On 29 September 2026 Amazon renamed its advertising platform Amazon Ads Agent and reduced it to three campaign types: sponsored ads, Full-Funnel Campaigns and DVA+. Full-Funnel Campaigns is already live for every advertiser in the United States and hands planning, channel mix and ongoing optimisation to Amazon's AI, while DVA+ begins rolling out in late October. For a seller running their own account, the question is not whether to adopt this — it is how much of your fourth-quarter budget you are willing to stop steering while the quarter is running.

What Amazon actually announced

On 29 September 2026, at its unBoxed conference in San Francisco, Amazon Ads published a post titled "Introducing Amazon Ads Agent, your AI-powered ad platform". The structural sentence in it is this one: "Amazon Ads Agent brings three campaign types—sponsored ads, Full-Funnel Campaigns, and Display, Video, and Audio—together in one simple, connected, AI-first ads platform." Display, Video, and Audio is the one Amazon shortens to DVA+.

This is the second half of a change that started a year ago. The same post notes that "Last year at unBoxed, Amazon Ads announced the unification of the demand-side platform and ads console into a single media buying experience." The 2026 announcement puts a name on that merged platform and collapses what used to be a long menu of campaign types down to three. Trade coverage of the event describes Amazon Ads Agent as replacing the former ads console and Amazon DSP experience with a single platform that includes a conversational chat layer.

Two availability facts matter more than the branding, and both are stated plainly by Amazon. "Full-Funnel Campaigns is now available to all advertisers in the United States." And "DVA+ will begin rolling out to advertisers beginning in late October." So one of the three is live for you today if you sell in the US, and the second arrives within weeks of this being published.

What Full-Funnel Campaigns asks you to hand over

The division of labour is not ambiguous. Amazon's own description: "Advertisers provide their budget, products, and creative, and Amazon's AI handles the rest on their behalf—planning, executing, and continuously optimizing campaigns." A Full-Funnel campaign is designed to run from premium streaming TV and audio at the top, through display and online video on Amazon and the wider internet, down to sponsored ads at the point of purchase — all inside one campaign, one objective and one budget.

Read that against what you actually do every week in a manually run account. You choose the match types. You read the search term report and add the negatives. You decide that a keyword converting at 2% does not deserve a top-of-search bid adjustment. None of those levers exist inside a campaign where planning and channel mix are the AI's job. You are not tuning a campaign any more; you are setting an allowance and a goal.

What you get back is a single reporting view. Amazon says: "In reporting, advertisers have a unified view of performance across formats, with standard metrics like impressions, clicks, and sales alongside new-to-brand and long-term sales metrics, giving clear visibility into campaign performance." Useful, and genuinely hard to assemble yourself across four ad products. But notice the level it operates at. Nothing in the announcement describes search-term-level visibility or control inside a Full-Funnel campaign, and you should not assume a capability that has not been stated.

The important counterweight, also in Amazon's words: "Throughout Amazon Ads Agent, advertisers have the option to use manual campaign building tools, let Amazon's AI optimize on their behalf, or employ a mix of both depending on campaign goals and advertiser needs." The manual path is not being taken away. Sponsored ads remains one of the three campaign types. This is an addition to the menu, not a forced migration — which is exactly why there is no reason to rush.

The results Amazon quoted are Amazon's results

Amazon's unBoxed round-up reports that beta testers of Full-Funnel Campaigns saw "67% higher long-term return on ad spend and 29% lower cost to acquire new-to-brand customers". The same post carries a named advertiser, Canine Naturals, saying that "41% of sales from Full-Funnel Campaigns were new to our brand".

Those are real published claims and we are quoting them accurately, but they are a vendor describing its own beta. No sample size, no category breakdown, no baseline definition and no methodology has been published alongside them, and "long-term return on ad spend" is a metric Amazon defines rather than one you can independently reconstruct from your own reports. Self-selected beta participants are also usually the advertisers with the budget and the creative assets to make an upper-funnel campaign work in the first place.

The practical rule: do not put those percentages into a plan. If you test Full-Funnel Campaigns, define success before you launch using numbers from your own account — your current blended TACOS, your current new-to-brand share, your actual contribution margin — and judge the test against those. A 67% uplift quoted by the platform selling you the campaign is a reason to be curious, not a forecast.

