For online advertisers, securing the lowest cost-per-click (CPC), cost-per-thousand impressions (CPM), and cost-per-lead (CPL) is paramount. However, the “lowest costs” are rising quickly.
Google, which accounted for 90% of search traffic in December, according to data aggregator Statcounter GlobalStats, saw its “average CPC across industries jump 12.9% year over year [throughout] 2025,” noted Mktg, a developer of marketing systems, in September. “Over the past three years, CPCs have risen nearly 40%, while conversion rates improved only 7.5%.”
Set by the advertising platform, the cost of posting an ad online is non-negotiable. To cut costs, companies are turning to technology to reduce overall marketing expenses. According to advertising agency MNTN, AI-generated ads “alleviate certain high-cost components of production, making it accessible to brands that previously couldn’t afford commercial content.” Risks include “lack of brand voice consistency, limited emotional resonance [and] potential consumer skepticism.”
Overcoming these potential issues is crucial. “AI-generated creativity isn’t going anywhere. It’s only getting smarter, faster, and more integrated,” said MNTN. “Expect more advanced personalization, where ad content adapts dynamically based on viewer data.”
Already here
The first attempt to use AI to create ads was in 2021, when Cadbury India used the technology to create personalized video promotions for each of their local shop owners. The ads featured an AI- generated avatar of Bollywood (India’s Hollywood) star Shah Rukh Khan endorsing their shop by name.
In 2022, Nike celebrated its 50 th anniversary with a fully AI- generated ad featuring a tennis match between Serena Williams when she debuted in 1999 and Williams in 2017. According to the blurb, the AI analyzed 130,000 matches to recreate her evolving playstyle.
Also that year, Heinz created an ad showing that every time they asked DALL · E 2, an AI software, to draw a ketchup bottle with different prompts, the results were nearly identical to a Heinz bottle. KitKat’s ad said when starting a prompt on any AI chatbot with “Have a break, and then…ask it a question,” it will generate more accurate results. The ad plays on KitKat’s slogan, “have a break, have a KitKat.”
Under Armour, a sports apparel brand, aired an ad featuring an AI avatar of Anthony Joshua, a British boxer, without disclosing they used AI to depict the athlete. Tool, the agency that created the ad, later explained on its website they were running short on time.
In 2024 and 2025, Coca- Cola aired a reimagining of its 1995 “Holidays Are Coming” ad campaign. For the first time, the ads relied entirely on AI. The 2024 version used the technology to render human characters. The second featured animals exhibiting human- like surprise and happiness.
Kalshi (an NBA Finals spot) was also fully AI- made. It featured ordinary people in mostly realistic settings, but didn’t disclose using AI to generate them. CNBC reported the 30- second ad was made “from idea to live in less than 72 hours and cost $ 2, 000.”
McDonald’ s (with a Christmas ad like Coca-Cola’s) and Popeyes (announcing their chicken wrap sandwich) used AI to depict familiar- looking settings and people. The Popeyes ad reportedly took three days to complete, with both the song and video footage created by AI algorithms.
There is little doubt the list of companies using AI will grow. “Chief marketing officers at many of the world’ s biggest brands have made artificial intelligence a centerpiece of their strategies,” reported Business Insider in December.
Sink or swim?
According to a Business Insider-commissioned survey of 6,000 consumers published in December, 39% held a negative opinion of AI-generated advertising, 36% were neutral, and 18% had positive views.
Matt Barash, chief commercial officer at the adtech platform Nova, told Business Insider, “When brands ask AI to invent stories from scratch, they don’t get innovation, they get an approximation of human emotion, and the result can make headlines for all the wrong reasons.”
Huei-Hsin Wang, a senior experience specialist at Nielsen Norman Group, a research firm, singled out AI-generated holiday ads as almost certain to fail. “AI holiday ads lack authenticity and emotional resonance,” she said. The McDonald’s and Coca-Cola ads faced “significant public backlash.” “The criticism stems from a mix of aesthetic, ethical, emotional and economic factors, often intertwined with broader societal concerns about AI’s role in creative work.”
That negative sentiment will likely persist. “Despite significant improvement in AI-video models over the past few years, generative AI still struggles to achieve the level of realism that aligns with real-world logic,” said Wang. “AI-generated visuals often provoke [an] uncomfortable sense of ‘almost human, but not quite.’ Once viewers notice these subtle distortions, the emotional connection begins to collapse.”
Accordingly, “to avoid exposing … limitations of AI, these ads rely heavily on workarounds: rapid montages of one-second clips, distant camera shots to conceal distortions and minimal close-ups to avoid uncanny details,” said Wang. “These choices mask the flaws in technology, but also undermine the creative storytelling.”
The result is that such ads are “often criticized as soulless, algorithmic, lacking human creativity or authorship,” Wong explained. “While humans may provide prompts and high-level direction, they do not fully control how those ideas are translated into the final output; much of the creative execution is determined by the model’s internal algorithm.”
Maintaining realism
Darya Buchakova, design director at Flowwow, an online marketplace, told Campaign Middle East in July the key to making AI-generated ads resonate and align with a brand’s identity is to “preserve the human touch in their campaigns.”
“AI-generated visuals [that are] more eye-catching and aesthetically appealing … don’t always mean [they are] trustworthy,” said Buchakova. “If an AI-generated image oversells a product like flowers, food, or home decor, and the real item doesn’t match, brand trust takes a hit.”
Sergio González Calmaestra, AI lead at fashion brand Desigual, who launched the brand’s first AI-generated campaign, told Campaign Middle East, “Behind every generated image, there [were] human decisions: carefully crafted prompts, art direction, studio management, retouching.”
“AI needs to be trained on the things that define your brand, [including] guidelines, color palette, typography, logo, and tone of voice,” said Buchakova. “That requires giving the AI model successful examples of past campaigns to train on.”
To “humanize your output,” she stressed, an advertisement should include a “narrative” before it’s prompted to the AI model. The AI-generated ad process also must include feedback loops in these prompts. “Always run visuals by a real creative team.”
Lastly, the AI-generated end product must be reviewed by artists and designers to “adjust details, enhance emotion or replace awkward elements,” Buchakova noted. “This editing must be done manually, with a clear understanding of the creative objective.”
Procedures and skills
Buckalova said companies relying on AI-generated ads need to establish standard operating procedures (SOPs) that ensure the AI generates multiple versions from each prompt and humans provide “creative reviews” of all options.
These SOPs should set “established standards” that ensure humans are “examining structure, lighting, texture, facial expressions and anatomy,” she said. “These checklists help preserve the team’s visual intuition and foster a refined design taste.”
Accordingly, teams need to be trained to identify “lifeless elements,” such as “how light behaves, what gives an image depth or emotion,” and common AI glitches, like plastic-looking hands or eyes.
Ultimately, “real magic happens when human intuition and machine learning work together,” said Buckalova. “The brands that will truly shine in this new era are those that experiment boldly, customize their tools, and always keep the human perspective at the core.”
