Amazon ads unleash creative potential with image-generation AI

Advertising has always combined elements of art and science, and it appears that the creative aspect is on the verge of significant enhancement. Over the past few decades, there have been remarkable advancements in digital advertising, thanks to data science, analytics, and artificial intelligence (AI). However, a survey conducted in March 2023 revealed that a majority of advertisers facing challenges in creating successful campaigns cited the primary obstacles as crafting ad content and selecting the appropriate creative format.

To address these challenges, a beta version of image generation powered by generative AI has been introduced by Amazon Ads. This innovative solution aims to eliminate the constraints on creativity and empower brands to create lifestyle imagery that can enhance the performance of their ads. For instance, placing a product, such as a toaster, in a lifestyle scenario, like on a kitchen counter next to a croissant, has the potential to boost the click-through rate for mobile sponsored brand ads by 40% compared to using standard product images.

This innovation is designed to be user-friendly, catering to advertisers of all sizes. It simplifies the process for those who lack in-house creative capabilities or agency support, making it equally accessible to larger brands seeking more efficient creative development.

Using this tool within the Amazon Ad Console is straightforward. Advertisers select their product, click “Generate,” and the generative AI swiftly delivers a set of lifestyle and brand-themed images based on the product details. Users can further customize these images with short text prompts and create and test multiple versions to optimize their ad’s performance.

Amazon Ads aims to gradually introduce image generation to select advertisers and expand its availability while refining the feature based on user feedback.

The company has taken an additional step by introducing an AI image-generation tool, allowing advertisers to create backgrounds based on product descriptions and themes. This tool is at present, in beta testing with select advertisers and is intended to become more widely accessible over time. Advertisers can upload an image, describe the desired background in text, select a theme, and click “Generate.” They can also refine the image further using additional text prompts and test multiple versions for performance optimization.

While the beta tool may have some imperfections, it has demonstrated the ability to create convincing backgrounds for images, as seen in the example of a toaster placed on a kitchen table adorned with fall decor.

In a broader industry trend, many brands are turning to generative AI to simplify the process of creating ads, which can be both expensive and time-consuming. Even major companies have acknowledged using software for this purpose.

In recent months, there has been an expansion of efforts in generative AI, with tools introduced to assist sellers in writing product descriptions and summarising customer reviews.

This functionality operates much like other open-ended text-to-image generators like OpenAI’s DALL-E 3 and Midjourney. Users provide descriptive text prompts, and the AI responds by creating an image that aligns with the given descriptions.

In a blog post, Amazon stated that they developed this new AI product image generator to make it easier for individuals and businesses lacking in-house capabilities or agency support to craft brand-themed imagery.

While the feature is at present, in its beta phase, it comes shortly after a similar announcement by Meta Platforms, which introduced a comparable tool for its advertisers on Facebook and Instagram.