Creative directors? Who needs those?
On Wednesday, Amazon introduced an agentic AI tool that brings all the image, video and audio generation capabilities of the company’s Creative Studio into its ad platform creative setup.
Advertisers can use the new free-to-use chatbot agent to run research reports, generate marketing concepts and directly operate Amazon’s other creative tools to fulfill user prompts. Theoretically, the product could develop entire marketing campaigns from scratch.
Amazon previously designed its Creative Studio to target small business owners who rely on search advertising. Although, these small and local advertisers – the bedrock of Google’s and Facebook’s account base – often face limitations on total budget, expertise and resources.
With this new tool, however, Amazon is setting its sights on more mid-market and even independent agency clients, Jay Richman, VP of product and technology at Amazon, told AdExchanger.
“There’s just so much creative that wants to be produced that simply isn’t today,” said Richman. “Everywhere you look, we see untapped demand.”
The new Creative Studio
Amazon’s new agentic creative system essentially combines the functions of all the platform’s previous generative AI tools, including its image generator, video generator and its audio generator. What’s new is the agent layer on top that performs deep reasoning and research capabilities, using other foundation models like Amazon Nova and Anthropic Claude.
First, the user initiates a chat-style conversation with the agent with queries or requests for the AI. Given a particular brand, for example, the user might prompt for a video ad with a goal to drive engagement.
The AI analyzes the brand’s online presence – product detail pages, visible social media accounts and previously published or web-scrapeable material – as well as whatever kinds of briefs or brand guidelines set manually by the advertiser. On top of that, the new creative tool kit can tap into proprietary Amazon shopping data, which, obviously, is not available elsewhere.
After that ingestion and research phase, the AI generates short text-based concepts and taglines for the user to select from.
Once the user selects a desired concept, the agent moves on to generating storyboards and visuals. That would eventually lead to multi-scene videos and display ads. It can also bring in featured music or voice-overs.
The final output, an example of which was shown to AdExchanger by Amazon, still carries some of the trademark elements of AI-generated imagery, such as over-warmed hues, odd textures and inconsistent minor details. That said, if the last time you watched an AI ad was when that Toys”R”Us spot made headlines last year, the Creative Studio’s quality is certainly higher caliber than that.
Even if it still looks a little … off.
After the strategy and creative is set, the ad is ready to be served across several products and formats in Amazon’s arsenal, including Sponsored Brands, Sponsored Display, Brand Stores and even the Amazon DSP.
But does it work?
One test of Amazon’s true AI creative chops is to consider whether a creative strategy or piece of content created by the Amazon Creative Studio could conceivably be repurposed for other uses, such as product listings on brand-owned stores or across social platforms.
Ultimately, that’s irrelevant. The generative AI creative is intended to improve sales and visibility on Amazon. (Case in point: Although the Creative Studio is technically free to use, an Amazon Ads account is required for full access.)
“This isn’t meant to be like an off-the-shelf creative suite that you can get from anywhere,” said Richman. “It’s very much steeped in understanding Amazon’s ad products and publisher environments.”
So far, added Richman, only half of the generative creative assets have actually been used, either in ad campaigns or store listings. However, every time a new model is released or updated, the adoption rate increases.
Those early adopters of the Creative Studio are also reportedly seeing better returns on Amazon, at least according to the retailer itself. On average, advertisers who use the AI tools see a 12% increase in sales, said Richman.
Of course, this raises the question: As these tools become more widely adopted, how will advertisers find ways to stand out against a sea of sameness that fully AI-generated assets might usher in?
Richman had an answer for that: While the AI agent is doing personalized research for each individual brand, it’s designed to determine a product’s unique qualities compared to everything else on the market.
The focus on personalization is also useful for fulfilling Amazon’s ultimate AI goal, which is to inspire more businesses to experiment with these new tools in the first place.
“We think that as the cost to produce this content goes down, the total amount of variations and experimentation will go up,” Richman said. “Which will result in a much more personalized creative experience.”
