My Labradorite Page Won A Gift Search I Never Targeted. So I Built One On Purpose

Oct 5

A few months ago, while I was drafting an email, I ran a quiet test on my own store. I opened ChatGPT and asked it the way a real gift buyer would: "recommend a unique handmade gemstone necklace for my wife who loves sculptural, one-of-a-kind art jewelry, around $500." Nothing. My brand did not come up at all.

Then I asked Google the exact same question. I was the first recommendation, ahead of Etsy. And the page earning that spot was not a gift guide, because I did not have one. It was my labradorite page: an education page about a stone I love, with a small gift-finder tool built into it.

I run two things. Red Pin Geek, where I help independent jewelry designers get found by AI, and Andrea Li Designs, my actual jewelry store, which I use as my test lab. So that split result, invisible to ChatGPT, first on Google, was not a curiosity. It was a clue. The page that won was the one a machine could actually read.

That is when the real question landed. If a page I never built for gifts can win a gift search on Google, what happens when I build one on purpose and aim it at the tools that could not see me yet?

This is that experiment, run on my own store, in the open.

Key takeaway. AI recommends the brands it can read. I found that out by accident when Google put my store first for a gift search, ahead of Etsy, through a page I never built for gifts. So I built a gift page on purpose and tracked it in the open. Three months on, two AI tools now recommend my store for that search and none of them cite the gift page. They cite the pages I wrote to explain something. The fix is not more beauty and it is not a technical overhaul. It is translating your "vibe" into plain facts a machine can repeat, and a tool, the Vibe Translator, does that translation for you.

Why does AI recommend some jewelers and skip others?

Here is what is actually happening. When someone asks an AI for a gift recommendation, the AI does not browse the way a person does. It does not scroll your feed, linger on a photo, or feel the weight of a piece in your hand. It reads. It pulls the brands it can actually describe and hands back a short list of names.

That single fact decides who gets named. The cruel twist for makers like us is that the things we pour the most into, the mood, the atmosphere, the feeling, are exactly the things a machine cannot see. The more wordless and beautiful your work is, the easier it is for an AI to skip you, not because the work is not good enough, but because nothing on the page told the machine what it was looking at.

If you take one idea from this piece, take that one. You are not competing on beauty anymore. You are competing on whether a machine can repeat what you do.

The clue: a stone-education page beat Etsy on Google

For most of my eighteen years making jewelry, almost every sale came from someone who already knew me. A friend, a referral, someone I met at a show. A stranger finding me online and buying was rare enough that I had made peace with it.

Then I rearchitected my entire website on this same pillar-page approach, and that peace got happily disrupted. Since then I have seen a measurable increase in sales from strangers, more inquiry-form submissions from people I had never met, and more Zoom consultations booked by names I did not recognize. The method was already working across my whole store, quietly, before I ever pointed it at gifts. So the experiment in this piece is not me hoping something untested will pay off. It is me aiming a proven approach at one specific, high-value search.

So when Google put my store first for that gift search, ahead of Etsy, I went looking for the reason. It was not my shop page. It was not a gift guide, because I did not have one. It was my labradorite page: a long, plain, factual page that explains what labradorite is, what it costs, how it wears, and why I use it, with a small gift-finder tool built into it. I wrote it to teach, not to sell gifts.

But "teach" is exactly what made it work. That page is dense with the kind of facts a machine can lift: the material, the stone, the price range, who it is for, where it ships from. Google had something concrete to grab, so it grabbed my store. ChatGPT, with less it could read from me, grabbed nothing. The beauty was never the signal. The readable facts were.

Why did Google find me while ChatGPT missed me completely? At the time I assumed the obvious answer: Google is out reading live pages constantly, while the chat tools lean more on what they were taught during training. That explanation felt clean, and I have since watched it fall apart. When I ran the same question again in September, the order had flipped. The chat tools named me. Google's AI answers did not. So the honest version is less tidy and more useful: these tools do not share one index, one schedule, or one set of rules about who gets named. Each one decides on its own, and which one finds you first is not something you can predict or control. What you can control is whether there is anything readable for them to find, because none of them can name a store they cannot describe. Build that once and you are eligible everywhere, whichever tool gets there first.

