Amazon Put Sponsored Ads and Streaming TV Under One AI Budget. Media Planning Just Lost Its Walls.

Amazon Ads has opened Full-Funnel Campaigns to U.S. advertisers, letting one AI system allocate a single budget across sponsored ads, display, video and streaming TV.

That turns media buying from a set of channel decisions into one optimization problem spanning awareness through purchase. The decision now is whether the lift from shared signals is worth giving Amazon control over channel mix, audience selection and the metric used to judge success.

The capability is deployed as an open beta, not merely announced. Any U.S. advertiser that sells on Amazon can provide products, creative, a budget and an optional return-on-ad-spend threshold. Amazon’s system then plans, executes and continuously adjusts the campaign while existing campaigns can continue alongside it.

One campaign now owns the funnel

Traditional media plans separate demand creation from demand capture. Streaming television and online video build awareness; display keeps the brand visible; sponsored ads convert active shopping intent. Each channel has its own budget, settings, reporting and specialist.

Full-Funnel Campaigns removes those operational walls. Amazon says its AI selects formats, moves money across funnel stages in real time, uses the company’s shopping signals to find audiences and optimizes toward total sales. When assets are missing, built-in tools can generate images and television-ready video for advertiser approval.

This is more than a conversational assistant making recommendations. The new capability owns allocation. The marketer supplies the commercial boundary; the platform decides which combination of exposures should do the work.

Shared signals are the actual capability

The strategic promise is not fewer campaign forms. It is that one format can teach the others. A streaming-TV exposure can influence how sponsored advertising performs later, while purchase behavior can reshape who receives upper-funnel media next.

Amazon reports that beta advertisers achieved 67% higher Long-Term ROAS than the blended performance of the same formats configured independently, while the cost of acquiring a new-to-brand customer was 29% lower. Those figures come from Amazon internal data collected from June through August 2026; they are vendor-reported, not independently audited, and individual results can vary.

Amazon’s Long-Term Sales metric also estimates the future value of new-to-brand customer actions across 12 months. That gives the optimizer a longer horizon than immediate conversion—but it also means the platform’s estimate becomes part of the business case for giving the platform more budget.

The bargain includes less control

Digiday reported on September 29 that Full-Funnel Campaigns buyers can approve creative but cannot choose individual channels or specific audiences. Amazon provides format-level visibility and says it is developing greater audience transparency. Its separate DVA+ product offers advanced programmatic controls, but that product begins rolling out later in October and should not be confused with the Full-Funnel open beta.

This creates a structural tension. Amazon can observe shopping, browsing, streaming and transaction signals that few advertisers can reproduce. It also allocates the money, serves much of the media and supplies the measurement used to evaluate the result.

Efficiency therefore needs an outside check. A higher platform-reported return can reflect genuinely incremental demand, better capture of demand that already existed, a shift from another channel or a different estimate of future value. Without holdouts or independent measurement, those explanations can look identical in a dashboard.

Media teams move from buying to governing

As execution compresses, the marketer’s job moves upstream. Teams must define which customer is valuable, what contribution margin permits, which claims creative may make, how much concentration is acceptable and what evidence is required before automation scales.

Agencies face the same shift. Their advantage becomes less about operating separate buying interfaces and more about designing tests, challenging platform recommendations, protecting creative quality and translating business economics into constraints an optimizer can use.

OrionPilot’s relevance here is not replacing Amazon’s buying engine. It is designed to keep business context, strategy, creative decisions, execution, connected data and weekly learning in one operating sequence. That independent record becomes more important when a platform internalizes the media plan: the objective, approved promise and lesson learned should still belong to the business.

Watch the evidence outside Amazon

Businesses testing Full-Funnel Campaigns should compare it with campaigns planned independently, not only with their own prior averages. Track incremental new customers, contribution margin, repeat purchase, creative fatigue, off-Amazon effects and total media concentration. Preserve a holdout where practical, and document every major budget, creative and objective change.

The next credible evidence is independent performance across categories, stable results after the novelty period and clearer explanations of how audience and channel choices are made. Marketing Dive confirmed the product’s U.S. availability and broader rollout details, while Digiday identified the reduced controls and transparency questions buyers will need to test.

Amazon has not eliminated the funnel. It has turned the funnel into a product—and asked advertisers to decide whether one optimization engine should own it.