Big Tech’s AI push is going to devour ad agencies. Or is it?

Like many creative types, Riley Shine has reservations about generative AI. That didn’t stop the BlueChew brand creative director from exclusively using the technology to make a nationwide TV commercial for his company’s erectile dysfunction medication. 

“My whole internal drive is like, ‘Let’s shoot real commercials. Let’s do the real thing,” Shine told Sherwood News. “And everybody’s like, ‘Let’s do this AI stuff.’”

In the 30-second video clip, Michelangelo’s David, newly confident after taking his BlueChew ED pills, has museumgoers and the subjects of other famous artworks, like Venus and Mona Lisa, swooning. It’s not a masterpiece, but it’s not bad either — especially for something that took Shine and two others only about 48 hours and the cost of the VEO 3 generation credits to make.

That said, he doesn’t think it could have been done without creative humans, and the tech has lots of limitations, so it couldn’t be used in every situation.

“You work within the boundaries of the tool — ’cause that’s all I think it really is: just another tool. It’s a different Photoshop,” he said. “You quickly find the incongruities, the weirdness of what AI is, and that’s when the creative soul comes out, because you’ve got to work with it. You see this thing in your brain and you want to recreate it in AI, like you would’ve Photoshop or shooting on a camera. You’ll get 80% of the way there — maybe even 90% — but you’ll never get that last 10% because that’s where the perfectionism comes in.”

Shine also said that when using text-to-video AI tools, it can be difficult to generate consistent images and there will be lots of small mistakes — faces not being exactly the same, strange angles. That’s why you’ll see so many AI videos using yetis or Stormtroopers (and why Riley’s team chose famous artworks for the commercial): they’re easily identifiable, fairly simple, and the corpus of images in the training data is plentiful.

“It’s not as good or the same necessarily as it would’ve if it was shot for real,” he said of the commercial. “But this way we can have a lot more fun with it and not feel so precious about the money we’re putting in.”

Doing more with less

The idea of doing more with less is obviously tempting for companies that envision their ad budgets going farther than before. It’s also potentially terrifying for highly skilled creatives — directors, actors, graphic designers, creative directors — who considered creating video ads their lone province and one of the last parts of advertising safe from computers. Importantly, the idea is also more possible than ever, as a slew of text- and image-to-video technologies — Google’s Veo 3, OpenAI’s Sora, Synthesia, Runway ML — are in a growing market and are getting better seemingly daily. What’s more, the Big Tech companies that for years have been gobbling up bigger shares of the ad market are baking AI video ad creation right into their platforms. 

TikTok and Amazon already offer marketers the ability to quickly make video ads from a text prompt. Recently, Meta announced it’s testing a tool that automatically converts a still image into a multi-scene video ad, to be out next year as part of its effort to fully automate every part of the ad process, including creation, using AI. 

As Meta CEO Mark Zuckerberg put it during a shareholder meeting in June:

“I’m into an AI business agent that delivers measurable results at scale. In the not too distant future, we want to get to a world where any business will be able to just tell us what objective they’re trying to achieve, like selling something or getting a new customer, how much they’re willing to pay for each result, and connect their bank account. And then we just do the rest for them.”

It can be hard not to see the writing on the wall as it seems like AI is taking over more and more ads. AI ads are flooding not just social media but also prime-time screens nationwide, for small firms and giant corporations alike.

An LA Dentist used an AI content creation agency to come up with a funny and viral ad, featuring a cursing, skydiving, Cors Light-drinking gorilla to promote its services. 

Prediction market Kalshi recently ran an AI-created commercial during the NBA Finals illustrating the wild things people can bet on. It featured scenes of people wrestling alligators, a bride fleeing the cops, and a farmer bathing in a pool of eggs — stuff that would be expensive to shoot but that was reportedly created in two days and cost less than $2,000 to produce. A Kalshi rep told NPR the company was “incredibly pleased with the outcome and effectiveness of the ad” and that it “generated a lot of buzz on social media.”

Of course, AI-generated content gets a lot of criticism, too. Last year people called Coca-Cola’s Christmas ad, which was an AI-made homage to its 1995 classic, “soulless” and “devoid of any actual creativity.” Coca-Cola’s head of gen AI Pratik Thakar, however, has since told Marketing Dive the ad was a success with consumers. 

