Amazon's Seller Assistant Goes Always-On, Acting On Pricing And Stock 24/7

The technical foundation behind the update combines several of Amazon's own AI investments. Seller Assistant runs on Amazon Bedrock and draws on both Amazon's own Nova models and Anthropic's Claude models, a combination Amazon describes as pairing leading foundation models with nearly 25 years of accumulated expertise from supporting sellers on its platform. That combination is what allows the system to move from simply monitoring account health and inventory levels toward actively developing strategic recommendations and taking authorised action, rather than only flagging information for a seller to interpret and act on themselves.

Several concrete use cases illustrate how that shift plays out in practice. When reviewing a seller's Fulfillment by Amazon inventory, Seller Assistant can proactively flag slow‑moving products before they start accumulating long‑term storage fees, then recommend whether the best course of action is to leave an item as is, mark it down, or remove it from inventory altogether. The system can also analyse demand patterns to prepare shipment and restocking recommendations, continuously monitor a seller's account for issues such as a rating drop or a listing that might violate newly introduced product safety regulations in a specific country, and automatically check that a seller's full product catalogue meets compliance requirements across every market where they sell, a task that becomes increasingly difficult to track manually as a seller's catalogue and geographic footprint grow.

Alongside the workflow upgrade, Amazon introduced a new plugin that extends Seller Assistant's underlying data and intelligence into tools sellers already use outside Seller Central, launching first with Amazon's own Amazon Quick product and, in beta, with Anthropic's Claude. That plugin lets sellers bring their Amazon sales, inventory and performance data directly into the AI tools they already rely on day to day, rather than requiring them to log into Seller Central specifically whenever they want to check or act on that data. Amazon is also offering a free 12‑month Quick Plus subscription to sellers adopting the new capabilities, intended to help them manage their business both on and off the Amazon platform from a single connected set of AI tools. Westmoreland has noted that roughly 90 percent of Amazon's selling partners already use some form of third‑party AI in running their business, a figure that appears to have shaped Amazon's decision to meet sellers inside tools they had already adopted rather than requiring all AI‑assisted work to happen inside Seller Central itself.

Amazon has also added persistent memory to Seller Assistant, letting it retain an understanding of each individual seller's business, including their specific products, preferences and pricing rules, over time, so that recommendations sharpen the longer a seller uses the tool rather than starting fresh with each interaction. Alfred Mai, founder and chief executive of ASM Games, described using Seller Assistant nearly daily since adopting it, saying it functions less like a generic support tool and more like a personal business consultant that already knows how to support his operations across FBA and sales reporting without him needing to dig around Seller Central to find what he needs.

According to Amazon, Seller Assistant already reaches more than 90 percent of its selling partners globally, and sellers accept its recommendations more than 90 percent of the time, figures the company has used to argue that trust in the system's suggestions is already well established heading into this expansion toward more autonomous execution. The tool remains free for any seller to use and entirely optional, and Amazon has said sellers retain control over how much of their personal and business data they choose to share with the agentic service.

The update reflects a broader pattern taking shape across large e‑commerce and enterprise platforms, as AI tools shift from answering questions to actively running parts of a business autonomously, provided sufficient guardrails and audit mechanisms are in place to keep a human able to review and, if necessary, reverse what the system has done. For Amazon specifically, embedding this level of agentic capability directly into Seller Assistant, and extending it into tools like Claude that sellers may already be using independently, positions the company to capture more of the AI‑assisted commerce workflow rather than ceding that ground to third‑party seller tools built on top of Amazon's own marketplace data. Whether sellers extend the same roughly 90 percent trust they have shown toward Seller Assistant's recommendations to fully autonomous, always‑on execution of pricing and inventory decisions, rather than keeping a human in the approval loop for the highest‑stakes actions, is likely to become clearer as adoption of the new workflows scales beyond this initial launch.