In the world of digital marketing, more and more evidence is emerging that artificial intelligence is not just a passing trend, but a real tool for improving efficiency. A recent A/B test conducted by kitchen appliance brand Cuisinart clearly demonstrated that AI video for advertising can outperform traditional production on key metrics. The brand, owned by Conair, compared a video created by Amazon’s neural network with classic studio creative, and the result was unexpected for many skeptics.
Generative creative showed better click-through and conversion rates, confirming that mass video generation by neural networks is already ready for real advertising campaigns. This case is an important signal for all those who doubted the capabilities of AI in content marketing.
Why AI creative didn’t fit the brand before
Cuisinart’s marketing team had long wanted to use generative AI, but the quality of images did not meet the brand’s strict standards. As Kelsey Smith-Eisen, Director of Amazon Marketing at Conair, explained, even static banners had errors: incorrect shade of green, distorted logos, inaccurate numbers and text on appliances.
«This even applied to simple banners,» she noted. However, with the rapid improvement in AI video quality, the company decided to test Amazon’s new tool — the «video creative agent.» Cuisinart became a beta tester for this solution, launched in 2024.
Flexibility vs. planning: the advantage of neural networks
Traditional marketing requires months of planning: booking studios, preparing locations. But the Amazon team works at a different pace — priorities change quickly, and it’s impossible to predict what will work in a week. Generative AI allows for quickly creating videos for current tasks, which is critical for dynamic platforms.
A/B test: AI video outperformed traditional
As part of the beta test, Cuisinart and its agency Global Overview launched two parallel campaigns with equal budgets: one used studio-shot video, the other used an AI-generated video. The AI version was refined with prompts, selecting from several options proposed by the agent. After a month, the results showed a clear advantage for generative creative.
Nikhil Naniwadekar, lead engineer for generative and agentic AI at Amazon Ads, explained: the agent doesn’t come up with radically new ideas but analyzes the brand’s past successful campaigns. It «relies on signals» and creates variations of what has already worked, but with novelty — this helps combat banner blindness among consumers.
Speed of progress: from static to nearly finished video in a year
Smith-Eisen emphasizes: the key is not just the test results, but the pace of technology development. A year ago, generating even a static image was unreliable and not worth the effort. Now AI almost fully creates video, requiring only minimal post-processing.
«In another year, if progress continues, we’ll be able to use videos immediately after generation, without refinements,» she predicts.
What’s next: personalization and new formats
Cuisinart plans to expand AI usage: creating not only product demos but also commercial videos. For example, if purchase data shows a person has moved to a new apartment, compact appliances can be shown; if a user starts cooking, relevant products can be suggested. The brand is also exploring the possibility of dialogues between AI characters, although for now it’s safer to use voiceover.
Importantly, generated videos only work within Amazon Ads — they can’t be transferred to YouTube or Instagram. But Smith-Eisen believes this isn’t necessary: each platform has its own audience, and differences in style are natural. Amazon Ads, in turn, is developing the tool: in June, a «one-click video generation» feature appeared.
Frequently Asked Questions
What is AI content production?
It’s the creation of advertising materials (videos, images, texts) using neural networks. Instead of studio shoots and a team of designers, AI generates videos based on text descriptions, saving time and budget.
How much does one AI creative cost?
The cost depends on the platform and complexity. On average, generating one video can cost from a few dollars to tens, which is significantly cheaper than traditional production, where a minute of video can cost thousands of dollars.
Can I order AI UGC for a crypto project?
Yes, many services offer the creation of AI avatars and videos in UGC (user-generated content) style for cryptocurrency projects. This allows for quickly scaling advertising without involving real bloggers.
What are Synthesia alternatives?
Besides Synthesia, there are HeyGen, Colossyan, Elai.io, and others. They allow creating videos with AI avatars and subtitles. Amazon Ads also offers its own tool, but it only works within the platform.
Conclusion
The Cuisinart test is another proof that AI video generation for advertising is becoming not just an experiment, but a working tool. Neural networks are already capable of creating videos that are not inferior to traditional ones in effectiveness, and surpass them in speed and scalability.
If you’re thinking about implementing AI in your marketing, start small: test generating one video and compare the results. Maybe it will change your approach to content. And if you need help implementing AI tools into your advertising strategy — contact us for a consultation.