AgentXray Blog

Amazon Ads Agent: What It Answers, and the Profit Question It Leaves to You

Amazon Ads Agent: What It Answers, and the Profit Question It Leaves to You

TL;DR

Amazon Ads Agent puts an AI assistant inside the ads console. What Amazon says it does, what its announcement does not mention, and how to judge its suggestions by profit.

✨ Written with AI assistance from Amazon's own announcement, and reviewed by Avanta Global EOOD before publication. Learn more

Short answer: Amazon Ads Agent is the new name of Amazon's unified advertising platform, with a conversational assistant that plans, targets, analyses and can apply its recommendations in one click. Amazon's announcement talks about reach, unique customers and cost per impression. It does not mention product cost, Amazon's fees or profit. Use it to run campaigns, and keep a separate answer to the question "does this make money after every cost?"

What Amazon announced

On 29 September 2026, at unBoxed, Amazon Ads introduced Amazon Ads Agent: one AI-first platform for sponsored ads, display, video and audio, and full-funnel campaigns (Amazon Ads). The assistant inside it covers four jobs:

  • Media planning: audience size, reach and budget allocation across channels.
  • Targeting: describe an audience in plain language and get product, category, keyword and audience recommendations, "ready to apply to campaigns".
  • Analytics: ask about performance and trends in plain language and get charts back.
  • Sponsored ads guidance: targeting, bid and budget recommendations you can "apply in one click".

Amazon says it is "backed by trillions of shopping, browsing, and streaming signals", and that the capabilities "will roll out over the coming months". Full-funnel campaigns are available to all advertisers in the United States now.

What it is built to answer

Read the announcement for the numbers Amazon chose to highlight. Advertisers using its natural-language targeting saw "more than 25% additional unique customers" and cost per impression lower by more than 10%. Those are reach and efficiency results, and they are useful ones: more of the right shoppers, at a lower price per view.

For a sponsored products seller, that means the assistant is strong at questions like which keywords should I add, where is my budget running out, and which targets have the best ACoS.

The question the announcement does not mention

The announcement does not mention product cost, Amazon's fees, or profit. As far as we can tell, the assistant works from the advertising console's data, and your product cost and settlement fees do not live there. We have not used it on a live account yet, so treat this as a reading of what Amazon published, not a test.

That matters because of one property of ACoS: your break-even ACoS is set by your price, your product cost and your fees, not by your ads. An assistant that judges a keyword by ACoS needs to know that line to tell you whether the keyword makes money.

A worked example (illustrative numbers)

Take an example product that sells for $60, costs $24 to make and ship in, and pays Amazon an example $24 in fees (a 15% referral fee plus $15 for fulfilment). Before advertising it leaves $12, so its break-even ACoS is 12 ÷ 60 = 20%. Any ad sale at an ACoS under 20% makes money.

Now run a deal at $54. Product cost and the fulfilment fee stay the same; the referral fee falls to $8.10. Profit before ads drops to $6.90, and break-even ACoS to about 13%.

A target ACoS of 18% was profitable at $60 and loses money at $54. Nothing in the ads report changed. The ACoS can even improve during the deal, because a lower price converts better, while every ad-driven sale loses money. An assistant that only sees the ads console sees the better ACoS.

Use both, judge by profit

Amazon's assistant can run the campaigns. What it needs next to it is a profit ledger that knows your break-even and checks afterwards whether a change worked:

  1. Know your break-even ACoS per product, at today's price. It moves every time the price or a fee moves. Our longer write-up on ACoS vs profit shows how a period with a better ACoS can earn less.
  2. Price every recommendation against it, whoever made it: you, an agency or Amazon's assistant.
  3. Judge a change after it settles. Look at profit after all ad spend two weeks later, not at the ACoS on the day you applied it.

This is the part AgentXray does. It reads profit from your own settlement report and your product costs, prices each PPC recommendation against your break-even, and follows every approved change for 14 days before judging it on net contribution. Your own AI (Claude or ChatGPT) can ask it the same questions over MCP. You can try the demo store without connecting an account, or see how it compares with other Amazon MCP servers.

FAQ

What is Amazon Ads Agent? The new name of Amazon's unified, AI-first advertising platform, announced at unBoxed on 29 September 2026. It brings sponsored ads, display, video and audio, and full-funnel campaigns into one place, with a conversational assistant for planning, targeting, analytics and recommendations.

Can Amazon Ads Agent change my campaigns? Amazon says its recommendations are ready to apply and some can be applied in one click. Treat an applied recommendation like any other change: check it against your break-even and look at the profit two weeks later.

Does Amazon Ads Agent know my profit? Amazon's announcement does not mention product cost, fees or profit. Your product cost is not part of Amazon's advertising data, so a profit answer needs a source that has it.

Do I have to choose between Amazon Ads Agent and AgentXray? No. Amazon's assistant runs and tunes campaigns inside Amazon. AgentXray measures what those campaigns earned after every cost. They answer different questions.

About this article

This article was researched and drafted with AI assistance. Before publication, it passed automated editorial review against Avanta Global EOOD's published editorial standards (factual accuracy, source attribution, voice & readability). Our editorial standards page documents exactly what we check. We continuously monitor published content for accuracy and update articles when new information emerges. Learn more about our editorial process and the team behind AgentXray.