Amazon has folded Rufus and Alexa+ into Alexa for Shopping, and your existing Sponsored Products and Sponsored Brands campaigns are already eligible to serve inside it. The lever is not a new ad format. It is whether your PDP and brand content are strong enough to win the conversation.
Your Amazon Ads Are Already in the Conversation: What Agentic Shopping Means for Advertisers
By Stephen Honight, Founder of Lmo7
Most brands are still thinking about Amazon’s AI shopping shift as something to plan for. The interesting part is that it has already happened to their ad campaigns, whether they noticed or not.
On 11 June Amazon Ads put out an advertiser-facing note about agentic shopping. The headline most people picked up was the consumer one. Amazon has merged Rufus, its shopping assistant, and Alexa+ into a single experience called Alexa for Shopping. Shoppers can now ask questions straight from the main Amazon search bar, build personalised shopping guides, pull dynamic product comparisons, see up to a year of price history, and hand off routine buying to the assistant.
The part that matters more for anyone running ads is quieter. Your existing Sponsored Products and Sponsored Brands campaigns are now automatically eligible to serve inside that conversational surface. No new format to buy. No new campaign to build. According to Amazon, it is already live.
That changes what the work actually is. Let me walk through the mechanism, then what I would do about it.
What Amazon actually announced
Strip out the launch language and there are a few concrete things in the announcement.
Sponsored ads can now surface inside Alexa for Shopping, recommending products and brands inside a natural conversation. Amazon says existing sponsored ads campaigns are automatically eligible to serve there.
Sponsored Products and Sponsored Brands can now run as prompts. These are conversational ad units that appear in shopping results and on product pages and open a conversation with Alexa for Shopping. Amazon describes the prompt as acting like a virtual product expert. Again, existing campaigns are automatically eligible to be optimised as a prompt.
There is also a new conversational ad experience on Alexa devices, which Amazon calls an industry first, and conversational tiles on eligible Echo Show and Prime Video surfaces. Advertisers already running sponsored ads are auto-enrolled into a lot of this.
Sitting underneath, Amazon has shipped an agentic tooling layer. A Creative Agent that builds ads in hours. An Ads Agent that compresses campaign setup and targeting from weeks into minutes. And an Amazon MCP server that lets external AI agents talk to the Amazon Ads API through prompts rather than custom code.
So the surface has expanded and the campaigns brands already run have been pulled into it by default.
The number that should make you sit up
Amazon shared two figures in the announcement that are worth holding onto, with one important caveat I will come back to.
The first. Nearly 20% of shoppers who interact with a prompt continue the conversation about that brand. The second. Adding prompts to a Sponsored Brands ad drove a 6% increase in conversions.
Those are Amazon’s blended numbers, not ours, and not a guarantee for any single brand. Treat them as direction, not a forecast. But the direction is the point. A prompt is not a banner that a shopper glances at. It is the start of a dialogue, and one in five people who engage carry that dialogue forward. That is a different kind of ad surface to the one most teams have been optimising for.
Here is where it gets interesting for the brands we work with. That continued conversation does not run on your ad copy. It runs on your content.
The mechanism most teams will miss
A normal Sponsored Products click sends a shopper to a product page and the job is mostly done. The page either converts or it doesn’t.
A prompt works differently. When a shopper engages with a Sponsored Products or Sponsored Brands prompt, they open a conversation with Alexa for Shopping. To answer the follow-up questions, the assistant pulls on the same content substrate that organic Rufus already reads. Your title. Your bullets. Your A+ content. Your brand store. Your Q&A. The structured attributes in the backend.
So the ad gets the shopper into the conversation. The content decides whether the conversation lands.
If your product page is thin, vague on what the product is for, missing the comparison points a shopper would ask about, or silent on the questions people actually have, the assistant has nothing good to work with. The shopper engaged with your prompt and the answer that came back was weak. That 20% continuation and that 6% lift do not accrue to a brand whose content cannot hold up its end of the conversation.
This is the same thing we have been saying about organic Rufus for a year, now showing up on the paid side. Paid prompts and organic answers draw on one content layer. You cannot buy your way out of a weak PDP. The ad buys you the opening. The content earns the recommendation.
My sense is that this is the bit most brands will get wrong in the first six months. They will see “automatically eligible” and assume the work is done. The eligibility is automatic. The quality in conversation is not.
What this looks like in practice
We have seen both sides of this play out across client work.
