WRKNG's AI Commerce engine took a woman-owned, kitchen-table jewelry brand from invisible in AI search to the most-cited brand in its entire category, across ChatGPT, Perplexity, Gemini, Copilot and Google AI. With zero ad spend.
Mack & Rex is a small, woman-owned business based in Utah, founded in 2020 by a creative mom and her two young daughters. The name comes from the girls' nicknames, McKinlee ("Mack") and Lexi ("Rex"), who were just 8 and 11 years old when the three of them dreamed the business up.
They'd learned to make Tila bead bracelets together and found the craft genuinely addictive: the kind of creative activity anyone can pick up in an afternoon. So they set out to make it easier for everyone else to find and buy high-quality Tila bead kits and supplies.
What started at their kitchen table grew into a full-scale Shopify operation, competing for attention against global marketplaces with a fraction of their budget.
By 2026 the front door to a purchase had moved. Shoppers stopped typing "tila bead bracelet kit" into Google and started asking AI which one to buy, then acting on the handful of brands it named.
Mack & Rex weren't in that answer. AI recommended marketplace listings and a few other stores instead. Of every citation AI made in their category, Mack & Rex accounted for 1%. Their competitors weren't the problem. The models simply didn't know or trust them enough to name them.
And it wasn't a problem they could spend their way out of. There's no ad auction for the inside of a ChatGPT answer. They were losing customers before those customers ever saw the store, and had no way to see it happening.
Getting recommended by AI takes two things, and most brands are missing both. AI has to be able to find your products, and it has to trust your brand enough to name it. We engineered both.
Layer one: the Foundation. We rewrote every product in the catalog for the way AI actually reads a feed: titles, descriptions, structured facts, materials, use cases and the real search phrases shoppers use. Not keyword stuffing. It is the specific data a model needs to match a product to a question.
Layer two: the Content Layer. When we switched the WRKNG AEO System on it began publishing the buying guides, comparisons and answer-shaped content AI reads, trusts and cites, on their own domain, so the authority accrues to them.
Then we measured every AI answer in their category, daily, across five engines.
Same store, 30 days apart. The slate bar is the "before." The green is now.
Flat for six weeks of measurement. Then the WRKNG AEO System went live, and AI citations, brand mentions and Google all broke upward in the same days.
Every direct competitor in the category, by how often AI cites them as a source.
Across every AI answer in their category, this one Shopify store is referenced more often than the three biggest sources on the internet, added together.
Citations sat at zero for six straight weeks of measurement, then went vertical within days of the system going live. AI visibility isn't a slow linear climb; it's flat until you cross a threshold of citeable content, then it accelerates. Brands who quit at week four never see it.
AI pulls from their collection and product pages (tila-bead-bracelet-stack, their starter kits) alongside their buying guides. That's the two-layer effect: the feed work makes products citeable, the content work makes the brand trustworthy. Neither alone gets you there.
Blog impressions went 334 → 14,069 and clicks 0 → 199 over matched 30-day windows. We never optimized for it. Whether Google rewards a piece depends on whether real search demand exists for it, which is why this lands differently for every brand, but it has never yet gone the wrong way.
There's no ad auction inside a ChatGPT answer. Mack & Rex beat competitors and out-cited the internet's biggest platforms on $0 of ad spend, because the position is earned through what you publish, not what you bid.
We'll run the exact questions your buyers ask AI, show you who it recommends today, and where you rank against your competitors. Same engine that did this.
Get Your Free AI Commerce AuditCitation, coverage and competitor data measured live across ChatGPT, Perplexity, Gemini, Copilot & Google AI (US), 2026. Google figures from Search Console, blog content only, matched 30-day windows. Prepared by WRKNG Digital · wrkngdigital.com
Sixteen days after launch, Petersen was cited more than every competing publisher combined.
Read case study → AI CommerceBefore we sold it to anyone, we proved it where agencies fight to be the one AI recommends.
Read case study → AI CommerceIn one month, eight of their own pages were cited by AI — and they passed two tracked competitors.
Read case study →