The technology is also creating a lot of AI slop.

“They are NOT GOOD!” Jon Elder, founder of Amazon seller consultancy Black Label Advisor, told me in an email about Amazon’s video from still-image tech. “Only the smaller brands with low budgets are using these tools. Once they can afford it, they are hiring an actual agency to create videos.” 

Doing more with more

Still, after each of these tech announcements, the stocks of the biggest ad agencies — Omnicom, WPP, Interpublic, Publicis — fall.

To be sure, people have been predicting the death of advertising for approximately forever. Meanwhile, global ad spending keeps generally going up. But the proportions of who’s eating that bigger pie is shifting and much of it is going to the Big Tech companies. 

When talking to people in the industry, the common refrain is that these types of AI tools are expanding the advertising market by allowing those who would have never had budgets for video ads — small and medium-sized businesses — to advertise. Meanwhile, the major corporations spending big bucks on advertising will mostly still rely on ad agencies, in-house or external, to carefully shepherd their brands. Interestingly, the top 350 brand advertisers make up only a quarter of total ad revenue, with the “long tail” of everyone else accounting for the other 75%, according to WPP Media’s “This Year Next Year” advertising forecast.

Ad agencies, to gird themselves, are employing more and more AI content generation themselves. 

WPP is using generative AI across its organization in everything from ideation to production, and even offers clients a number of the AI tools that are supposed to eat its lunch, like Veo 3 and Sora, within its WPP Open platform.

“AI is transforming how we create,” said WPP Chief Creative Officer Rob Reilly, who thinks of AI as a tool that benefits ad agencies like his.

“ I think it’s going to level the playing field for a lot of companies, but it’s also going to raise the bar for creativity,” he said in an interview. “The real premium is, who’s coming up with that idea that is so breakthrough or so fun or different or interesting, or makes you laugh or makes you cry? Being able to poke at a human emotion is still the most important thing.” 

In other words, a good ad is a good ad, no matter what mechanism was used to make it, and AI slop is still slop.

Execs at Omnicom are equally as bullish on AI.

“ Essentially we’re giving superhuman intelligence to everybody in the front office,” Jonathan Nelson, CEO of Omnicom Digital, told Sherwood. That includes using AI to analyze data and ad performance, but also in content production and personalization at scale  — altering ads to fit the customer or the climate they’re in.

Using AI to personalize ad campaigns, or “programmatic creative,” was a big area of excitement.

Personalization could mean anything from quickly changing an existing ad into different languages or swapping out background or characters to better advertise to a given person’s age, demographic, or climate. Much like Netflix swaps out the lead image of a show to most appeal to whoever is looking, entire ad campaigns could be converted to be germane to whomever it ends up in front of.

“Why do I keep seeing ads for cars on snowy mountain roads?” he said. “I live in Texas.”

Nelson also argued that tech companies can’t replicate all of what ad agencies do, which is handle, from end to end, complex ad campaigns that cross nations and platforms. They advertise on Meta and Google, of course, but not just there. 

“ To think that that’s going to replace everything that we do shows that you don’t understand what we do,” Nelson said. That includes describing and finding an audience, creating content that fits, distributing it across online platforms and in the real world, and finally gauging how well it performs.

Smaller ad agencies say AI makes them able to compete with the big dogs. Content creation is only a small part of its usefulness, said Michael Duda, managing partner of brand consultancy Bullish, which uses it for campaign production, content testing, and personalization as well. 

“ AI from a workflow perspective is incredible,” Duda said. “As a smaller agency, in the scheme of things, this is great ’cause we can weaponize AI and not have to hire and all that.” 

BlueChew is continuing to work on real ads, using real cameras and actors and scripts, in addition to its much cheaper forays into AI. The effectiveness of the ads and the public’s continued appetite for them is uncertain. What Shine is sure of is that, at least for now, he and people like him still have jobs, even if those jobs are different than they used to be.

“Now that I’ve gone in and I’ve been in these AI trenches,” he said, “you realize this is just going to allow us to  explore areas that we know we couldn’t afford to spend the time on earlier, or it’s going to allow us to iterate on things, but it’s still going to be a tool that needs to be wielded by people who are just as creative as ever.”