Set up Amazon Seller Assistant workflows once, in plain English from Seller Central, and Amazon's upgraded AI agent keeps watching your prices, stock levels and account health around the clock — then acts on what it finds, even while you are logged out. That is the core of the announcement Amazon published in its newsroom this week (reported 23 September 2026 by outlets including Retail Gazette and The Star): Seller Assistant, the AI helper already available to sellers on the US marketplace, has been rebuilt as an agentic system that runs continuous, standing instructions rather than answering one question at a time. It is free, it is optional, and — per Amazon's announcement — it is live for all sellers in the US store now, with other countries following in the coming months.
📺 Watch: Automate Anything using the NEW AgentOS System!
🔥 Get the Agent OS as a free bonus: AI Profit Boardroom members get the full Agent OS zip, prompt libraries, daily tutorials and weekly live coaching calls. → Get inside
What Amazon Seller Assistant Workflows Actually Do
A workflow here is a standing instruction that stays active and fires when your business conditions change. Per Retail Gazette's report of the announcement, sellers can tell the agent in plain language to do things like flag when a product's ratings fall below a set level, monitor competitors' prices, or identify when stock needs replenishing — and the agent keeps executing those instructions without the seller being logged into Seller Central. Amazon's own newsroom post fills in the detail across five capability areas:
- Inventory and fulfilment. The agent proactively monitors stock, flags slow-moving products before long-term storage fees accrue, recommends price adjustments or removal, analyses demand patterns, prepares shipment recommendations and coordinates FBA fulfilment centre allocation, per the announcement.
- Account health. It continuously monitors account status, scans listings for compliance issues — Amazon's example is unapproved pesticide function claims — explains what triggered a problem, recommends fixes and implements the ones you approve.
- Compliance navigation. It alerts you to missing certifications and documentation and walks you through the requirements step by step.
- Advertising. A Creative Studio generates ad concepts from conversational prompts; Amazon cites one seller whose AI-generated ads produced a 338 percent increase in click-through rate and a 121 percent return on ad spend.
- Growth strategy. The agent analyses sales patterns and customer behaviour, recommends new product categories and builds plans for seasonal peaks, per the announcement.
Under the hood, Amazon says the system runs on Amazon Bedrock and uses both Amazon Nova and Anthropic Claude foundation models, combined with 25 years of seller-support data.
How Much Do Amazon Seller Assistant Workflows Cost?
Nothing beyond what you already pay to sell. Amazon's announcement states the agentic capabilities come at no additional cost, and Mary Beth Westmoreland, Amazon's Vice President of Worldwide Selling Partner Experience, is quoted confirming the service is available to sellers at no additional cost, remains optional, and lets sellers control how much information they share through it. For context on adoption, Amazon has said Seller Assistant has already reached more than 90 percent of its selling partners worldwide, counts hundreds of thousands of active users, and sees its recommendations accepted more than 90 percent of the time, according to the company's statements reported alongside the launch. Free plus already-deployed is an aggressive combination — it means the default Amazon seller in the US now has an always-on agent available whether or not they ever sought one out.
If you want to put AI agents to work in your own business the way Amazon is putting them to work for sellers — real systems, not theory — check out the AI Profit Boardroom and get the full agent playbook here. Want a personal roadmap instead? Book a free SEO strategy session and plan it 1-on-1 with Julian.
Guardrails, Audit Trails and Who Stays in Control
The obvious worry with an agent that touches live pricing and inventory is an automation running wild on your account. Amazon's answer, per the announcement and Retail Gazette's coverage, is a permission dial: you decide whether the agent only makes recommendations or takes autonomous action, every action it takes is recorded in an audit trail, and you can set your own limits and approval requirements before anything executes. The system also builds persistent memory of your pricing patterns, inventory cycles and growth goals, so recommendations become more personalised over time, per the same reports. Seller Alfred Mai, founder of ASM Games, is quoted in Amazon's announcement calling the assistant his own personal business consultant that answers immediately rather than requiring him to dig around. The sensible way to adopt it mirrors how you would deploy any agent: start in recommendations-only mode, audit what it suggests for a few weeks, then progressively grant autonomy on the tasks where its judgement holds up — the same staged-trust approach covered in the guide to running a fleet of AI agents without losing control.