When we ran a 60-day AI search engagement with Trip Drinks, the gains did not come from spending more. They came from site signals, content upgrades on the product pages, and fixing the sources that the models lean on when they form an answer. Average position across the major models moved from seventh to third and AI referral traffic rose by about a third. The lesson that carries straight into this announcement is that the content work is what moves the recommendation, not the impression count.
On Amazon specifically, the Veloforte work was content at the listing level. Bullet matches, titles brought live, A+ content built out across the catalogue, ASINs improved scrape over scrape. That is exactly the substrate a prompt conversation now pulls from. A catalogue that has had that treatment is ready for the conversational surface. A catalogue that hasn’t is going to underperform in it, regardless of how the ad budget is set.
So the practical read is simple. The brands that have already done the boring content work are the ones positioned to benefit from the conversational ad surface on day one. The brands that haven’t have a clear, fundable job in front of them.
Quick wins and the long game
I always try to be honest about the two tracks here, because it is easy to oversell the fast fixes.
The quick wins are real and you can ship them in weeks. Make sure your Sponsored Products and Sponsored Brands campaigns are actually set up to run as prompts, not just left on default. Audit your top SKUs for prompt readiness, which means clear titles, fit language, comparison points and honest answers to the questions shoppers ask. Seed Q&A on the listings that drive the most revenue. Tighten the bullets so they read as answers, not keyword lists. Get the A+ content and brand store saying something a virtual product expert could actually quote.
The long game is harder and slower. Authority is still the ceiling on what any AI surface will cite and recommend, and that is built through reviews, third-party mentions, expert sources and category-relevant presence over months, not weeks. The content work makes you eligible to win the conversation. Authority is part of what tips the assistant towards recommending you over a competitor when the shopper is comparing. Anyone telling you a content sprint alone fixes AI visibility is selling you half the picture.
For this announcement specifically, the quick wins are where the immediate return sits, because the ad eligibility is already switched on and the content is the bottleneck. The authority track runs in parallel and matters more the more competitive your category is.
A note on the agentic tooling
The Creative Agent, the Ads Agent and the Amazon MCP server are worth a flag even though they are not the main event for most brands yet.
The direction of travel is that campaign building and creative production are getting faster and more automated. That is good. It frees up time. But it also means the differentiator stops being who can stand up a campaign quickest, because soon everyone can. The differentiator becomes the quality of the inputs. Better product data, sharper content, cleaner claims, a catalogue an agent can actually understand and sell. The tooling commoditises the setup. It does not commoditise the substance.
If you run a multi-brand portfolio, the MCP server angle is one to watch, because it opens the door to connecting your own AI tooling to the Amazon Ads API without a custom build. That is a capability conversation for a later date, not a this-week action for most teams.
So what should you do next
The honest answer depends on the size and shape of your business, so here is how I would route it.
If you run Amazon ads at any real scale and you have a brand registry, the immediate move is an audit of your top SKUs for prompt readiness, then a check that your Sponsored Products and Sponsored Brands campaigns are configured to serve as prompts rather than sitting on defaults. This is the Challenger Amazon Stack work, Ads at one thousand pounds a month paired with Content Optimisation at seven hundred and fifty, because the ads and the content have to move together for this to pay off. Doing one without the other leaves money on the table.
If you are a founder-led or lean team and budget is tight, start with the content layer on your hero products. The ad eligibility is already on, so the highest-return fix you can fund yourself is making the listings strong enough to win the conversation. If you want the data to do this properly without a full retainer, our DaaS tier gives you direct access to the source data across Amazon and the other channels at two hundred and fifty pounds plus VAT a month, with onboarding so your team can read it and act.
If you run a multi-brand portfolio and need to bring several stakeholders along before anyone acts, this is an Enterprise conversation. A workshop to get your Amazon, content and media teams aligned on what the conversational surface changes, then multi-brand tracking so you can see where each brand stands as the surface develops. The point at that scale is shared understanding before spend, not a quick campaign tweak.
Whichever route fits, the underlying job is the same. Amazon has already put your ads in the conversation. The work now is making sure the conversation goes well. That is content, not budget, and the brands that understand the difference will pull ahead while everyone else waits for a new format that has, in effect, already arrived.
If you want a quick read on whether your catalogue is ready for the conversational surface, that is exactly the kind of thing we can put a number on. Worth a look before your competitors get there first.
Stephen Honight is the founder of Lmo7, an AI-native agency helping consumer brands win in AI-powered discovery and agentic commerce. Lmo7 works with brands including Trip Drinks, Veloforte, Brown-Forman, Haleon, Pelotan and Symprove across Amazon and D2C. Before Lmo7, Stephen led ecommerce and Amazon operations at Unilever, Mars and The LEGO Group.