How to use AI to generate images

Question 1

What is AI image generation?

Answer

AI image generation uses generative AI to create visual content, such as images and videos, using machine-learning algorithms. Generative AI is a type of AI that can produce new content and ideas, including conversations, stories, images, videos, and music. AI image generation can serve as a powerful tool for brands to produce engaging and diverse creatives. This guide provides an overview of AI image generation, available in the Amazon Ads ad console, to help brands of all sizes to easily produce ad-ready creatives.

The Amazon Ads image generation (beta) capability is a free generative AI solution that enables brands of all sizes to easily make lifestyle and brand-themed images based on product details, and refine them through short prompts or by applying seasonal and lifestyle themes.

Question 2

How does AI image generation work?

Answer

AI image generation works by using generative models like diffusion models or generative adversarial networks. These models are trained on data sets of existing images to learn the patterns and distributions of pixels that make up different types of visual content.

The Amazon Ads image generation capability presents an accessible, user-friendly generative AI tool that empowers advertisers to generate lifestyle visuals with just a click and seamlessly integrate into their Sponsored Brands and Sponsored Display campaigns. Image generation simplifies creative building to a few actions, removing the need for technical or design expertise.

Question 3

Why should I use AI image generation on Amazon Ads?

Answer

Whether it’s helping reach new audiences on a website or boosting conversions for your Brand Store on Amazon, quality creative remains key. Image generation can produce product- and brand-oriented creative in seconds, speeding up creative production and campaign activations. With image generation, you can also compose infinite variations to test and learn which creative drives the strongest engagement and better return on ad spend.

Question 4

How does AI image generation help optimize my Amazon Ads campaigns?

Answer

Enhanced campaign content can help create better brand experiences and help advertisers more effectively showcase their brand. Image generation lets you experiment with creative without expense. House as many creatives as needed in the creative assets library with no asset cap. And after assigning your AI-generated images to campaigns, you can measure click-through rates, advertising cost of sales, and return on ad spend in the advertising console.

Question 5

How can I create engaging creative with AI image generation?

Keep it concise and to the point. Only specify what you want to see in the generated image.
Reference the product as already-present in the product image: "A box of cans on a table," "A sandal on a beach," or "A face wash in a bathroom." Avoid any instructions or verbs like "Put the," "place the," "now add a."
Be specific: Describe exactly what you want the model to create. Make sure to specify what you want the scene to include (the subject), where the elements of the scene should be placed, and how you would like the scene to be composed (description of the scene).
Use precise adjectives to help the model understand the mood or style of the image. These should follow the scene description in your prompt.
Avoid ambiguous language and conflicting terms which can lead to confusion and may result in the model generating images that don’t match your vision, for example, "Black and white spotted dog" is ambiguous because it can be interpreted as a "spotted dog in black and white style" or "a dog with black and white spots."
Use positive language in your prompt. Negative language such as "do not include" is typically ignored by the model.
Test and revise your prompt’s structure to optimize the generated images. If you like certain elements, you don’t have to completely rewrite your prompt. It may help prompt development to keep a notepad of inputs and results. This way, if you get results that absolutely don’t work, you can record that, and change your approach.
Avoid subsequent, add-on inputs as each input is treated as independent to the model. Revisions should include all information as previous inputs are not remembered.

Do: On the first try, input "A sandal on a beach." On the second try, input "A sandal on a beach with a tree next to it."
Don’t: On the first try, input "A sandal on a beach." On the second try, input "Now add a tree next to it."

Enter prompts to refine your branded creatives

Answer

Generative AI outputs can make mistakes. Follow the recommendations below to help produce high-quality creative using the image generation capability:

Question 6

How do I use AI image generation in Amazon Ads campaigns?

Step 1

Choose from three entries in the advertising console:

Click the creative tools icon in the sidebar and click Image generator.
Create a new campaign for Sponsored Brands > Product Collection > Creative > Custom images.
Create a new campaign for Sponsored Display > Creative > Add image.

Step 2

Select products (ASINs) to promote.

Step 3

Access image generation by clicking “Choose an AI-generated image.”

Step 4

The menu will automatically present pre-generated enhanced product images.

Step 5

Use prompts and themes to generate or refine creative variations.

Step 6

Save custom images to your creative assets library for future use, or add them directly to a campaign.

Answer

Access AI image generation to start generating in six easy steps:

Question 7

Who can use AI image generation on Amazon Ads?

Answer

Image generation (open beta) is available in the U.S., Canada, Mexico, France, Germany, Italy, Spain, and the U.K. Access the tool from the advertising console or from the Sponsored Brands product collection or Sponsored Display campaign workflows.