DVA+ lands in late October, which is the worst fortnight of the year to restructure

DVA+ consolidates what are currently separate buys — Sponsored Display, Sponsored TV and programmatic display, video and audio — into one campaign type with one objective and one budget. Amazon's framing is that it "simplifies campaign creation and optimization for advertisers of all sizes" while keeping an advanced path: with "advanced settings", experienced programmatic buyers have full control. Event coverage describes the same thing as a simplified mode and an advanced mode sitting side by side.

The timing is the problem, not the product. A new campaign type appearing in your console in late October arrives in the middle of the highest-traffic, highest-CPC weeks of your year, when your historical data is least representative and a structural change is hardest to read. Any campaign you launch in that window has to learn on your most expensive traffic, and any dip you see afterwards will be impossible to attribute cleanly between the new structure, peak-season competition and your own inventory position.

So treat the rollout as information rather than an instruction. When DVA+ appears, open it, look at how your existing Sponsored Display and Sponsored TV activity is represented, screenshot or export your current structure and settings, and then leave it alone. Migrating display spend into a new unified campaign type is a January project. Peak is for running the thing you already understand.

The transparency question is the sensible one to ask

The trade press at the event put the uncomfortable question better than any vendor material would. Digiday's write-up frames the appeal as less operational complexity, and the accompanying question as how much visibility advertisers keep as Amazon's technology takes on more of the decisions. An Amazon executive quoted in that piece said: "We are figuring out how to give them more transparency on the audience side." The same report notes Amazon did not commit to disclosing how individual winning auction prices are determined.

That is not a reason to refuse the product. Automated bidding has been the default in paid search for years and plenty of accounts are better for it. It is a reason to be precise about what you are buying: you are buying outcomes measured with Amazon's metrics, inside an auction whose pricing logic is not disclosed, optimised toward a goal you set once at the start. If that bargain is acceptable to you, the campaign type is a legitimate tool. If you cannot explain your results to yourself afterwards, you have no way to improve them.

The honest version of the trade-off is a labour question. Full-Funnel Campaigns removes a large amount of weekly work from your calendar. If the hours you currently spend on search term reports and bid adjustments are producing a better return than the AI would, automating them is a loss. If those hours are producing nothing — and in plenty of small accounts, poked-at-weekly manual campaigns genuinely do underperform — then handing them over is a gain. You can only know which by measuring the account you actually have.

The agent layer is coming for your reporting as well as your bids

Three smaller announcements from the same event point in one direction. The conversational Amazon Ads Agent experience — asking questions of your account in plain language rather than building a report — is described as available in expanded closed beta. Amazon also announced an Amazon Ads MCP Server (Lite), which it says "simplifies how agents interact with Amazon Ads", starting with a small set of core tools that an AI agent uses to discover further capabilities as it needs them. Alongside it, an AMC Expert Tool turns a plain-language question into an Amazon Marketing Cloud SQL query.

Taken together, the direction is that the mechanical skills of running Amazon ads — pulling the report, writing the query, bulk-editing the bids — are being commoditised faster than the judgement around them. That is not a threat if you are the person with the judgement. It is a threat if your entire process is mechanical.

The practical consequence for a seller is to stop valuing your own time by the reports you can assemble and start valuing it by the decisions only you can make: what margin you will accept on a given ASIN, which products are worth defending, when a listing problem is being misdiagnosed as a bidding problem. An AI that can answer any question about your account is not much use if nobody is choosing the right question.

What to do this week

First, go and look. If you sell in the US, Full-Funnel Campaigns is live in your account today, and a campaign type you have never opened is worth twenty minutes of reading before someone recommends it to you in a webinar. Note what it asks for, what it will not let you set, and what its reporting actually shows.

Second, protect the quarter. Do not move money that is currently working out of campaigns you can steer and into an automated campaign type mid-peak. If you want to test Full-Funnel Campaigns now, fund it with incremental budget you are willing to lose for the sake of learning, not by cutting Sponsored Products spend that is already converting. Write down, before you launch, the new-to-brand share and blended TACOS you would need to see to call the test a success.

Third, prepare for late October without acting on it. Export your current Sponsored Display and Sponsored TV structure, bids and budgets now, so that when DVA+ appears you have a record of what the account looked like before. Diary the actual migration decision for January, when your data is representative again.

Fourth, keep doing the one job none of the three campaign types does for you. Every one of them sends traffic to a detail page you own. An AI that plans a perfect media mix into a listing with a weak main image, thin bullets and unanswered objections will simply buy you expensive proof that the listing is the problem. In a quarter where traffic is at its most expensive, conversion rate is still where the leverage is.

Sources