Meet Amelia: Amazon’s New AI Assistant for Third-Party Sellers

Amazon is launching an artificial intelligence tool to help third-party sellers quickly resolve account issues and get sales and inventory data.

The company announced Thursday that it will launch the product, called “Amelia,” in beta for select U.S. sellers, with plans to roll it out to the public broadly later this year. Amazon describes it as an “all-in-one, generative-AI based selling expert,” and will make it accessible through Seller Central, its internal dashboard for third-party sellers.

Amelia is the latest generative AI tool that Amazon launched last year to capitalize on the hype generated by OpenAI’s ChatGPT. The company has launched an AI-powered shopping assistant called Rufus, an enterprise chatbot called Q, and Bedrock, a generative AI service for its cloud customers.

As CNBC previously reported, Amazon also plans to upgrade its Alexa voice assistant with generative AI capabilities, and the company has invested billions of dollars in OpenAI competitor Anthropic, its largest venture deal to date.

CEO Andy Jassy told investors earlier this year that the “generative AI opportunity” is nearly unprecedented and that increased investment is needed to capitalize on it.

“I don’t know if any of us has seen a possibility like this in technology in a really long time, for sure since the cloud, perhaps since the internet,” Jassy said on the company’s first-quarter earnings call in April.

Google and Microsoft have launched competing products to ensure relevance in a market that is expected to exceed $1 trillion in sales within a decade.

AI is also increasingly prevalent on Amazon’s e-commerce platform. The company now displays AI-generated summaries of product reviews and is introducing AI capabilities for third-party vendors to help it build listings and generate photos for ads.

Amazon also announced Thursday that it will release tools to enable sellers to create AI-generated video ads and use AI to bulk-create product listings based on their entire catalog. The company said it is starting to use generative AI to show personalized product recommendations and listings based on users’ buying history. For example, Amazon will show the phrase “gluten free” in the description of a cereal box if a shopper would typically search for a product that contains that phrase.

Amazon made the announcement at its annual Seller Conference in Seattle. Third-party sellers are central to Amazon’s dominant e-commerce business. Since about 2017, they’ve made up at least half of all items sold on the site. In the second quarter of this year, that figure rose to 61%.

More and more sellers are using the company’s AI services, Dharmesh Mehta, Amazon’s vice president of global partner services, said in an interview with CNBC. He said more than 400,000 of Amazon’s 1 million third-party sellers are using the AI ​​selling tool, up from 200,000 in June.

With Amelia, Amazon is using generative AI to help third-party sellers with the crucial issue of troubleshooting their accounts. The company has an extensive team that helps sellers resolve account holds, manage inventory issues, and build their business on the site. Sellers have long complained that it’s hard to get a quick resolution or contact a human when unexpected issues arise with their accounts.

The company said Amelia can help investigate account issues and in the future “solve the problem on the seller’s behalf.” Mehta explained how sellers can ask Amelia to file claims on their behalf instead of filling out out-of-stock forms, and how the tool can automatically resolve issues.

“There are going to be places where, hey, instead of chatting with seller support or getting on the phone with someone, maybe Amelia is able to do that and do that faster,” Mehta said. “I don’t need to send an email to someone and wait for a response.”

Amelia said it uses a software tool that gives users access to large language models from Amazon, Anthropic and Stability AI. Mehta said Amelia is trained using public data from the internet, as well as information gleaned from Amazon seller resources, FAQs and other public websites.

Mehta said the model is not trained on seller-specific data, which is confidential.

Amazon said the tool uses Search Augmented Generative (RAG), a popular industry AI framework that combines generative AI with long-established information retrieval techniques. This allows seller-specific information to be pulled from Amazon’s internal systems without being stored or included in the model’s training data.