Product Listings Are Being Written By Machines
On a marketplace or a quick commerce app, the product page is where the purchase actually happens. Yet product listings, the titles, bullet points and descriptions on those pages, are increasingly written by generative tools. A brand with hundreds of products across several platforms has thousands of fields to fill and refresh, and generation made that volume manageable.
It also turned the most commercially important words a brand publishes into its least considered.
Why product listings went first
The work was always tedious and usually outsourced. Catalogue teams or vendors wrote listings to platform templates, stuffed with search terms, rarely read by anyone senior. Handing that task to a model looked like an obvious saving, and in many cases the output is no worse than what it replaced.
Platforms themselves now offer sellers tools to draft listings, which makes the shift faster still.
The ad is approved by five people. The page where the customer decides is often approved by none.
Where it goes wrong
Sameness is the first problem. A model asked to write a persuasive listing for a face wash will produce much the same listing for every face wash, and on a results page where products sit side by side, the brand becomes interchangeable in exactly the place it most needs to be different.
Accuracy is the second. Generated listings can describe features a product does not have, overstate a benefit, or carry a claim that the brand's own advertising would never have been allowed to make. The listing is still the brand speaking, and the rules on misleading claims do not stop at the edge of the marketplace.
Who should own the words
Someone accountable for the brand, not just the catalogue. That means a short set of approved product facts and claims that every listing must draw from, a voice that distinguishes the brand from the shelf next to it, and a regular check of what is actually live on each platform.
Brands spend heavily on retail media to get shoppers onto these pages. It makes little sense to spend so little on what those shoppers read when they arrive, and generation is a reason to pay more attention to the page, not less.
More from The AI Shift.