Pulse News

**How Meta and Google Benefit From AI-Created Ads—Made With Other Firms’ Technology**

By Catherine Perloff

Mar 5, 2025, 8:50am PST

Five months ago, app studio Sociaaal, which creates apps that help people identify their celebrity look-alike or suggest pick-up lines to score a date, started using generative artificial intelligence to create ads for running on social media, including YouTube. The AI-created ads included some Sociaaal's marketing team say they never would have thought of, like a video of an older woman encouraging people with Gen Z lingo to download Sociaaal's dating app Rizz God.

More importantly, Sociaaal and its ad agency were able to cheaply ramp up the number of ads they create, to 400 a month from 40. Using AI tools from startups such as HeyGen and Poolday AI, Sociaaal created enough variations of images, texts, and videos to appeal to every type of potential customer. The tactic worked—CEO Patrick Stuart-Constant credits the company's 65% growth in revenue over a four-month period—mostly subscriptions to its apps—to the increased variety of ads it devised.

**The Takeaway**

- New AI tools allow marketers to make many different versions of an ad

- Explosion in ad quantity is driving higher returns—and more ad spending

- Google is paying advertisers to test other firms' AI tech

The good news for tech firms Meta Platforms and Google: Sociaaal is reinvesting much of that increased revenue in more ad spending, Stuart-Constant said. Sociaaal isn't alone. Shamanth Rao, CEO of Sociaaal's ad agency, RocketShip HQ, said he had several other clients also using generative AI from firms like HeyGen, Midjourney, and OpenAI's Dall-E to make their ads. Those clients have increased their spending on advertising by 30% to 40% recently.

The experiences of Sociaaal and RocketShip show how generative AI technology can boost ad revenue for big tech companies, even when advertisers aren't using the companies' creative ad tools. Google has gone out of its way to tell its advertisers they're happy with whatever generative AI ad tools they want to use as long as they're spending money on Google. The company's advertising division has even set up funds to reimburse advertisers for the costs of trying generative AI technology, even if Google doesn't make the tools, said an ad firm executive and a person familiar with Google's thinking.

A Google spokesperson said the company has several initiatives that support the ecosystem in applied AI innovation and integration with Google Ads specifically. The spokesperson said that Google believes in the value of generative AI improving ad performance.

Google's gambit may pay off, given that seven brands and ad agencies said generative AI made their ads work better. If Google can convince more advertisers to adopt generative AI on their platform, and advertisers spend more on Google as a result, its revenue could continue to climb. But their competitors might benefit as well.

Google and Meta are likely to particularly benefit from advertisers' generative AI strategies because their superior technology for personalized advertising makes it easier for the abundance of ads generative AI creates to reach the right person. Meta didn't have a comment.

**More Is Better**

The primary benefit of generative AI in ad creation is the ability to customize ads to appeal to more types of audiences. Sociaaal has been able to run ads in different geographies, using the generative AI tools to create ads featuring actors speaking the local language, Stuart-Constant said.

"Before it would be prohibitively expensive to [create 400 ads] with [user-generated] videos, content creators, actors," said Stuart-Constant. "These ads are superfast to produce and much cheaper to produce."

By making more ads, marketers increase the likelihood that consumers will respond to one they see by downloading an app or buying a product. "I can find winners much faster," said Rao. "The more shots on the goal I have, the easier it is for me to come up with a winning ad."

Hatch, which makes devices that help people sleep, used generative AI to create a campaign targeting three different personas who might care about their sleeping habits—a biohacker, a wellness enthusiast, and a stressed professional. Hatch's agency, Monks, used Google's Gemini and its own internal AI tools to quickly identify these three personas, said Monks' executive vice president of digital media, Brittany Blanchard. Previously, to identify the best potential target customer, Monks would have had to spend hours with focus groups to figure out what types of consumers would be most interested in Hatch's devices, she said.

Once Monks came up with the personas, it used Gemini and an AI tool it had developed to make more than 60 different ad variations to appeal to those individuals. For the biohacker, for instance, an ad might talk about how Hatch can improve someone's sleep score. For the professional, the ad could call out how Hatch makes you more alert for a big presentation in the morning.

Blanchard said using generative AI cut in half the time required to make the campaign from 12 to six weeks. Advertisers often judge the success of a campaign partly on how much it ends up costing to get someone to buy something. For this campaign, the advertising cost per Hatch device purchased was 31% lower than similar campaigns Hatch ran that did not use the generative AI strategy, Blanchard said.

Having enough versions of an ad to appeal to very specific audiences works better if there is also technology to target them. Apple and government regulation have over the past several years made it harder for advertisers to target individuals. But both Google and Meta have introduced new technologies to identify the right consumers when data is more scarce. Those tools are particularly effective with the abundance of personalized images and text that generative AI can create, Blanchard said.

"The way we used to target and get really specific is simply not possible anymore," she said. "The platforms have gotten much smarter with the signals that they're able to target. Creative variety and volume, plus the AI-powered campaign delivery, [works]."

Catherine Perloff is a reporter covering the intersection of advertising, media, and artificial intelligence. You can reach her at catherine@theinformation.com or on Signal at 216-650-0155.