faster — but only with the right balance of oversight and human intervention.
artificial intelligence

Every once in a while, the public gets a rare glimpse into how the world’s largest marketers plan to shake up their categories with the help of artificial intelligence (AI).

At the 2024 Consumer Electronics Show in January, Walmart unveiled its plan to sell an AI-powered replenishment service that is trained to know when to restock frequently purchased household items and can place the fulfillment order without human assistance. Last fall, Procter & Gamble said it would use AI to develop smart fragrance design software that can identify and help eliminate foul odors from laundry detergents, while Coca-Cola teased a limited-edition Y3000 Zero Sugar drink that illustrated the power of AI to propel brands into the next millennium. 

These news developments barely scratch the surface of AI’s sweeping impact on the industry. ChatGPT’s Taylor Swift-level popularity has created an avalanche of use cases for generative AI models across the entire value chain, with its groundbreaking technology now touching virtually every aspect of marketing. 

And yet even this is just the tip of the iceberg, says Dan Temby, senior vice president of technology analytics at agency DAC in Toronto, Canada. “Every day I am talking to marketers and telling them to turn to ChatGPT for assistance with something they’re working on, and they’re still surprised it can be used effectively in that situation.”

More Than a Wordsmith Machine

Marketers tend to think of the natural language processing models that power programs like ChatGPT as being limited to written content (e.g., drafts of emails, ideas for copywriting, headlines, etc.). But while they are not built for mathematical calculations, per se, they can answer complex numerical questions and help solve a wide range of analytical problems. 

As an example, Temby recalls working on a recent pitch for a consumer brand looking to reach $50 million in e-commerce revenue by next year. “We were trying to show that there are more levers we could pull than simply buying media, and how those things working in concert would drive ambitious goals,” he says. “Normally we’d run a VBA [Visual Basic for Applications] script, but now I tell my colleagues to go to ChatGPT and explain the problem to the model just like you explained it to me.”

Temby says that brand marketers are now discovering what their analytics teams have known for some time: With the right guidance, generative AI models can accrue benefits that go well beyond speed or efficiency. 

“Marketers still have to be good at marketing in order to use AI effectively, in the same way that software developers still have to be good engineers,” Temby says.

Marketing Automation: A Roadmap for AI Integration

One area that may provide a roadmap for effective AI integration lies within the field of marketing automation. Currently, more than three in four (76%) companies employ some type of marketing automation software, according to HubSpot. Many of these products are now built with an AI foundation of predictive analytics and content creation capabilities that help brands deliver personalized messages at scale.

The automation space includes a growing number of platforms, like Klaviyo, that position themselves as a more flexible alternative to legacy SaaS companies, thanks to their ability to aggregate the marketer’s historical first-party data into a centralized repository for building consistent campaigns across owned channels such as email, SMS, mobile apps and web push notifications.

Many of Klaviyo’s AI tools help create the building blocks of campaigns, but require input from the marketer in order to run the model. For example, with Email AI, marketers provide parameters for the kind of promotional email they want to send (e.g., 20% off our new blue jogging shorts), and the model returns a template with suggested blocks of text, fonts, colors and image placeholders.

“Many of our AI tools remove mundane or repetitive operational tasks from the user, which allows marketers to devote more time to the strategic aspects of the campaign,” says Jessica Schanzer, product marketing lead at Klaviyo.

Klaviyo envisions a future of “autonomous AI” in which the models not only make recommendations, but can also suggest next phases of campaigns and carry out the work all on their own. “This next stage of AI innovation is not just about predicting or creating content, but offering a self-optimizing marketing platform that runs on an autopilot in the background,” Schanzer explains. “You turn it on and you see a pop-up element that says, ‘Hey, we noticed this campaign didn’t perform as well as the previous one. Would you like us to make a change by segmenting to this new audience?’ It’s about identifying new opportunities.”

Keeping a Human Eye on AI

As marketers delegate more tasks to AI, the need for transparency and human oversight increases. This is especially true when publishing written ideas or design concepts that may raise issues of copyright infringement. “There’s no shame in using ChatGPT for creating content, as long as you put rigor into the outputs and disclose [the sources and methods of] your work,” says Dominik Heinrich, a former IPG marketing executive and assistant professor of AI Design at the Pratt Institute in New York.

Heinrich says that AI is often framed as a false choice between productivity and creativity, and he reminds marketers that AI is designed to augment human ingenuity, not replace it. For example, he recently consulted with a global consumer products marketer at his Creative AI Academy on ideas for a liquid soap dispenser prototype that would improve the usual slow and messy process of transferring soap from the refill bottle to an original dispenser. “Gen AI helped me create and simulate hundreds of ideas, but none of these were working. They were either not known, or had already been tested and had no market adoption.”

That’s when his passion for space exploration produced a lightbulb moment: How would the refill work in a zero-gravity environment, he wondered? This “simple” thought process led to an idea for a vacuum-operated refill pack. “You basically press the refill pack together, screw on the soap dispenser, and by releasing air in the soap refill, it will then fill up the soap dispenser,” he explains. “Airflow can be adjusted to bottle volumes to make this perfect, of course.”

A soap refill dispenser that works quickly with minimal handling and no cleanup or waste? Huh. Now why didn’t AI think of that?

This ad will auto-close in 10 seconds