How AI changed ecommerce search — and why your product pages stopped winning
Nothing announced it. Rankings reports still look fine. But the page a buyer reads first is no longer a page you own, and most stores are still optimising the layer underneath it. Here is exactly what moved, in order.
One question, two completely different journeys
Search → links → your page
Search → answer → maybe a click
Why engines skip product pages almost every time
It is not a penalty and it is not personal. It follows from what an answer engine is trying to do.
An engine only cites what it does not already know
It has read the manufacturer's description on two hundred other stores. Repeating it back adds nothing, so there is no reason to name you as the source of it.
Product pages are built to sell, not to answer
Features, shipping, reviews, an add-to-cart button. A buyer asking "which lasts longest" is not answered by any of it, and the engine can tell.
Guides answer the question directly
So guides get quoted — including guides written by competitors, marketplaces and magazines that do not even sell the product.
The citation, not the ranking, is now the prize
Being named in the answer is what puts your brand in front of the buyer. Position one underneath it is a consolation.
Four things every quoted page has
This is the whole formula, and almost nobody in ecommerce is doing all four.
A direct answer, immediately
The question answered in the opening two sentences, the way an expert would say it out loud — not four paragraphs of preamble before the point.
Specialist vocabulary
The precise nouns a professional uses: materials, failure modes, measurements, standards. Breadth of correct terminology is a large part of why a page reads as authoritative.
Something first-hand
What you see when a product fails. The mistake customers repeat. A comparison you actually ran. This is the part a competitor cannot copy, and the part that earns the citation.
Structure a machine can read
Clear headings, real FAQ markup, one idea per section. An engine cannot quote what it cannot cleanly extract.
And search stopped happening only on Google
Each of these is a search engine with its own index and its own ranking rules. Most stores post to them and never optimise for search on them.
TikTok
Buyers search it before Google. It indexes spoken words and on-screen text.
How we do it →YouTube
Playlists rank on their own, and AI answers quote how-to videos.
How we do it →A planning engine. Boards rank, not just pins.
How we do it →Profile name, bio and alt text are indexed fields, not decoration.
How we do it →Where this leaves an ecommerce store
Not "write more blogs". Four specific moves, each of which we run as a service.
Find the questions you have not answered
Your published content, matched against what buyers actually ask. What is missing is your list.
See the service →Upgrade what you already published
Most stores gain more from fixing existing posts than from adding new ones. Audit first.
See the service →Fix the pages with real demand
Category pages carry more search than SKUs and are usually a title and a grid.
See the service →Cover the other engines
TikTok, YouTube, Pinterest and Meta each need their own keyword work.
See the service →The ones we get asked most
Is traditional SEO dead?
Do we need to stop optimising product pages?
Which engine matters most for ecommerce?
How is this different from hiring a content agency?
Can AI just write all of this for us?
See what AI engines say about your store today
The scan reads your live catalogue and shows which questions in your category you already answer, which you do not, and where a competitor is being cited instead of you. No call required to get it.