Amazon makes it so easy to buy on its marketplace through third-party selling that while waiting for the keynote by Senior Vice President of Amazon Ads Colleen Aubrey to begin I bought espresso pods for my Nespresso machine in less than a minute.

Aubrey opened the keynote with one of the most gracious lines directed to the generations — Gen X and Baby Boomer.

She called it an “incredible privilege” to live through two “massive transformations” — the internet and artificial intelligence. I agree.

Today, products from third-party sellers represent more than 60% of the units sold across the marketplace.

But the company wants to do more than offer these sellers ad services. It wants to provide a way for them to build creative and automatically generate images. On Wednesday, Amazon Ads launched a GAI creative tool to Sponsored brand advertisers at no cost.

The image generation, available in beta, aims to remove creative barriers and enable brands to produce lifestyle imagery that helps improve their ads’ performance. 

In the Amazon Ad Console, advertisers select the product and click “Generate.” The tool uses GAI to deliver a set of lifestyle and brand-themed images — based on product details — in a matter of seconds.

The image can be altered by entering short text prompts, while multiple versions can be quickly created and tested to optimize performance.

Aubrey said this is just the beginning of Amazon’s move into GAI for advertising using AWS Bedrock, a foundational model based on GAI. The goal is to launch a series of services based on AWS’s technology, including chips.

Ryan Hamilton, founder of Universimm, a sort of social platform with niche topics that launched formally in July, uses Amazon Rekognition, a tool supported by GAI for video and image analysis that flags inappropriate content such as sexually explicit material and partial nudity across the site, so moderators can remove them.

Half of large global firms will experiment with customer-facing GAI in 2024, according to Forrester Research, but companies that succeed will test GAI with employees before deploying it in customer-facing experiences.

Machine learning and artificial intelligence has helped Amazon identify challenges within its own business. At one time, the advertising business ran ads for products that customers could not buy directly from Amazon at the time — such as washing machines and refrigerators. In one week, the team saw a spike in the number of non-approved ads. Sellers were creating ads that linked to a product page in Amazon Store, not their websites.

This showed that Amazon was solving the wrong problem — and that was the birth of Sponsored Products.

More than seven in 10 advertisers abandoned the campaign creative process because they struggled to choose a format. 

Today, products from third-party sellers represent more than 60% of the units sold across the marketplace.