The experiment: I built a gift page on purpose, and a cluster around it

A clue you stumble into is interesting. A result you can repeat on command is a method. So I set out to test it deliberately.

In June I built a gift page on my own store, at andreali.com/gifts/. But I did not stop at one page, and the reason is the whole point. I worked backward from the gift buyer's mindset and mapped the questions they actually move through: the fear at the start (will I get this wrong), the decision in the middle (which stone, which color), and the objection at the end (what if it is not right). Then I built a high-intent blog post for each stage and linked all three into the pillar:

how to read her style when you are frozen at the start,
how to choose a stone by color without asking her,
and what happens if she wants to return or resize it.

I chose those three on purpose, not by guessing. Each one answers a real stage of the gift-buying journey, and each one feeds the pillar. This matters because a pillar with a cluster of specific, readable, on-topic pages pointing into it gives an AI far more it can confidently grab and repeat than a lone page does. The architecture is the argument: more readable, corroborating facts, and more chances to be the brand that gets named.

Now the honest part, because the alternative is the kind of overselling I refuse to do. I published the pillar and the three posts in June, and Google has indexed them. But I am writing this as an experiment in progress, not a victory lap. I am tracking the same query that started all of this, the one ChatGPT could not answer with my brand, across ChatGPT, Perplexity, Gemini, Google's AI Overview and Google's AI Mode, and I would rather show you what actually happens than what I hoped would.

Here is the honest read, on September 23, three months after the gift page went live.

Two of the five tools I checked now name my store for that question. In June, none of them did. Perplexity does more than name me: it put me first in both runs and closed by narrowing its own list down to me. ChatGPT names me too, though as a second thought rather than its top pick, and it named a specific piece, my Imperial Jasper necklace at $425, with the details right down to the Oregon sunstone. I ran that one in a temporary chat with personalization turned off, so it was not my own browsing history being handed back to me.

Google's AI answers, both the Overview and AI Mode, still do not mention me at all. Gemini did not either, though Gemini hides most of the sources it used, so I count that one as unproven rather than a clean no.

So the tools moved, and they moved in the opposite direction from the one this story started with.

Now the part I did not expect, and the part worth your time. Not one of them cited my gift page. The pages they quoted were my necklaces collection, my wearable art page, a studio story about the Tamar collection, and one product page. Search Console, the free Google report that shows which searches your pages turn up for, says the same thing from the other side: the gift page has 33 impressions and zero clicks in the last month, and nearly every search I can actually see is somebody typing my own name. Google hides the smallest ones, so I cannot claim the rest are the same, only that nothing in the visible half looks like a gift search.

I built a gift page on purpose, and it is not the page winning the gift search. The pages winning are the ones I wrote to explain something.

I want to be careful about how much weight that carries. This is a small number of runs, on one question, on one day. Ask again next week and the specific pieces named will move around. But it is the second time the same thing has happened to me, and the first time was the accident that started this piece. A page that teaches something specific gets quoted. A page that mainly gathers products to buy does not, at least not yet, and not on its own.

That does not make the gift page a mistake. It means the gift page still needs what the labradorite page already had: something on it worth quoting.

Underneath all of it is one move, and it is the move I want you to be able to make today. Here it is in a table. The left column is how we instinctively write. The right is what a machine can actually repeat.

What you wrote (the vibe)	What AI can repeat (the facts)	Why it matters
"Born from a season of stillness"	"Handmade 14k recycled-gold necklace with a vintage chrysocolla stone, one-of-a-kind, ships from Denver"	The AI can now name the material, the type, and the origin
"Bold, unapologetic, a conversation piece"	"Sculptural statement labradorite necklace, gift-ready, around $500"	It matches a gift search with a real price band
"For the woman who is the center of attention"	"Statement gift jewelry for a partner who loves bold, one-of-a-kind pieces"	It matches the words a gift buyer actually types

Nothing in the right column is less beautiful. The poetry still lives on the page. The facts just go first, where the machine can find them.

What "translating your vibe into facts" actually means

You do not have to become technical, and you do not have to lose your voice. The whole move is translation, not replacement.

