Amazon Seller Assistant Evolves Into AI Operating System

Amazon is repositioning its artificial intelligence offering for third-party sellers from a conversational assistant into what amounts to an operating system for running an e-commerce business.

Mary Beth Westmoreland, vice president of Worldwide Selling Partner Experience at Amazon, revealed in an interview with PYMNTS that the company has expanded Seller Assistant from a tool that answers merchant questions into a system that continuously monitors business conditions, reasons across multiple operational domains, and executes actions with permission.

"We've gone from a smaller set of capabilities to onboarding just hundreds of them," Westmoreland said Wednesday. The system now integrates inventory management, pricing strategy, advertising optimization, demand forecasting, and compliance monitoring into a unified intelligence layer.

Why it matters

This shift represents a fundamental change in how AI serves enterprise users. Rather than compressing individual tasks like writing product descriptions or summarizing data, Amazon is targeting managerial attention itself as the scarce resource. For millions of third-party sellers who generate more than half of Amazon's retail sales, this could transform AI from another dashboard they check into the infrastructure through which they operate their businesses.

From reactive answers to proactive operations

The architecture Westmoreland described moves beyond traditional enterprise software, which assumes humans will interpret signals and decide what to do next. Seller Assistant now builds what she called a "personalized understanding" of individual merchant businesses, including their goals, patterns, and operating preferences.

In one scenario, a homewares seller asks how to prepare for fourth-quarter demand. The AI identifies opportunity for a specific planter, then recommends changes to product descriptions, keywords, pricing, and promotions based on current demand, historical performance, and category conditions. The seller reviews scenarios through a dynamically generated visual interface before authorizing changes.

Merchants can also describe automations in plain language and leave them running continuously. A seller might instruct the system to monitor top products for rating declines and prepare responses, or watch for competitive openings in a category and shift advertising spending accordingly.

"We wanted to provide sellers the ability to be able to automate all of this so that it just works for them 24/7," Westmoreland said.

Surrendering the interface to own the infrastructure

Amazon is also introducing a selling partner plugin that makes its seller intelligence available through Amazon Quick and, in beta, Anthropic's Claude. The move acknowledges that roughly 90% of Amazon's selling partners already use third-party AI tools.

Rather than forcing merchants to work exclusively inside Seller Central, Amazon is positioning itself as the intelligence and execution layer underneath whatever interface sellers prefer. In one example, sneaker-accessories seller Proof Culture used an AI agent to identify an upcoming release, check inventory, recommend restocking, draft a purchase order, and prepare internal communications about expected demand.

The trust question

How much authority businesses will actually delegate to AI systems remains the critical question for enterprise agent adoption. Amazon's approach offers one answer: sellers choose what information the plugin can access, can revoke that access, and must explicitly approve consequential actions. Seller data remains within Amazon's infrastructure.

Westmoreland said the vast majority of selling partners now use Seller Assistant, and they accept more than 90% of its recommendations. That acceptance rate suggests the path toward greater AI autonomy may depend less on model intelligence than on how much authority merchants are willing to delegate.

The details were first reported by PYMNTS in an exclusive interview with Westmoreland.