The Agentic Post
Breaking
Digital Twins and Physical AI  Â·  Humanoid Robots in Manufacturing  Â·  AI Data Centers and Water Usage  Â·  The AI Chip Supply Chain, Explained  Â·  What Is Fine-Tuning? A Plain Explainer  Â·  Meta and Sierra Want to Give AI Agents a Front Door to Stores  ·  
Home/Business/Enterprise Adoption
Amazon Opens Seller Central to Outside AI Agents

Amazon Opens Seller Central to Outside AI Agents

Enterprise Adoption

Amazon launched a Selling Partner plugin letting sellers manage inventory, pricing and listings from inside Anthropic's Claude or Amazon Quick, opening Seller Central to outside AI agents for the first time.

Amazon opened its Seller Central tools to outside AI agents at its Accelerate conference on September 23, 2026, starting with Anthropic’s Claude. A new Selling Partner plugin lets sellers manage inventory, pricing, listings and analytics from inside Claude or Amazon’s own Quick assistant, without opening Seller Central at all. The plugin is in beta for US sellers, with international expansion to follow.

What changed, exactly?

Three things shipped together. Seller Assistant gained persistent memory of each seller’s pricing patterns, inventory cycles and growth goals, and that memory carries across into Quick and Claude. It gained always-on workflows that run continuously in the background, monitoring conditions even when the seller is not logged in, rather than waiting to be asked. And it gained the Selling Partner plugin, which exposes that intelligence to outside AI tools.

Seller Assistant itself is not new. Amazon launched it in 2023 and it already runs on Claude models through Amazon Bedrock. Amazon says it has reached more than 90 percent of selling partners worldwide in their native language, has hundreds of thousands of active users, and that sellers accept its recommendations more than 90 percent of the time. Seller data used by Seller Assistant stays inside Amazon’s infrastructure and is not shared outside Amazon, the company said.

Eric Burns, field CTO at Anthropic, framed the seller-side value directly: the plugin brings Amazon’s data-driven insights into a seller’s Claude account alongside their existing accounting and supplier data connections.

Why is this a notable shift for Amazon?

Amazon has historically kept sellers inside its own interface. Letting them run their business from a third-party assistant loosens that grip, though Amazon still decides which platforms get supported. The contrast is sharp on the consumer side: days before this announcement, Amazon blocked Meta’s Muse, an AI agent that shops on behalf of consumers, from its store. Amazon’s position is that outside agents must identify themselves and follow the rules of the sites they use.

So the door is open for agents that help sellers spend money with Amazon, and closed for agents that help buyers shop across it. That is a coherent commercial position rather than a contradiction, but it is worth naming plainly. Amazon is also a major investor in Anthropic, which makes Claude the natural first partner.

What do sellers get right now?

  • Selling Partner plugin in beta for US stores, working with Claude and Amazon Quick
  • A canvas feature, a visual workspace generated from live seller data, for comparing pricing strategies or diagnosing ad campaigns
  • Free 12-month Amazon Quick Plus subscription for every primary account holder globally, plus two co-workers, available through December 31, 2026

Amazon says roughly 90 percent of sellers already use outside AI to run parts of their operations, which is the real argument for the plugin. The work was happening anyway, just with sellers copying data between tools by hand.

The wider pattern

This is the enterprise version of a shift we have tracked across the year: platforms rebuilding so that agents, not humans, are the primary way business logic gets executed. It is the same direction as agents moving into customer support and the reason MCP-style connections have become infrastructure rather than a novelty. Worth noting the backdrop: seller fees brought Amazon 46.8 billion dollars in the second quarter, and those fees are among the issues in the FTC’s antitrust case against Amazon, set for trial in March 2027.

See Amazon’s own announcement and GeekWire’s report from Accelerate.

Up Next
Three AI Labs Are Building Their Own Regulator

Three AI Labs Are Building Their Own Regulator

Big Tech

Google, OpenAI and Anthropic have approached former White House AI advisor Sriram Krishnan to lead a proposed self-regulator modeled on FINRA, tentatively called the Frontier AI Standards Agency.

Google, OpenAI and Anthropic are building their own regulator. According to The Information, the three labs have approached Sriram Krishnan, the former Senior White House Policy Advisor on Artificial Intelligence, to serve as chief executive of a body tentatively called the Frontier AI Standards Agency. The name is not fixed: The Information also renders it as the Standards Authority for Frontier AI. The model is FINRA, Wall Street’s industry-funded self-regulator.

Why is the choice of CEO the striking part?

Because Krishnan spent his time in government arguing against exactly this kind of institution. He served in the second Trump administration from January 2025 until June 2026. On the way out he said "there will not be an FDA for AI," arguing that a centralised agency requiring "a team of lawyers before you can get a model out" would put "sand in the gears" of the AI revolution.

The three biggest frontier labs want to build a regulator, and they want it run by someone on record opposing government regulators for AI. Read generously, that is a signal the body will be pragmatic rather than obstructive. Read sceptically, it is a signal about how much teeth it is meant to have.

What would the agency actually do?

The stated purpose is developing common standards for testing and safety across advanced AI systems, operating independently to fill gaps left by government oversight. Rather than each lab running its own evaluation approach, a shared framework would let governments, businesses and researchers compare how models were tested. OpenAI Chief Global Affairs Officer Chris Lehane confirmed in a Washington briefing on September 15 that the three labs had been coordinating on safety protocols for several weeks.

This is not the first attempt at industry coordination. Google, Microsoft, OpenAI and Anthropic launched the Frontier Model Forum in July 2023 with broadly similar aims. What is different here is the FINRA framing, which implies a body with membership, verifier status and some enforcement function rather than a research consortium.

What has not been answered

  • No founding charter, budget, or membership list has been published
  • No stated powers to actually halt a model launch
  • Who funds it, and whether verifier status requires paying the founding three
  • What formal role, if any, NIST or the US National Labs would have in the testing pipeline
  • Whether Anthropic or Google publicly confirm participation on their own terms, rather than the story landing through OpenAI-adjacent channels

Krishnan has not publicly accepted. His answer, plus a founding charter or CEO announcement before year end, is what would distinguish a live self-regulator from a trial balloon.

Why the timing is not subtle

This surfaced the same week Amodei and Altman told the UN Security Council that the industry urgently needs global oversight, and the same week Australia disclosed that an OpenAI agent breached a government health portal and took three months to report it. Announcing a self-regulator while asking the UN for regulation is a coherent strategy if you want to shape the rules before someone else writes them. Whether a body funded by the three companies it oversees can credibly discipline them is the question that will follow this wherever it goes.

Original reporting by The Information.