"customers can smell fake from a mile away and the second they realize its AI they're out." That warning came from a Reddit post in a Shopify community, and it captures the tension of this holiday season in one line. Shoppers are happily asking ai agents and AI assistants what to buy, who to buy it for and where to get it delivered in time. Then they punish any brand that looks thin, generic or fake when they arrive.
The headlines keep landing on the same shift. The gift search no longer starts with a search bar or a walk past a shelf. It starts with a question typed or spoken to an assistant: a gift for a father who cooks, a stocking filler under a set budget, a skincare set that suits sensitive skin and arrives before a date. The assistant answers with a short list. If your brand is not on that list, the rest of your marketing never gets a chance to work.
Why are ai agents sending shoppers to some brands and not others?
An assistant can only recommend what it can read, check and trust. It pulls product facts from your pages, your feeds and the places your products are listed. It weighs reviews for recency and detail. It looks for clear answers on price, contents, delivery and returns. Where the facts line up, it can recommend you with confidence. Where the facts are missing or contradict each other, it moves on to a brand it can describe accurately.
This is not a beauty contest. A polished homepage with vague copy loses to a plainer page that states what is in the product, who it suits and when it arrives. The assistant is not being disloyal to your brand story. It simply cannot sell a story it cannot verify, and it will not risk a recommendation built on a guess.
That is why the fake problem matters so much. Shoppers use AI to find gifts, then apply a human test to what they find. Stock imagery that could belong to anyone, reviews that read like templates and product claims with nothing behind them all fail that test. The brands that win this season will look real and read clearly at the same time.
What should a small CPG brand fix before the season?
Start with the facts an assistant reads first, and make each one complete, current and consistent everywhere your product appears.
Product facts an assistant can read. State ingredients or materials in full, with sizes, weights, quantities and allergen information where it applies. Put them on the product page in plain text, not only in an image or a downloadable file. If a gift buyer asks what is in it, the answer should take seconds to find.
Gift suitability, said plainly. Say who the product is for, what occasion it suits and what problem it solves. Add the price band in words a shopper would use. Assistants match gifts to people, so vague claims about quality help far less than a clear line about the recipient.
Gift guides and bundles written the way people ask. Build one guide around real questions: gifts for new parents, gifts for runners, gifts that travel well. Group products into bundles with a named purpose and a total price. Write the guide in question and answer form where it helps, because that is the shape of the request the assistant is trying to answer.
Reviews where assistants look. Recent, specific reviews carry weight. Ask recent buyers for detail about who they bought for, how the product performed and whether delivery matched the promise. A short run of honest, dated reviews on your own pages and your retail listings beats a large pile of old, generic praise.
Stock and delivery honesty. Publish your order cutoffs for holiday delivery, by region where they differ. Keep stock status accurate, and say so plainly when an item has sold out. Nothing burns trust faster than an assistant recommending a gift that cannot arrive, and shoppers remember the brand, not the tool.
Much of this work overlaps with the move from shelf to screen set out in the retail to DTC playbook, where the facts that once sat on packaging now have to work in product data too.
Not sure what an assistant finds when it reads your brand?
The Growth Audit checks the gaps that stop a small brand getting recommended, from thin product facts to missing reviews, and maps the fixes in order.
What did seasonal retail teach about being findable?
I built Liquiproof from zero to 3,000+ retail doors, including Adidas, IKEA, Selfridges and Burberry, and exited profitably in under six years. In stores like Selfridges and IKEA, the seasonal lesson was always the same. The product that sold was the one a busy shopper could understand in seconds: what it was, who it was for and why it was worth picking up. If the display left questions open, the shopper moved to the next shelf.
An AI assistant is that shelf now, only faster and less forgiving. I learned in retail that honesty about stock and delivery was not admin, it was the sale itself, because a disappointed gift buyer rarely gives a brand a second chance in the same season.
What happens if the facts stay thin?
The assistant does not leave a gap. It fills the gap with another brand, or with an older version of your facts pulled from a retailer page or a review site. Prices drift, old bundles resurface and discontinued lines get recommended. Meanwhile the shopper who does reach your site meets copy written to sound impressive rather than to answer questions, and the fake test kicks in.
There is a second cost. Gift buyers are often new customers buying for someone else. A good first order, delivered on time and as described, is the cheapest route to a repeat buyer in the new year. A bad one closes that door before it opens. Getting found by AI in the answer only helps if what the assistant found was true.
What should you do this week?
Keep it small and concrete. This is a prep list, not a rebrand.
Pick your five gift products. Choose the items most likely to be given, not your full catalogue. Fix those pages first and let the rest follow.
Run the assistant test. Ask the main assistants the questions your buyers ask, and read what comes back. Note every wrong fact, missing detail and competitor named instead of you.
Close the gaps in one pass. Update the product facts, the gift guide, the delivery cutoffs and the review requests together, so every surface tells the same story on the same day.
Check again after a week. Assistants refresh what they read at different speeds. A second test shows whether the fixes landed and where an old listing still contradicts you.
The Growth Audit exists for exactly this kind of gap between a good product and a brand an assistant can recommend. If you run a small CPG brand and want the fixes mapped in order before the season peaks, start with the Growth Audit for CPG brands and work the list while there is still time to matter.