Generative AI in advertising: Is it worth it?

Since the release of ChatGPT in 2022, marketers have used generative AI, powered by large language models, to create text, audio and images in their ad campaigns. 

Fast forward to 2024 and the use of generative AI in marketing and sales has doubled, with 65% of people now saying their company is regularly using generative AI. 

Despite its widespread adoption, consumers are unsure about the use of AI in advertising. 45% of people say they think brands should use it, 36% of consumers say they don’t think brands should use it, and the rest are undecided. 

In this post, I’ll take a look at the use of generative AI in advertising and unpack the questions: How useful is it? Do consumers want it in their ads? Is it coming for our jobs? And do the pros outweigh the cons?

Since its release, notable brands have been experimenting with AI. Let’s take a look at two great examples of generative AI in advertising. 

1. Heinz: A.I. ketchup 

One of the reasons consumers may feel uncomfortable with AI in ads is because of a phenomenon known as uncanny valley. The uncanny valley describes the uneasy, anxious feeling we get from celebrity faces with too much plastic surgery — they make us uncomfortable because they don’t quite “read” as human. 

If you type in “AI ads,” into YouTube, you’ll find several creepy spoof AI ads in which the tech has tried to replicate the human face. 

Take a look at one user’s comment under an AI-generated beer commercial: 

Rather than using AI to imitate the human, Heinz and ad agency Rethink put generative AI front and center of their creative concept in an ad for the brand’s ketchup:

The creatives behind the ad tell the viewer how they used AI to create images of ketchup with the AI image-generator Dall-E 2. The ad flicks through several different images of ketchup based on a prompt: “ketchup.” Viewers see a line up of pictures including ketchup street art, ketchup impressionism, and ketchup stained glass. Each bottle shows up with the famous Heinz branding. 

The tagline reads, “Even AI knows that ketchup is Heinz. No matter how we asked for it.”

Rethink says: 

“Heinz Ketchup is a cultural icon — it’s not just a ketchup, it is the ketchup. This summer, text-to-image AI generators took the Internet by storm. 

People were creating weird and wild images from written prompts like “Napoleon riding a Harley Davidson,” or “R2D2 getting baptized. 

We saw an opportunity to participate in an exciting cultural conversation while proving to a new, younger audience of tech and art fans that Heinz is the definitive ketchup.”

A masterclass in brand recognition and brand equity, the playful dig suggests that, “even the robots,” know Heinz is the best brand of ketchup. 

Despite consumer’s ambivalence towards AI, the ad went viral and improved social engagement by over 1495%. It also went on to win the 2023 Clio Gold Award for Product/Service.

Why does the ad hit the right notes with consumers? 

One of the reasons why the ad might be so effective is because consumers are in on the joke rather than the butt of it. Ad execs aren’t trying to pass their content off as the work of human creativity — they’re not trying to replicate the human face or voice or trying to get the ad to pass as someone else’s art. They’re not trying to fool the consumer, they’re bringing them in on the joke. 

This playful transparency could be why the ad was so popular. Speaking to Forbes, AI expert Michelson says: 

“When it comes to AI, trust is something that must be earned, and part of that involves transparency and understanding.”  

And the stats back up the idea that transparency and trust are essential to the effective use of AI in advertising. One study found that ads that were honest about using AI saw a 47% jump in ad appeal.

2. Nike: Never Done Evolving

 In 2023, Nike partnered with AKQA studios to create the “Never Done Evolving,” campaign. The ad shows how tennis professional Serena Williams has refined her technique over time. 

Featuring a match between an AI-generated younger Serena, playing her first Grand Slam at the 1999 US Open, versus an older AI-generated Serena at the 2017 Australian Open.

Here's AKQA shared, “Machine learning was able to model each era’s playing style: decision making, shot selection, reactivity, recovery and agility based on archival footage. Building on the vid2player technique developed by Stanford University, Nike were able to bring the models of Serena to life by re-rendering the players from each generation into an entirely new scene and have them appear to be playing and responding to each other.”

They add, “The result was seeing both Serena’s play each other for 130,000 games and 5,000 matches, enough tennis matches to stream for an entire year if played back-to-back.”

