Amazon Canvas brings interactive dashboards and scenario planning to third-party sellers

Amazon launches AI-generated 'canvas' for sellers as e-commerce platforms race to add AI tools

Amazon is adding a new layer to its AI tools for online sellers, rolling out a feature that provides customized, interactive dashboards and scenario planning in real time. The new "dynamic canvas" in Amazon's Seller Central displays data, charts and different options for sellers in response to user prompts. It expands the existing Seller Assistant, an AI chat tool that Amazon introduced in 2024 and later upgraded with agentic capabilities. "It's the difference between giving someone a better calculator and giving them a financial advisor who really understands their business inside and out," said Mary Beth Westmoreland, Amazon's vice president of Worldwide Selling Partner Experience, in an interview. The announcement comes as e-commerce platforms race to embed AI into their tools. Shopify offers its Magic and Sidekick AI assistants. Walmart built a tool called Wally for merchants and partnered with OpenAI and Google to let consumers use chatbots to shop. The financial stakes are high for Amazon: Independent sellers account for more than 60% of sales in Amazon's store, and third-party seller services generated more than $172 billion in revenue for the Seattle-based company in its most recent fiscal year. Not all sellers have embraced the AI tools. An Amazon Seller Central forum post announcing an update four months ago drew 36 thumbs-down votes against 5 thumbs-up. Sellers complained about broken links, generic recommendations and, in one case, an AI response that advised a seller dealing with false policy violations to seek legal action against Amazon. Asked about the forum feedback, Westmoreland cited internal data showing that sellers accept Seller Assistant's recommendations nearly 90% of the time, as an indication of its effectiveness. The new canvas for Seller Assistant goes further to provide "a lot more context to sellers in a visual way" that helps them understand recommendations more deeply, she said. Sellers can ask questions like "How are my products performing?" and the canvas generates charts showing sales trends, traffic and inventory health. They can also test scenarios such as "What if demand drops 10%?" and see projected impacts on revenue and cash flow before committing to a decision. Much of the interface is generated dynamically by the AI rather than pulled from preset templates. The system runs on Amazon Bedrock using Amazon's Nova models and Anthropic's Claude. Westmoreland said Amazon built the architecture to swap between models as needed. The Seller Assistant canvas will be available starting Tuesday to all sellers in the U.S. and U.K. at no additional cost. It launches with performance analysis tools, and Amazon plans to add marketing optimization, inventory planning and product launch features in coming months.

Amazon Gives Sellers an Agentic Window Into Their Business | PYMNTS.com

By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. The timing matters. More than 60% of sales in Amazon's store now come from independent sellers. As that share has grown, so has the complexity of running a business on the platform. Sellers must manage inventory, track marketing performance, plan new product launches, and respond to shifts in customer demand, often without large teams to support them. Canvas is Amazon's answer to that challenge. Mary Beth Westmoreland, vice president of worldwide selling partner experience at Amazon, described the launch as more than an incremental product update. "This is really about the next phase in reimagining the entire seller experience," she said. "Not just adding new AI features to existing workflows." Westmoreland explained the goal was to move sellers from checking static reports to actively engaging with their data in real time, comparing the shift to "the difference between giving someone a better calculator versus giving them a financial adviser who knows their business inside and out." Canvas is built on the same agentic AI architecture as Seller Assistant, Amazon's existing chat-based tool for sellers, which is powered by Amazon Bedrock and uses models including Amazon Nova and Anthropic Claude. Where Seller Assistant responds to seller questions through conversation, Canvas extends that capability into a visual, interactive workspace. Sellers can ask a question or choose from suggested prompts, and the system assembles a personalized dashboard that brings together sales data, inventory levels, customer traffic trends, and recommended actions in one view. The experience is designed to be conversational. A seller can ask Canvas to analyze their sales performance, and the system will generate charts and recommendations based on their specific data. If they want to go deeper, they can ask follow-up questions or request a different perspective, and Canvas updates in real time. The same logic applies to inventory decisions, where Canvas can model multiple restocking scenarios and show the projected impact of each on revenue, cash flow, and storage costs. Sellers can pose hypothetical questions such as "What if demand drops 10%?" and see updated projections before committing to any action. For marketing, Canvas can analyze campaign spend, impressions, conversions, and sales lift, then propose forward-looking strategies with projected outcomes. Sellers can adjust constraints on the fly, such as asking Canvas to focus only on products with excess inventory, and the recommendations update accordingly. Canvas also helps sellers think through new product launches by pulling together historical trends, category demand signals, and competitive data to map out expansion options and their tradeoffs. Early users have responded positively, said Westmoreland, citing examples of Amazon Sellers who found that the tool compressed hours of work into seconds, whether it was shifting around inventory or introducing a new product line. Sellers can access Canvas directly through Seller Central by opening Seller Assistant and typing a question or selecting a suggested prompt. The system handles the rest, assembling a personalized workspace based on their specific business data. Westmoreland noted that seller confidence in acting on AI-generated recommendations has been building steadily. Amazon says sellers currently act on Seller Assistant's recommendations roughly 90% of the time. "We see Canvas as the next step in the journey where we can provide additional context and visuals to help sellers have more context, thereby giving them more confidence in taking action," she said. She was also clear that human judgment remains central to the experience. "Like any good adviser, the AI makes recommendations and then the seller takes the final action," she said. "They make the final decision." Over time, Amazon says Seller Assistant will take on more execution tasks on behalf of sellers, such as updating prices or creating restock orders, making the path from insight to action more automated. Canvas is available now to all sellers in the U.S. and U.K. through Seller Central at no additional cost. The initial rollout focuses on sales performance analysis, with Amazon planning to add capabilities for inventory management, marketing optimization, and product launch planning in the months ahead. The company also expects to expand Canvas to additional countries and make it available in languages beyond English later this year.