Jacob Tinyassistant professor of marketing at the Kellogg School, sees huge potential for genetic AI to help with new forms of market research, shake up brainstorming sessions and personalize advertising on a scale that simply wouldn’t have been possible before.
“One of the biggest advantages of genetic AI is its ability to provide marketers with negligible insights,” says Teeny. “ChatGPT is good for breaking the functional fixity where you only see something from a certain angle.”
Here, Teeny offers three ways marketing teams can work with generative AI today, balancing unconventional insights generated by algorithms with the human judgment always needed to evaluate ideas and keep brands on track road.
Know your customers
Marketers are increasingly using AI models in a variety of ways to help with customer research.
First, AI can help teams generate ideas about customer segments that might use a product or service, and then it can help create a description of the relevant characteristics—a persona—for each segment.
“When you’re working in marketing and you’re trying to launch a new product, you come up with all these names for different targets, like ‘urban peacock’ or ‘hard-working Jane,'” says Teeny. “Artificial intelligence is very good at profiling consumers who might be interested in what you’re selling.”
For example, online travel aggregator Kayak recently used genetic artificial intelligence to develop a better sense of who might be most interested in its service. In partnership with Supernatural, a creative company that combines human expertise with insights from artificial intelligence and machine learning, Kayak’s marketers have created a unforgettable advertising campaign it appeals to men in their 20s and 30s who appreciate humor in their ads and are bothered by political polarization. The ads mocked a character who is a “kayak denier,” insisting that “kayaking isn’t real.”
Genetic AI can also be used to enhance the design of consumer research, drawing on principles from market research to create questions that can produce more insightful results.
“A lot of research has been done so poorly,” says Teeny. “GPT could ensure that you design your research in a way that is fluid for respondents and gives you data that is valuable, interpretable and actionable that you might not have come up with on your own.”
Another promising avenue for marketers, Teeny says, is using artificial intelligence to collect consumer data through chatbots. Consumer information gathered from these automated interactions can help offset data loss in other areas as tech companies like Apple and Google stop using the once-ubiquitous third-party cookies that tracked users as they browsed the web.
“As a research tool, GPT is getting more and more capable to identify your personality or certain customer segments just by having a conversation with you,” Teeny says. “Any of these genetic AIs could be a good source to learn about a consumer in a very naturalistic way.”
He cautions, however, not to take the information gathered by AI at face value. “It’s up to the dealer to determine if that’s valuable evidence that could be verified.”
Boost your creativity
At its most basic level, genetic AI can free up a creative team to focus on larger projects while meeting routine daily content needs. AI chatbots can quickly write usable copy for social media posts, a time-consuming and seemingly endless task. Teeny also expects to see genetic AI used to develop brand experience guidelines, the playbooks brands use to set guardrails and expectations.
“You can train a GPT to understand your brand personality: you’re fun, you’re outgoing, you don’t engage in political discussions, you really like cats,” Teeny says as an example. “Within seconds, it can pull up tons of potential social media posts.”
Genetic AI can also create imagery, although brands must be careful that this work does not cause consumer backlash. Lego had to back down after AI images on her website drew criticism from her fans.
What might work better right now, says Teeny, is to use genetic AI in the pre-production process, where it can quickly create fully-planned scripts and ideas to pitch ideas for a brand or ad, rather than the half-finished ones. sketches commonly used in early planning meetings.
“With genetic AI, you could repeat a lot of examples that look very professional,” says Teeny. “For agencies, that could be a big boon.”
For now, however, some brands have banned agencies from using AI — both because of concerns about bias and because they may not believe work generated quickly by a chatbot is worth paying high fees. However, regardless of whether marketers actually use AI-generated materials in their campaigns, Teeny believes they should use the tool to enhance creativity and brand strategy in even more transformative ways.
“Because these AI models use a probabilistic basis, you can get any kind of association with an idea that you’re talking about,” says Teeny. “In psychology, we call this the nomological net—how different concepts relate to each other. In your mind, there are certain limits to these associations. Generative AI extends that.”
For example, in a recent study comparing Wharton MBA students to OpenAI’s GPT-4 innovative product ideascustomers in the target market stated that they were more likely to purchase the products recommended by GPT-4.
That said, people should still be the tasters to determine if a creative idea is also useful and relevant. After all, it’s not just the craziest idea that wins the market — it’s the one that fills a real consumer need.
“This is a tool that helps you refine your ideas or come up with new ideas,” says Teeny. “You still need someone to look at it and say, ‘This is good. This is not good.'”
Importantly, Teeny does not want to see the spread of artificial intelligence further erode the use of strategy in marketing. “Your brand strategist needs to make sure a cool idea aligns with your brand personality and campaign goals.”
Personalize your offer
Generative AI has set the stage for a new level of personalization, from ads to social media content. Teeny’s own research found that AI can be used to effectively tailor advertising based on a consumer’s personality traits, such as extroversion.
For example, AI would emphasize bright colors and social relevance of a product advertised to extroverts, while emphasizing more subdued colors and intrapersonal advantages when advertising the same product to introverts.
“It comes up with messages that, on average, are more persuasive than non-personalized ads,” he says. And while previously it was expensive and time-consuming to tailor a message to an individual consumer, artificial intelligence can complete the task in seconds.
AI is also capable of personalizing content based on data from customer transactions. Carvana, a used car retailer, used genetic artificial intelligence to create more than 1.3 million unique videos for its buyers, “celebrating their car buying journeys”. Each video included information such as the date a customer purchased their car, the make and model, and where it was purchased.
Generative AI can be useful in a globalized context for designing messages for all kinds of consumers, whether that means translating a message into an additional language or suggesting cultural references for customers in unfamiliar regions. But there are limits to how far marketers may want to go with personalized appeals.
“For some of the big brands, the ones that have their own personality, you don’t want to personalize too much,” Teeny says. For a brand like Harley-Davidson, whose heavyweight motorcycles are often bought by men over 35, trying to create AI-tailored content for 18-year-old women could confuse the brand’s identity and alienate traditional customers.
On the other hand, Teeny explains, Coca-Cola has a much broader brand identity focused on happiness, so it might be easier to use genetic AI to personalize what happiness means to many different types of customers.
However marketing teams decide to approach personalization, their success will depend on how authentically AI is able to represent their brand.
“A big part of the success of personalization depends on trust,” says Teeny. “We live in a post-truth world and when everything is so personalized, you can start to wonder, is this reality or is this just what I’m being served?”
So when genetic AI is used for marketing, he reiterates, the human touch will continue to matter.
“You shouldn’t rely solely on artificial intelligence,” says Teeny. “We have to look at it as a tool, just like we look at Google searches or Wikipedia. Artificial intelligence can help enable or accelerate our own ideas and designs.”