Check out the AI model working behind the scenes: 

And the finished ad:

What’s so good about this ad?

Nike shows us how AI can give consumers the chance to experience something they love in a new way. Without AI analytics and generation, tennis fans would never get to see a younger Serena go head-to-head with 2017 self. 

Experiment with a new kind of creative

You can use AI to experiment with different content and create a new viewing experience. Generative AI can also give you new ways of looking at the creative process.

Researchers at The University of Leeds say:

“GenAI can produce what seems like diverse and original outputs. It can create content through capturing nuances in language, based on patterns which we may not have seen before in the data they were trained on.”

Coca-Cola was the first big-name brand to use generative AI in a commercial. The brand partnered with creative agency Bain and Company, who used OpenAI’s DALL-E2 text-to-image AI model and ChatGPT, to create the ad “Masterpiece.” In the ad, the creative team playfully uses AI to create a more immersive storytelling experience. 

In the commercial, famous pieces of art, personified through AI, share a bottle of Coke in a gallery. The ad is almost a satirical nod to how we can use AI to play with new ways of “bringing to life,” human being’s artistic creations  — in both the commercial and non-commercial space. 

While many creatives worry that AI will replace them with cheap, lackluster ad content, Coke’s ad shows you how you can use AI to support your creative vision and create a new viewing experience. 

Hyper personalize your ads

One of the best uses of AI is personalization. 

From first names to buying behavior, generative AI can bring together several sources of data on consumer behavior and preferences to make personalized ads tailored to your audience. AI can predict what consumers want and switch out elements of a commercial, such as the music to suit the viewer.

The Jen A.I. campaign from Virgin Voyages, VMLY&R and AI company Deeplocal is a great example of personalization through AI. 

The ad features an AI-generated Jennifer Lopez inviting people on a Virgin Voyages cruise. The ad switches between “Jen A.I.,” and the “creators” behind her — who do comedic impersonations of the singer. 

The ad shares the tagline, “Invite your crew to voyage,” inviting watchers to create their own personalized video invites on the Virgin Voyages site.  

Repurpose your content 

60% of marketers reuse their content two to five times. But without AI, it can take a lot of trial and error (and lengthy analysis) to work out what does and doesn’t work for your audience. 

Mike Kaput, Chief Content Officer at Marketing AI Institute, notes the usefulness of AI for content repurposing:

“You can take a single ad, give it to an AI tool, and it will spin that ad off into a number of different variations. Those variations could include different ad sizes and formats to adhere to different platforms. Or they may include different designs and creative based on all the various campaign ideas you and your team have come up with.”

Hand the boring work over to AI 

One of the best ways you can use generative AI is to automate all the boring necessities of your day job — giving you more space and time to be creative. 

McKinsey reports that gen AI could add up to $4.4 trillion in yearly global productivity — with marketing and sales getting 75% of this value. 

Generative AI could automate tasks that take up around 60-70% of employees time. It can also improve productivity by more than 40% in small to medium-sized businesses. 

Use AI to create project plans, meeting schedules, and reports — freeing up your time.

Turn consumer insights data into testable concepts 

“Ten years ago, people in brand manager roles would wait for data. Then they would spend two hours before the meeting getting it ready. The launch of agile research and the ability to test assumptions quickly means that you can do this often, but you still have to deal with data tables and charts — and putting them into your PowerPoint campaign. 

Gen AI is going to help you do that. It’s more efficient. It’s going to have you put more of your effort and time into thinking about the work after the work than the work. And that’s fantastic.”

- Ryan Barry, President of Zappi

Zappi brings together different AI agents to help you gather consumer insights data and turn this data into new ideas and creative assets. 

You can use our AI-based platform to guide the creation of ad or innovation concepts that match the needs, interests and preferences of your audience.

By combining consumer insights data, Zappi understands what your customer needs and wants. It then uses gen AI to turn these insights into recommendations on what to do with your data. 

Zappi relies on several different AI agents to help you do this. The analytical AI agent makes recommendations based on consumer data and shares the “thinking” behind its ideas so you can see why it’s made the creative decisions it has.  The facilitator agent organizes the process and may ask the design agent to create creative assets like images that reflect the research. You can then share these assets with your audiences for real-time feedback. 

