Amazon Launches Project Amelia, a Commerce Chatbot for its Third-Party Sellers
Amazon’s latest generative artificial intelligence tool, Project Amelia, is set to ameliorate the seller experience.
The e-commerce giant announced Project Amelia, a chatbot aimed at answering sellers’ queries and sharing insights to expand their businesses, at its Accelerate conference in Seattle on Thursday.
More from Sourcing Journal
The beta version rolled out to a select group of sellers on Thursday, and the company plans to bring access to additional U.S. sellers over the next few weeks. Down the line, Amazon plans to make it available in other countries and in languages other than English.
Project Amelia has several key functions. It can answer knowledge-based questions, which may include queries about how to prepare for an upcoming shopping event, like Amazon Prime Day or Black Friday. A fashion and apparel seller might ask Project Amelia, “What are the top three items I should add to my store to attract new sellers on Black Friday?”
The system can also provide updates on metrics and data based on a brand or seller’s sales over time. For instance, a seller might ask, “How are sales on pink shirts?” and receive a response about sales volume, comparing the sales of pink shirts in September of this year to September 2023.
Though the third function Amazon boasts about Project Amelia has yet to launch, the system is slated to be able to handle problem resolution with sellers, whether about inventory issues or other potential incidents. The company said Project Amelia will be able to solve the issue independently for a seller or will connect the seller to human support personnel as needed.
Amelia is built on Amazon Bedrock, which brings together multiple large language models and makes them accessible to those building AI systems.
Mary Beth Westmoreland, vice president worldwide selling partner experience at Amazon, said Project Amelia provides a new level of depth to the seller experience.
“Project Amelia delivers a personal Amazon selling expert that understands a seller’s unique business and provides personalized recommendations, insights and information to help sellers thrive,” she wrote in a blog.
Tools for listing and consumer customization
Though Project Amelia may be the most buzzed about piece of Amazon’s new generative AI tools, it has also introduced a tool for upping the game on consumer-facing product recommendations. Earlier this year, Amazon announced its Rufus chatbot, targeted at answering shoppers’ questions in real time. But the newest tool from the company helps hyper personalize the experience based on product titles and descriptions; the changes consumers see are almost imperceptible.
“Based on a customer’s shopping activity, Amazon reviews the customer’s preferences to create personalized recommendations categories on our homepage and throughout the shopping journey, as well as provide personalized product descriptions that are more relevant to individual customers,” Westmoreland wrote.
That means that, if a consumer had been actively searching for a wedding guest dress via Amazon, the app might offer them a category that says, “Formal dresses for a fall wedding,” rather than simply saying, “More like this.”
Alternatively, if a consumer repeatedly searches for a specific attribute over and over, Amazon may move that particular attribute up higher in the product title or description. In her blog, Westmoreland uses the example of a consumer who consistently looks for gluten-free food items.
“If a customer regularly searches for gluten-free products, the term ‘gluten-free’ may be added to a relevant product description to help the customer find the best product to meet their needs,” she explained.
Though this feature may be aiding and easing consumers’ shopping journeys, it also ensures sellers’ products end up on consumers’ radar as needed; instead of actively searching through long listings and reviews, the information consumers need about a product will be quickly accessible and available to them.
Already, Amazon uses generative AI to help sellers create one-off product listings with titles and descriptions the model perceives will attract consumers; in the coming months, Amazon plans to make that possible for bulk listings. Westmoreland said sellers will be able to import full spreadsheets with “a sparse set of listing details,” which Amazon’s AI models will turn into full descriptions for multiple listings simultaneously.
Gen AI for product imagery and video
When it comes to product listings, consumers have indicated that images and videos are of particular importance to their decisioning around making purchases. Amazon wants to make it easier for sellers to show off their products, whether in beautiful settings or via video.
According to Westmoreland, “rich image carousels and comparison charts on…product detail pages,” or A+ Content, as Amazon calls it, can help sellers’ sales increase by up to 20 percent. Sometimes, though, creating robust image sets can be costly or too resource intensive, particularly for smaller sellers.
In its slew of new announcements, Amazon noted it’s looking to make strong product imagery more accessible.
“Amazon is enabling brands to capture the essence of their brand by inputting a few descriptive product terms they want to highlight, like ‘quality leather’ or ‘versatile for any occasion,’ and the tool will generate rich narrative content that highlights the product’s unique qualities,” Westmoreland said. “By uploading an image of their product, sellers can also generate captivating background imagery and scenes that bring their products to life and more effectively showcase their brand identity.”
Amazon has also taken product shots one step further; it announced Thursday the launch of a new tool called Video Generator, which enables sellers to upload a still photo of their product which the generative AI-powered system turns into “visually rich video content in a matter of minutes.”
In the blog, Westmoreland noted that Project Amelia, Video Generator and each of the other behind-the-scenes, generative AI-powered tools it launched Thursday are meant to give sellers a new lease on their businesses, empowering stronger sales and better listings.
“These impactful new tools are the latest examples of how Amazon is leveraging generative AI to transform nearly every experience we know—including how we support independent sellers in creating thriving businesses by selling in our store,” she said. “We’re super excited about the innovation and value that these new offerings provide for sellers, and we will continue to develop additional, innovative AI tools to further enhance the success of Amazon’s selling partners.”