At unBoxed 2026 Amazon folded its ads console and DSP into one agent and opened Full-Funnel Campaigns to every US advertiser. Here is what launched and how to test it without handing over your budget.
The short version: At unBoxed 2026 (September 28 to 30, San Francisco), Amazon folded its ads console and DSP into one conversational tool called Amazon Ads Agent. It opened Full-Funnel Campaigns, which put one goal and one budget behind sponsored ads, display, video, streaming TV and audio, to all US advertisers.
It also introduced ad formats built for Alexa for Shopping and a paid Review Requests product. The opportunity is real. So is the risk of handing your budget to automation before you know how it behaves. Below is what launched and a test plan with a guardrail for each.
Amazon Ads Agent is a single conversational buying tool that replaces the separate ads console and DSP experiences, covering Sponsored Ads, DVA+ and Full-Funnel Campaigns. Advertisers plan, build audiences, analyze and optimize in plain language, with both manual and automated controls, according to Amazon. The conversational experience is in expanded closed beta.
The DSP name is going away. Programmatic buying is not. If your team has spent years learning two interfaces, the next six months are about learning one and making sure nothing you depend on got lost in the move.
Sponsored Ads, DVA+ and Full-Funnel Campaigns now run through one conversational tool. The DSP name is retiring; programmatic buying moves inside the agent.
The numbers are promising and unaudited: in beta, Amazon reports 67 percent higher long-term ROAS and 29 percent lower cost to acquire new-to-brand customers (Amazon-reported beta results). Full-Funnel Campaigns, now available to all US advertisers, let you set one goal and one budget, and Amazon's AI handles the channel mix across sponsored ads, display, video, streaming TV and audio, plus creative and audiences.
Those are strong numbers. They are also Amazon grading its own homework, measured with Amazon's own long-term sales metric. That does not make them wrong. It means you test them on your account before you move real money.
Strong numbers, graded by Amazon on its own long-term sales metric. Test them against your own baseline before you move real money.
Amazon announced six other changes worth a brand operator's attention:
Test one tool at a time, on a defined slice of spend, against a control, with a stop rule written down before launch. Here is the plan we are running across our accounts this quarter.
| Tool | Test scope | Guardrail | Success metric | Read after |
|---|---|---|---|---|
| Full-Funnel Campaigns | One hero brand, 10 to 15% of its monthly ad budget | Pause if TACoS rises 2 points over the control for two straight weeks | New-to-brand cost and long-term sales vs. current mix | 6 weeks |
| Ads Agent (conversational) | Planning and reporting only | No budget or bid changes without human approval | Hours saved on weekly reporting | 4 weeks |
| AI creative | Fill gaps on SKUs with no lifestyle or video assets | Brand review before any asset goes live | Coverage and CTR on those SKUs | 4 weeks |
| Review Requests | ASINs under 1,000 ratings with a clean product record | No follow-up messaging that steers buyers | Rating trend, not just review count | 6 weeks |
| Branded Conversations | Draft briefs now, launch when available | Claims in briefs must match the listing | Conversion on conversation-driven sessions | At launch |
Because the job of the ads manager has changed. When an agent can build, launch and optimize a campaign from one sentence, the skill that matters is no longer clicking through a console. It is knowing your margins, setting limits the agent cannot cross, and spotting when it optimizes for the wrong thing.
An agent told to maximize ROAS will happily spend on shoppers who were going to buy anyway. An agent told to grow new-to-brand will happily buy expensive reach. Neither is wrong. Both need someone who knows the brand's contribution margin by SKU to say which one matters this quarter.
Start with four moves:
Access, baseline, list, brief. The first three take an afternoon; the brief is the one that keeps paying off.
We run ads, catalog and creative as one system with senior operators on every account. We treat every new Amazon tool the same way. Test small, measure against our own baseline, then scale what earns it. unBoxed handed brands a lot of automation. We are here to make sure it works for your margin, not just for the dashboard.
Want the test plan above built for your account, with your baseline and your guardrails? Book a meeting for a free Marketplace Assessment and Account Audit. A senior TopRank operator reviews your account, and you leave with a written priority list.
Sources: Amazon Ads, "unBoxed 2026 news and announcements" (advertising.amazon.com/library/news/unboxed-2026-news-announcements); Amazon Ads, unBoxed event page (advertising.amazon.com/events/unboxed); Amazon Ads, "Get product reviews faster with Review Requests" (advertising.amazon.com/resources/whats-new). Secondary coverage: AdExchanger, MediaPost and Adweek, September 29, 2026.
About the author: Josh Phillippi is Founder and CEO of TopRank Partners, a performance-based Amazon and Walmart marketplace agency with no annual contracts.
Book a meeting for a free Marketplace Assessment and Account Audit. A senior TopRank operator reviews your account, and you leave with a written priority list.
Book your free assessment