Ross Walker is the Director of Retail Media at Acadia.
Meta: Amazon just moved Sponsored Products Prompts to general availability and started charging for them. Here's what that means, why it matters, and what you should do about it right now.
-
What if 30% of Amazon's ad spend flowed through Rufus prompts in the next 6 months?
Amazon didn’t make a big deal about this. But on March 25, 2026, something pretty significant happened: Sponsored Product Prompts and Sponsored Brand Prompts officially moved from beta to general availability in the U.S.
If you take a step back, this isn’t just another incremental ad product update. Our read is that this is early infrastructure for how Amazon wants discovery and advertising to work in a more AI-driven environment.
If you advertise on Amazon, this affects you. Starting now.
Here's what's actually happening, why it matters, and what brands should be thinking about as this rolls out at scale.
How This Actually Started
The first indication that something was happening here was pretty easy to miss.
A new sub-tab appeared inside the Sponsored Products workflow in Campaign Manager. Initially, it was completely blank. No data, no explanation, just clearly a placeholder for something that hadn’t been rolled out yet.
Then, over time, data started to populate.
First impressions, then clicks, then sales. At that stage, there was no spend associated with it, which is usually a signal that Amazon is testing the mechanics before fully turning on monetization. You could see attributed revenue without any cost against it.
Now, CPC billing is live. And more importantly, there’s now a standalone report where you can pull prompt performance across campaigns, not just within a single campaign view.
This Is a Placement, Not an Ad Type
The natural instinct is to try to categorize this within the existing structure: Sponsored Products, Sponsored Brands, Sponsored Display.
We don’t think that’s the right way to think about it.
The better framing is that this is a net-new placement layer that sits on top of those existing ad types. It’s much closer to how Sponsored Products expanded over time (search, then product detail pages, then featured placements, then off-Amazon) rather than being a distinct format.
What Amazon is doing here is adding a new surface where ads can appear, not redefining the ads themselves.
That distinction matters because it suggests this won’t stay isolated to one campaign type. Over time, it likely becomes a shared placement across the ecosystem.
What Makes Prompts Different
If you look at the prompts Amazon is surfacing, they tend to be framed as questions or comparisons. Things like “what’s the best option for [use case]?” or “which product has the best [feature]?” That’s a different layer of intent than a standard search term.
In a typical Sponsored Products environment, you’re matching a product to a keyword. The relationship is fairly direct: a user searches for something, and your ad shows up because you’re bidding on that query.
With prompts, there’s an extra step in between.
Amazon is interpreting what the user is trying to accomplish, generating a structured prompt around that intent, and then inserting products into that response. So instead of just matching to a query, your product is effectively being positioned as a potential answer.
A prompt placement can still drive to a specific product, so in that sense, it behaves like Sponsored Products. But at the same time, many of these prompts are implicitly comparing options or evaluating brands, which starts to look more like Sponsored Brands.
And layered on top of that, the overall experience feels closer to a recommendation engine than a traditional ad slot.
So you end up with a single placement that is doing multiple jobs at once: product discovery, brand positioning, and guided evaluation.
That’s what makes it difficult to neatly categorize, and also what makes it potentially more valuable. It’s not just inserting your product into a list of results. It’s placing it inside the logic of how Amazon is helping the user make a decision.
Where These Prompts Are Showing Up
In terms of where this actually lives in the experience, it’s starting to show up in a few different places.
You’re seeing variations of it in:
- Autocomplete-style suggestions under the search bar
- Rufus-generated prompt recommendations
- The Rufus results module that appears within search
Within those Rufus results, there’s already a mix of organic and sponsored placements. They’re not always explicitly labeled in a way that makes that distinction obvious, but functionally, that’s what’s happening.
Our expectation is that this expands across multiple surfaces as Amazon continues to push prompt-based interaction more heavily.
The Larger Shift Amazon Is Driving
The more important point here is what Amazon is trying to do from a behavioral standpoint.
They’re not just introducing a new placement; they’re trying to reorient how people search on the platform.
Instead of a purely keyword-driven interaction, the push is toward more conversational, prompt-based discovery. That’s a meaningful change. Users are very ingrained in how they currently search, and shifting that behavior takes time.
Right now, if you look at actual usage, most people still default to typing a query and hitting enter. But the interface is increasingly surfacing prompts as an alternative path, and over time, that tends to shape behavior.
So part of the uncertainty here isn’t about the ad product itself, it’s about how quickly users adopt the interaction model that supports it.
What You Can and Can’t Control Right Now
From an advertiser's perspective, the current state is fairly constrained.