The Claude Plugin: Your Seller Data Inside Other AI Tools
The second half of the announcement matters just as much for anyone building an AI-run business. Amazon launched a Selling Partner plugin that connects seller accounts to external AI tools through its Amazon Quick platform — currently in beta with Anthropic's Claude, per Amazon. The plugin pulls your listings, inventory levels, sales analytics and performance metrics into the outside assistant, which can then act on the account, with your approval, the way Seller Assistant does inside Seller Central. In practice that means the AI stack you already run — research agents, content agents, finance dashboards — can finally see your Amazon business directly instead of working from exported spreadsheets. If you are new to wiring tools together like this, the primer on AI agent orchestration for beginners explains the plumbing, and the Agent OS resource shows how to structure a whole business around connected agents rather than isolated chatbots.
📺 Watch: These GPT-6 Astra Automations Run My Business
What This Means If You Sell on Amazon
The money case is straightforward: the tasks these workflows absorb — repricing checks, restock timing, storage-fee triage, compliance chasing — are exactly the hours that eat most sellers' weeks without growing revenue. Handing them to a free agent converts admin time back into product research, sourcing and marketing time. The announcement's storage-fee example alone is a direct margin play: a workflow that flags slow movers before long-term storage fees accrue is money recovered on autopilot. Three practical moves follow from the announcement:
- Turn on the highest-certainty workflows first. Stock replenishment alerts and account-health monitoring have obvious right answers; let the agent prove itself there before it touches pricing.
- Write your standing instructions precisely. The system takes plain-language instructions with thresholds — ratings floors, price bands, stock levels — so vague instructions produce vague automation.
- Use the audit trail as a training feed. Reviewing what the agent did each week tells you which decisions to automate next, which is the fastest route to a genuinely hands-off operation.
If Amazon selling is one of several income streams you are automating, the round-up of the best free AI tools for beginners to make money in 2026 pairs well with this — Seller Assistant now effectively joins that free-tools list for anyone with an Amazon account.
The Bigger Trend: Always-On Agents Replacing Dashboards
Amazon shipping agentic workflows to hundreds of thousands of sellers is the clearest mainstream signal yet that the interface for running an online business is shifting from dashboards you check to agents that check things for you. The same week's news cycle reinforced it — per aiagentstore.ai's weekly round-up, Amazon's seller workflows landed alongside OpenAI adding voice-driven work agents to the ChatGPT mobile app and Alibaba launching its enterprise AgentCore platform, all inside a few days. For sellers, the competitive implication is blunt: when monitoring and reacting is free and automatic, the edge moves to whoever configures their agents best and feeds them the best strategy. That is a skills question, not a budget question. The full walkthrough of going from a single agent to a coordinated system lives in the guide to AI agent workflow automation from first agent to fleet, and if you are weighing up which AI models to trust with business-critical work, the Goldie Bench write-up covers how the leading agent brains compare in hands-on tests. For a look at the community where operators trade these configurations daily, there is an honest AI Profit Boardroom review on this site too.
The bottom line on Amazon Seller Assistant workflows: free, live in the US now, genuinely agentic rather than a rebranded chatbot, and conservative by default because you control every permission. If you sell on Amazon, switching it on in recommendations-only mode this week is close to a no-brainer; the sellers who learn to direct it well will simply out-operate the ones who do not.
If you want to build the kind of AI-run business that benefits from every launch like this — with the Agent OS zip, 1,000+ done-for-you workflows and five live coaching calls a week — check out the AI Profit Boardroom and join the operators inside. Or start with a conversation: book a free SEO strategy session and get your automation plan mapped for free.
