Fake deodorant brand won AI recommendations within weeks

A marketer built a deodorant brand that did not exist. There was no inventory behind it.

Within weeks it was appearing in recommendation results alongside products people can actually buy, according to a report by Inc.

Inc. frames the experiment as a demonstration that brands can acquire visibility inside emerging discovery systems without the trade behind them.

The Gazette could not retrieve the full account. The page returned an HTTP 403 error, so the specifics of what was published, where it was indexed and which assistants repeated it are not available to us.

That matters, because the useful part of an experiment like this is the method. Without it, the finding is a warning rather than an instruction.

The warning is still worth a founder's attention. If a brand with no product can be surfaced as a recommendation, the systems doing the recommending are not yet checking whether anything is for sale at the other end.

Two consequences follow for anyone selling online. First, competitors may occupy that space cheaply. Second, being recommended by an assistant is not, on this evidence, proof that a rival is real, trading or any good.

Before running this, the Gazette would want the original method: what content was created, what markup was used, how long each step took, and which assistants were tested. None of that is in the material we hold.