You can see which prompts you’re showing up for. You can measure performance. But you don’t have the ability to directly target or exclude specific prompts, and there’s no prompt-level bidding mechanism yet.
Everything is being auto-curated by Amazon.
There is a workaround where you can take the prompt data and bid on similar long-tail queries in your keyword campaigns. But that’s not a one-to-one relationship. Showing up for a keyword doesn’t guarantee placement within a prompt result, because those are being generated and selected differently.
This will evolve. The likely path is some form of bidding or prioritization within a defined set of prompts, rather than completely open-ended targeting. But that hasn’t materialized yet.
What the Early Data Suggests
Even with limited data, there are some early directional signals.
If you think about the economy of on-site Amazon advertising, top-of-search has always been the most coveted and valued placement. It’s hyper above-the-fold. It’s the first thing people see. It commands premium CPCs from both a Sponsored Brands and Sponsored Products perspective.
Now there’s a new inventory surface in that same above-the-fold real estate (Rufus prompt results), and advertisers are going to find themselves clamoring for it as consumer adoption grows.
Engagement metrics, particularly click-through rate, are trending closer to top-of-search placements than standard search results. At the same time, conversion rates are lower, which aligns with the idea that users engaging with prompts are often earlier in their decision-making process.
Inventory is also inherently limited. These prompt results have to be curated enough that they’re actually readable and useful, which means you’re not going to see the same scale as traditional search placements.
All of that points toward a relatively constrained, high-value inventory environment, which typically leads to higher CPCs over time.
How to Think About This in the Funnel
There’s a bit of tension in how to position this from a funnel perspective.
On one hand, some shoppers use Rufus at the top of the funnel to figure out what their options are. Others use it at the bottom, to resolve objections, clarify value propositions, or get a final nudge toward purchase. So in raw usage terms, Rufus is full-funnel.
On the other hand, Amazon’s own KPI reporting tells a different story. Amazon has put out data showing a significant positive impact on conversion rate for Rufus-assisted shopping journeys. They’re not marketing this as an awareness vehicle.
The cleanest way to reconcile that is: This is behaviorally a full-funnel interaction, but from an optimization standpoint today, it’s closer to bottom-of-funnel.
That may change as usage evolves, and we think the real clarity here is going to come from Amazon Marketing Cloud once prompt interactions start showing up in path-to-conversion analysis.
That’s the data that should drive bidding decisions. If I see high traffic volume on a particular auto-curated prompt, and AMC shows me that the prompt is consistently three to five steps upstream from conversion, I might still bid up heavily on it, knowing the attribution is going to lag. If I see a prompt that’s consistently the last touch before purchase, I’m bidding on it aggressively from day one.
What Brands Should Be Doing Now
Even without full control, there are a few things that are worth doing immediately.
The first is simply pulling the Prompts report and understanding where you’re showing up. That’s foundational. Most brands aren’t doing this yet, and there’s an advantage to building that baseline early.
The second is treating prompt language as a signal. The way Amazon is framing your products in these prompts is effectively a reflection of how its AI understands your catalog. If you’re consistently appearing in certain types of queries (feature-driven, use-case-driven), that language should be reinforced in your product detail pages.
The third is making sure you’re set up in Amazon Marketing Cloud. Once prompt interactions are incorporated into pathing data, that’s where the real strategic insight is going to come from.
Where This Is Likely Going
There are two timelines to consider.
On the product side, things like bidding controls, improved reporting, and better transparency are relatively near-term. Those are solvable problems, and Amazon has a clear incentive to build them.
On the behavior side, it’s slower. Shifting users toward prompt-driven interaction is not immediate, and it depends heavily on how aggressively Amazon pushes that experience in the UI.
But at a broader level, there’s fairly strong alignment across the industry that AI-driven, agentic discovery is where commerce is heading. Prompts are Amazon’s early implementation of that.
Our expectation is that within six to twelve months, this becomes a core part of search strategy, not something experimental or optional, but a standard placement that requires its own approach to budgeting and optimization.
The Bottom Line
We’re in the very early stages of understanding what sponsored prompts are going to look like at full deployment.
The control is limited, the data is sparse, and the consumer behavior shift is still in progress. But those aren’t reasons to wait. There are reasons to get in now, establish the baseline, and be positioned to move fast when more optimization levers arrive.
Six months from now, the brands that started pulling the Prompts report in April are going to look very different from the ones that didn’t.
We're watching this closely, running analysis across our client base, and will continue sharing what we're seeing. If you want to dig into what your current prompt exposure looks like and what it means for your Amazon strategy, reach out to your Acadia team.
Read more from this author: