350 Million People Shopped Through an Assistant Last Year. Your Storefront Is Quietly Becoming a Data Feed.
· CX Pulse
Alexa for Shopping hit 350 million users, and the customers using it spend 40% more per order. They never see your product page.
350 Million People Shopped Through an Assistant Last Year. Your Storefront Is Quietly Becoming a Data Feed.
On Amazon's Q2 earnings call last week, Andy Jassy said more than 350 million shoppers have used Alexa for Shopping in the past twelve months. Active users nearly doubled year over year. Interactions were up five times.
The number worth stopping on is a different one. US customers who use the assistant spend about 40% more per order than customers who do not.
For roughly a decade, "nobody actually shops by voice" was a fair thing to say. It stopped being fair sometime in the last year, and the people who switched are not the low-value end of the market.
What changes when the assistant is the interface
A shopper using an assistant does not see your product page. They see whatever the assistant decided to say about your product.
That answer is assembled from structured data. Titles, attributes, specifications, availability, price, return terms, review summaries. The layout you paid for, the photography, the copy someone agonised over, the carefully built comparison table, none of it is in the room.
This is a quieter shift than it sounds, because nothing on your side breaks. Your site keeps working. Traffic looks normal. What changes is that a growing share of buying decisions now happen upstream of your site, in a summary you did not write.
Where this actually bites
The businesses that struggle here are not the ones with bad websites. They are the ones whose product data was always a bit approximate, because a human could fill the gaps.
A dimension left blank because the photo made it obvious. A variant described in the title instead of a field. Stock levels that are roughly right. Return terms that live in a PDF. A spec sheet that says "contact us for details."
A person browsing forgives all of that. An assistant answering a direct question does not. It either has the fact or it does not, and when it does not, it says something vague or recommends something else.
What this means if you sell anything
Three things follow, and none of them require adopting any of this yourself.
Your product data is now customer-facing copy. It used to be back-office plumbing that fed a page. It is increasingly the thing the customer actually hears. Nobody proofreads a feed the way they proofread a homepage, and that gap is going to start showing.
Gaps are worse than errors. A missing field does not produce a wrong answer, it produces a hedge, and a hedge loses to a competitor whose data was complete. Silence is a ranking signal now.
Post-purchase is in scope too. The same assistants are moving into order status, returns and support. If your systems can only answer those questions to a human on a phone, they cannot answer them to the software the customer is actually using.
The part nobody budgeted for
Most organisations have someone responsible for the website and someone responsible for the catalogue, and those are usually not the same person, and the catalogue one is usually more junior.
That made sense when the catalogue fed a page a human would read. It makes less sense when the catalogue is the page.
None of this requires a strategy or a platform. It requires somebody looking at the product data the way a customer now encounters it, which is as an answer rather than a listing, and noticing every place it says nothing at all.
Source: Amazon Q2 2026 earnings, reported by CX Dive, August 3, 2026.