Because of AI’s speed and efficiency — the discovery and research process that once took days or weeks can take minutes. 

To learn more about these features, reach out to us. 

Rather than AI being more good than bad or bad than good, it’s an interwoven maze of both the good and the bad. It creates human problems as much as it solves them. 

Elena Osadchaya and her research team (quoting researchers Mick and Fournier and M.W. Lewis) say that technologies are often seen as inherently paradoxical. This means, in the words of M.W. Lewis, that they contain, “Contradictory yet interrelated elements—elements that seem logical in isolation but absurd and irrational when appearing simultaneously.” 

But some people think the pros outweigh the cons. 

Jessica Apotheker calls generative AI a “productivity revolution.” She says AI can free up creatives to focus on right-brain tasks (those that need creativity) by up to 40%. In the words of Zappi’s President Ryan Barry, AI brings efficiency and automation to your workflows — giving you more time for creative storytelling. 

On the other side, you need to make sure the platforms you use are reliable. You also need data checks in place. 

That’s because one of the biggest risks of using generative AI is its inclination to make mistakes or get things wrong. 

In a survey by McKinsey, inaccuracy was ranked the biggest risk of using generative AI, narrowly followed by issues around intellectual property infringement:

Model hallucinations are one of the main reasons for inaccurate content. When a large language model hallucinates, it sees patterns that don’t exist and creates false or incomprehensive content in response. 

Take Google’s Gemini, which hallucinated and told users that the James Webb Space Telescope had taken the world’s first photos of a planet outside our solar system. 

Researchers at the University of Leeds stress that people must remember these three things when using generative AI: 

1. Although Gen AI models appear to understand the content they use and generate, they don’t actually understand it

2. The data that Gen AI models use for training have lots of inaccuracies and biases in them 

3. Gen AI can easily create fake news, misinformation, and ‘deep fakes’

AI is only as good as its data. 

And with so many marketers relying on AI to bring together and analyze several datasets, how much time would they need to double check the source of their third-party data to see if it’s reliable and correct? With AI pulling from millions of datasets — is it even possible to accurately trace the source of each dataset? 

Jarred Cinman says another one of the biggest issues surrounding AI is potential copyright infringement. 

He says that most agencies are asking, “Who owns the copyright on AI-generated content? If I make an image on Mid Journey or write some copy on ChatGPT, do I own that or does the tech provider? Does the generated content infringe the copyright of the content owners on which the model was trained?”

Right now, he says, there are no clear answers to these questions because the tech and its regulation are still in its infancy. 

And when it comes to creative content — is AI even that good? 

Many creatives believe that AI will never be able to make ads as good as human beings can. And they aren’t the only ones. No matter the format, many consumers feel like there’s still a grating inauthenticity and blandness to AI content. 

Here’s one Reddit user, in a YouTube creator’s community, talking about how most AI-generated videos are “unwatchable”: 

Generative AI’s talent for making “unconsumable” content was playfully poked at by Wunderman Thompson who took a highlighter to the awfulness of AI content and copy in this KitKat campaign:

Wunderman Thompson’s chief creative officer João Braga says:  

“AI is revolutionizing our industry and beyond. But the reality is not every brand has something relevant (or fun) to say in that space. KitKat has breaks, and AI gives us more of those. So, we thought we’d have a crack at it ourselves and poke some fun at AI — while we can.”

While most marketers would never use AI’s creative output without heavily refining it, editing AI content and copy may take more time than just writing the taglines ourselves. 

And when it comes to metrics — does AI content even perform better? 

A study from professor Jochen Hartmann found that online ads performed better when the images were AI created — but only if they didn’t look like AI,. wWhile ads failed to perform better when the caption was written by AI. 

While AI can’t replace creatives, the magic behind the humor in Virgin Voyage’s "Jen A.I.", the storytelling aspect of Coke’s "Masterpiece" and the playfulness in KitKat’s campaign may be where it really shines in its efficiency and ability to take on the boring, but essential elements of the creative process.

Curious about how Zappi’s AI-powered platform can give you more time for creative storytelling? Talk to our team to find out more or see if you qualify to be a Beta user of Zappi's AI powered concept testing solution.