For twenty years brand discovery has had two front doors: search, where a person typed a need and scanned a list, and the feed, where a person scrolled and something caught them. Creators built businesses in the second door, and a sponsorship was, at bottom, a way for a brand to get into a feed a person trusted. Both doors are still open. But a third has appeared, and it works differently. A person asks an assistant, in plain language, which laptop to buy or which skincare works in a humid climate, and the assistant answers. It does not surface a list. It makes a judgment call.
Marketers at large brands have started to say this out loud. On a panel this summer with executives from IKEA, Marriott, Unilever and Adobe, the framing was that brand discovery has always happened in places companies do not control, but that what is new is the intermediary itself choosing. When the assistant decides, brand fundamentals matter more than ever, and credibility has to be earned at the moment a person first encounters a brand through an AI rather than through an ad.
That framing was for brands. For creators it raises a sharper question. In a feed, a creator is the intermediary. In an assistant, what is a creator for?
How the answer gets formed
Assistants do not invent their recommendations. They draw on what has been written and said about products, weighted by how credible and how consistent it appears, and they increasingly cite where the answer came from. A product that has been reviewed, compared, tested and discussed by people who seem to know what they are talking about shows up in the answer. A product that exists only in its own advertising does not, or shows up with a caveat.
This is the point creators should sit with. The raw material of an AI recommendation is precisely the kind of content good creators already make: the honest review, the side-by-side comparison, the long video that actually uses the thing. What changes is that the content now works twice. Once for the people who watch it, and again as evidence when an assistant is asked the question the video answers.
Why the region’s creators have an edge
Global assistants are trained heavily on English and on markets where reviews are abundant. Ask one which phone to buy and it has a great deal to work with. Ask it which phone plan is best in Kuwait, which abaya brand holds up, which Cairo restaurant is worth the drive, or which Arabic-language course actually teaches Gulf dialect, and it has far less. The regional, multilingual, specific knowledge that a creator in Jeddah or Tunis or Beirut carries is exactly what the systems lack, and exactly what they will reach for when someone asks a regional question.
The raw material of an AI recommendation is the content good creators already make. What changes is that the content now works twice: once for the viewer, and again as evidence when an assistant is asked.
That makes the local creator something like a human version of the assistant, a filter that knows the market from the inside, and the brands that understand this will value regional creators more, not less, as discovery moves to AI. A brand cannot write its own way into a Gulf-specific answer. A creator who has actually compared the products in the Gulf can.
Being the source the answer cites
None of this happens automatically. Content that is easy for an assistant to find, read and trust is content that is titled clearly, says what it compared and how, states conclusions plainly, and lives somewhere a system can index. A twelve-minute video with the verdict buried in the last minute and a title designed for curiosity is fine for the feed and nearly invisible to an assistant. The same review with a clear title, a written summary in the description, a plain verdict and a published transcript is discoverable in both doors.
The written layer matters more than most video creators are used to. Descriptions, captions, a companion post, a written version of the review on a site the creator owns: these are what the systems read. Creators in the region who publish in Arabic and English, with the dialect named, are giving the assistant precisely the signals it needs to match a regional question to a regional answer.
And the content has to be trustworthy in a way a system can detect. Sponsored content presented as independent, verdicts that contradict each other across videos, praise for everything: these read as noise. A creator with a clear track record of honest verdicts, including negative ones, becomes a source. A creator who has recommended every product they were ever paid to hold does not.
What this means for brand deals
The brands that grasp the shift will start asking creators for something different: not just a post that reaches a feed, but content that will still be found when a person asks an assistant six months later. That is a different product, with a longer shelf life and a different value, and creators should price it that way.
It also changes what a brand should want from a creator. A creator whose honest verdict on the brand’s product ends up in the answer is worth more than one who was paid to be enthusiastic, because the assistant weighs the former and discounts the latter. The most valuable creator relationship a brand can have in the coming years may be with someone it does not fully control.
The move now
Make the verdict findable. Write it down, not just say it. Publish somewhere you own. Name the market and the language. Keep a record of what you have recommended and why, so that the pattern a system sees is one of judgment rather than availability. The feed will still be there, and the sponsorship will still be the main business. But the third door is opening, and the creators who are already standing in it will be the ones the answer points to.