Lead each page with what the piece is, in plain facts: the materials, what it is, who it is for, the price range. Then write as beautifully as you always have. "Handmade 14k recycled-gold necklace with a vintage chrysocolla stone" comes first. "Born from a season of stillness" comes right after. You lose nothing, and you gain a reader the machine can quote.

That is the entire idea behind the tool I will point you to in a minute. It keeps your words exactly as you wrote them and shows you the plain facts to add in front of them, so the piece becomes readable by the tools doing the recommending without losing an ounce of your voice.

One more thing worth saying plainly, because the people with budget for fine and custom work are the ones already shopping this way: 74 percent of households earning over $100,000 use AI tools, compared with 53 percent under $50,000 (Menlo Ventures with Morning Consult, "2025: The State of Consumer AI," April 2025, a nationally representative survey of 5,031 US adults). The fact that you may not personally shop with AI is not evidence your buyers do not. They are not you.

How to do it on your own site, without getting technical

Here is the method, the same one I ran on my own store. It works on Squarespace, Shopify, or anywhere else. The platform changes the buttons, not the idea.

Lead every product and gift page with plain facts. First sentence: what it is, the materials, who it is for, a price range. Keep all your poetry, just put it second. This alone is most of the work.
Build one page on purpose for the search you want to win. Pick the query, a gift pillar is the obvious one heading into the holidays, and write a real page for it that links to your relevant pieces. Intent compounds. A page built for a search beats a page that wins it by accident.
Make the facts match underneath. The plain facts on the page should also live in your structured data, the small behind-the-scenes tags that tell Google and AI what a page is. When the visible facts and the hidden facts agree, the AI trusts and repeats them.

Do this before shoppers start gift-searching in earnest, not during. Those pages take a few weeks to get picked up by Google and the AI tools, so the work you do now is what gets you named in October and November.

Why now: the October window

I want to be careful here, because the last thing I will do is manufacture urgency. This is not a scare tactic. It is a lead-time reality, and it is the one piece of timing that actually matters.

Gift searches are seasonal. The questions people ask an AI in October and November, "a unique handmade gift for my wife," "a one-of-a-kind necklace under a budget," are the ones that send buyers to checkout in December. That part is obvious. The part most makers miss is that the page answering those questions has to exist, and be read, before the searching starts.

Remember the mechanism. An AI recommends the brands it can read. For it to read your gift page, the page has to be live long enough for Google and the AI tools to pick it up, which takes weeks, not days. A gift page you publish in early October has time to be found and start surfacing by the time shoppers are actively asking. A gift page you publish in December is technically perfect and effectively invisible, because the season is already underway and the page has not been read yet.

So you are not guessing at the pace, here is the rough timeline I plan around after publishing a pillar like this. Treat it as a map, not a guarantee, since the exact speed depends on your site and how often the tools recrawl it.

Stage	Roughly when	What you are watching for
Indexed	A few days	Google has added the page to its records. You can confirm it in Search Console.
First search traction	1 to 3 weeks	The page starts appearing for its target searches, and early impressions show up.
AI answers catch up	2 to 4 weeks or more	The page becomes eligible to be pulled into AI Overviews and chat answers.

Work backward from that. If you want to be in the answer when shoppers are searching in late November and December, the page needs to be live and read by early-to-mid October. That is the runway, and it is the whole reason I am writing this now instead of in November. I built my own gift page in June, on purpose, so it would be old news to the AI tools by the time it mattered. You do not have that much runway left for this year, but you have enough. The work you do in the next few weeks is the work that gets your name into the answer while the queries are live.

One honest caveat: this is not a one-and-done. A gift page built now keeps working next year, and the year after, getting stronger as it is read more. That is the quiet upside of doing the foundation work instead of chasing a seasonal post that is gone in a day. You are not renting attention for the holidays. You are building a page that earns it, every season, on its own.

See your own translation: the Vibe Translator

Reading about translation is one thing. Seeing your own copy translated is what makes it click.

The Vibe Translator takes one of your product or gift descriptions, keeps your voice exactly as written, and builds a facts-first version in front of it from a few details you fill in. You see, side by side, exactly which facts a machine was missing and where they go. For most makers, that single view is the moment the whole idea stops being abstract.

