Companies are tantalized by the many new possibilities artificial intelligence offers, particularly due to recent advances in generative AI. Yet a case study of several brands’ early AI experiments reveals that so far, the technology has produced mixed results for businesses, providing a boon for some while posing a reputation risk for others.
What made the difference in whether a company’s use of AI succeeded or failed was exactly how and why the technology was employed—with the “why” being especially important, say Harvard Business School professors Julian De Freitas and Elie Ofek. Those businesses that were overly focused on using AI to capture short-term economic value without considering the sensitivities of their customer base—and how this would affect attitudes toward their brands—often ended up with failed AI initiatives.
“Companies often jump on the AI bandwagon pretty quickly,” says De Freitas, an assistant professor in the Marketing Unit. “One of the reasons is, there’s a lot of pressure from shareholders and other executives at these companies to have an AI strategy. But brands are fragile, because their equity lies in the eyes of consumers. So, one wrong move could destroy brand equity that you’ve been building up for years.”
In the past two years, AI has gained significant traction in customer-facing settings, with companies eagerly adopting the technology to help with everything from sending tailored marketing messages to automating communications with customers, all while hoping to lower costs and speed up workflows. Indeed, the market for AI grew to more than $184 billion in 2024, a $50 billion increase from the year before—and many expect the market to continue to flourish, according to the case study “AI and Brand Management: Promises and Perils,” written by De Freitas and Ofek.
Yet with many companies experiencing public pushback for their use of AI, business leaders should take a more cautious approach to ensure the technology creates value instead of liabilities, the researchers advise.
When AI backfires on a brand
The case details 13 companies that failed to use AI effectively. Here are some high-profile examples of companies apologizing for and rethinking their AI use after receiving criticism:
In February 2024, users found that Gemini AI , the AI tool, would create historically incorrect images, including depicting people of color dressed as Nazi soldiers, leading to public outrage. Months later, the company’s Olympics ad showed a father helping his daughter use AI to write a letter to an Olympic athlete, prompting criticism that the company was promoting AI to replace personal expression and human creativity.
While Google has aimed to be an AI leader, these displays showed the company fundamentally misunderstood some users' cultural sensitivities toward AI, the researchers say.
"If you're trying to touch on an issue that deals with authenticity in humans and in human creativity or in human connectivity, that's when using AI, which is considered emotionless, can often lead you astray and become a problem," says Ofek, who is the Malcolm P. McNair Professor of Marketing.
Levi Strauss
partnered with startup Lalaland.ai to create AI-generated models from diverse backgrounds to supplement human models. However, the plan backfired when critics saw the move as a cheap shortcut, instead of a real commitment to hiring human models from diverse backgrounds. People questioned whether the company was using AI to replace human jobs in the modeling industry.
"Levi’s intended to enhance diversity in their marketing but ended up undermining their perceived authenticity because AI was viewed as replacing the jobs of real, diverse human models,” De Freitas says. “The technology appeared to be primarily a cost-cutting measure, contradicting the company's stated diversity goals."
Lego
angered fans after they discovered AI-generated images on the company's website. The subtle flaws typical of AI art—characters with misshapen hands or wearing the wrong hair accessories—prompted a fierce reaction. Many said the practice went against LEGO’s brand identity, which was built on human creativity and craftsmanship.
“Yes, it was a very small change. The hand looks a little different and the headband, too,” Ofek says. “But once you suddenly discover that, you're like, have they been cheating me elsewhere? Are they not authentic with me in other places? It raises a bar of suspicion.”
Intuit
decided to add an AI-powered assistant to TurboTax. However, the tool provided incorrect tax advice to users, as evaluated by tax experts, a critical misstep for a brand that prides itself on providing highly accurate tax services.
“Your existing brand and the vertical it sits in are both important. In the case of Intuit, your entire brand promise is that you won't make mistakes on tax returns," Ofek says. “So brands should ask themselves: What could go wrong specifically with respect to their core promise?"
When AI boosts brands
Companies that have successfully incorporated AI into their marketing strategies have been more calculated about harnessing the technology, rather than simply jumping on a popular trend or using it to cut costs, the researchers say. Some success stories include:
Orange
The French telecom company cleverly used AI facial editing technology in a 2023 advertisement that ran during the Women's World Cup. The ad first depicted what looked like the French men’s soccer team members showing off their skills before revealing that the players were actually the women’s national team, whose faces had been digitally swapped using AI.
“Orange succeeded because they understood what people were saying about AI in the wider culture, including that AI could be used to mislead people. They demonstrated that this same ability of AI to alter appearances could also be used for good,” De Freitas says. “They used the technology to make a point about gender bias in sports, not as a way to cut costs. And the timing during the Women's World Cup gave them an interested audience.”
Most importantly, the ad wasn’t really about AI. “AI is not the hero of the story,” he says. “The hero of the story is the women's soccer team."
Nike
used AI and 3D printing to create custom sneakers for 13 professional athletes, including French basketball player Victor Wembanyama and American sprinter Sha'Carri Richardson, who competed in the 2024 Olympics in Paris.
“Nike played its hand correctly by leveraging the Olympians it already supports,” explains Ofek. “The bespoke sneakers reflect the unique journey of each Olympian, and instantiate notions of using AI to help push boundaries, personal excellence, and innovation—key themes associated with the Nike brand.”
IKEA
Kreativ, the design app, has used AI to scan customers’ rooms and show virtual renderings of the company’s furniture in their homes. This technology aimed to tackle one of retail’s biggest challenges: helping shoppers see how products will look in their own spaces before buying—including accounting for size constraints and even the room’s lighting. The app also enabled the company to provide personalized recommendations based on people’s preferences and past shopping behaviors.
Some of the best uses of AI in marketing involve augmenting human creativity rather than replacing it—in many cases, accomplishing something that humans alone cannot, the researchers say. “Here, the benefits of AI to the customer journey are undeniable,” notes De Freitas. “And the way AI is used to provide a personalized solution communicates that the brand cares about meeting the customer where they are and solving their specific needs.”
How to use AI without hurting your brand
De Freitas emphasizes that brands can be brittle and customer trust is often fleeting. “You have to manage brands very delicately because they’re meaning-based assets, whose value is partly determined by subjective perceptions,” he says. He and Ofek outlined five tips for effectively using AI without falling into common traps:
Think long-term.
Ask whether AI will truly strengthen your brand over time or just save you time and money in the short run.
Read the room.
Understand how customers currently feel about AI and how that might affect how they perceive your use of it.
Solve real problems.
Identify how AI will make specific customer experiences better in line with your brand image.
Keep humans in the picture.
Avoid making AI the main focus of customer interactions, which can feel threatening rather than helpful.
Anticipate problems.
Develop a communication plan for potential AI mistakes before launching a new initiative, particularly when the technology has the potential to affect your brand’s core promise.
The two hope the case will help managers think more deeply about the potential consequences of using AI and whether they’re comfortable taking those risks before jumping in. “There are real risks to mixing AI and brand management—brand risks, customer relationship risks, liability risks—and the case reveals the predominant reasons for why these risks may be realized,” De Freitas explains. “Just because everyone is leaping into the AI waters, doesn’t mean there aren’t sharks lurking”.
Image credit: Ariana Cohen-Halberstam with assets from Unsplash and AdobeStock.
Have feedback for us?
AI and Brand Management: Promises and Perils
De Freitas, Julian, and Elie Ofek. "AI and Brand Management: Promises and Perils." Harvard Business School Case 525-021, October 2024. (Revised February 2025.)