It is not a magic button, and anyone promising that AI will instantly recommend you is overselling. The mechanism is simpler and more honest: AI reads your words for the feeling and your facts for the facts, and when the facts are there and accurate, you get understood, and recommended, more often.

Frequently asked questions

How do I get my jewelry recommended by AI for gift searches?

Lead your gift and product pages with plain facts (what the piece is, its materials, who it is for, the price band), build one page intentionally for the gift search you want to win, and make those facts match your structured data underneath. AI recommends what it can read, so the goal is to make the gift-relevant facts easy for a machine to repeat.

Does building content on purpose really beat content that ranks by accident?

The accidental version already worked once: Google put my store first for a gift search, ahead of Etsy, through a page I never built for gifts. The deliberate version, a gift page plus a cluster of high-intent posts built on purpose, is designed to do that reliably and to close the gap on the tools that could not see me yet, like ChatGPT. I built it in June and checked again in September. Two of the five AI tools now name my store for that search when none of them did before, and none of them cite the gift page itself. I am reporting that as it stands rather than claiming a cleaner win.

Will this ruin my brand voice or require technical skills?

No. The method is translation, not replacement. Lead with one sentence of plain facts, then write as beautifully as you always have. The few more technical pieces, like matching your facts to your structured data, can be handled with a simple tool or done for you.

I am on Squarespace or Shopify. Does this apply to me?

Yes, on every platform. The steps differ a little, the idea is identical: put the facts about your work where a machine can read them, and build on purpose for the searches you want before shoppers start looking.

Sources

Menlo Ventures with Morning Consult, "2025: The State of Consumer AI," April 2025 (nationally representative survey of 5,031 US adults): 74% of $100K+ households use AI tools vs 53% under $50K. https://menlovc.com/perspective/2025-the-state-of-consumer-ai/
Andrea Li, first-party (June 2026): the gift-query test (ChatGPT returned no brand match; Google ranked Andrea Li Designs first, ahead of Etsy, via the labradorite education page with a built-in gift-finder tool); the deliberate andreali.com/gifts/ gift pillar plus three high-intent supporting posts, published and indexed June 2026.
Andrea Li, first-party (2026-09-23 re-check, same query): Perplexity named Andrea Li Designs first in both runs; ChatGPT named the brand on both prompt wordings in a temporary chat with personalization off; Google AI Overview, Google AI Mode and Gemini did not cite the brand. None of the five cited /gifts. Google Search Console, 28 days to 2026-09-22: /gifts at 33 impressions, 0 clicks. A small number of runs on one query on one day, reported as a dated check rather than a measurement.

Continue Learning

The Foundation Gap: why AI search skips independent jewelers: the bigger picture this gift experiment sits inside, why beautiful sites get passed over.
AI Knows You Exist. But Can It Describe You?: being found is not the same as being understood, the step right before this one.
Score your own store on the free AI Visibility Score: see, in two minutes, how readable your store is to AI right now, the same gap this gift experiment closed.
Subscribe to the Red Pin Geek Substack (free): weekly inside-view content for designers building AI visibility on their own. The Premium tier walks through the Vibe Translator step by step, with a real client case study.
Work with me on your specific store (Readiness Audit / Snapshot $97 / Full Audit $597 / Full Audit + 1:1 call $997): the private read. I run this same gift-and-vibe check on your store and tell you exactly what AI can and cannot see.

Andrea Li is the founder of Red Pin Geek and has spent 18 years in the jewelry industry. She designs and sells handcrafted gemstone jewelry at andreali.com and helps independent jewelry designers get found by AI search at redpingeek.com. Every strategy she offers clients, she tests on her own store first.

Want Red Pin Geek in Your AI Answers?

Google now lets you choose preferred sources that get labeled in AI Overviews and AI Mode. If this content helps you navigate AI visibility for your jewelry business, add us and you'll spot our links faster next time Google answers your questions.

Add Red Pin Geek as a Preferred Source

Takes one click. Works in Google Search, AI Overviews, and AI Mode.