A Jewelry Store Now Gets a Quarter of Its Online Revenue From AI. Here's What Changed.

In February, this store earned nothing from AI shopping assistants. Not a dollar. The channel did not exist for them.

Last month it produced 24% of their online revenue — with zero advertising spend behind it.

Nothing about the products changed. What changed is that a machine could finally read the catalog.

The store

Brownlee Jewelers is a Charlotte jeweler: seven locations, ninety years, three generations. Exactly the kind of business that has every advantage offline and, until recently, almost none of it online.

Their first order from an AI assistant is dated 10 May 2026. Here is where it stands now:

• 24% of online revenue now comes from AI shopping assistants
• 29x more likely to buy — AI visitors versus the average visitor
• 550% ChatGPT growth inside a single 30-day window
• Zero advertising spend behind any of it

The multiple is the number worth sitting with. A visitor arriving from an AI assistant has already described what they want, been shown a specific piece as the answer, and clicked. That is not traffic. It is a pre-qualified referral, and it converts like one.

Orders rose 300% and average order value rose 77% over the same period, while sessions moved barely 15%. Almost none of this came from more people showing up. It came from the right people showing up, already sold.

Why this is not a jewelry story

It would be easy to read that as a lucky category. The industry data says otherwise.

Adobe reports that AI-driven traffic to US retail sites grew 393% year over year in Q1 2026. By March that traffic converted 42% better than non-AI traffic — a complete reversal from a year earlier, when it converted 38% *worse*. Revenue per visit runs 37% above non-AI sources.

That reversal is the signal. Twelve months ago this channel was worse than what you already had. Now it is materially better, and it is compounding faster than any other acquisition channel most retailers own.

The part almost nobody has fixed

Here is the sentence that should make you uncomfortable.

Adobe scored US retail sites on how machine-readable they are. Homepages averaged 75%. Category pages, 74%. Individual product pages — where the buying actually happens — averaged 66%. In Adobe's words, much of that content is currently invisible to LLMs.

Roughly a third of the average retailer's product content cannot be read by the systems now sending the best-converting traffic on the internet.

This is not a content problem or a marketing problem. It is a data problem, and it is unglamorous:

• Missing identifiers. No GTIN or MPN, non-unique variant IDs, empty brand fields. These are named exclusion reasons — the product never enters consideration.
• Titles a machine cannot parse. A bare SKU like "14K-R-0382" is invisible no matter how beautiful the piece is.
• Attributes trapped in prose. These systems read structured fields: stone, shape, carat, metal, setting, size. A paragraph of lovely copy with no structure underneath reads as an empty product.
• Crawlers blocked below the surface. Bot-protection and WAF rules block AI crawlers far more often than robots.txt does — a layer beneath where anyone thinks to look.
• Products never published to the sales channel. Live and selling on the website, never eligible for the surfaces. Nothing errors. Nothing in your admin tells you.

Every one of those was present in the jeweler's catalog. Fixing them is what produced the number at the top of this article.

What is actually running out

Two things are moving at once, and they move in opposite directions.

The surfaces are being handed to you. Shopify now syndicates catalogs to ChatGPT, Copilot, Gemini, and Perplexity from your own admin. You do not have to build the pipe, and you should be skeptical of anyone who offers to sell you one.

But syndication is not the same as working. It pushes your catalog out; it does not make your catalog good. If a third of your product data is unreadable, you are syndicated and invisible at the same time — and nothing in your dashboard will tell you which.

Meanwhile the competitive window is closing on its own. Fewer than 1% of online retailers use advanced AI agents today; roughly a third are projected to by 2028. The advantage available right now is the advantage of being legible while your competitors are not. That is a temporary condition.

What we would do

Nothing exotic. We read your catalog the way an agent reads it, then fix what is stopping it.

• Grade every product against the actual feed specifications
• Check whether the AI crawlers can reach you at all, including at the CDN layer
• Verify price and availability match between your feed and your live pages
• Ship structured product data your product pages currently lack
• Turn on each surface and confirm it with live queries, rather than assuming

Then we measure what actually happens for the searches your buyers really run. See the full approach for any e-commerce catalog, or the version built for jewelers.

One thing we will not tell you

We cannot promise placement. Nobody outside OpenAI, Google, Microsoft, and Perplexity controls what those systems surface, and their rules change regularly. Treat any agency that guarantees a ranking as a red flag.

What is controllable is whether you are eligible, technically reachable, and accurately represented. Those are engineering problems with known answers. The jeweler's results followed six months of catalog work — they are evidence this pays off in retail, not a forecast for your store.

Find out where you stand

The first useful question is not what this costs. It is whether an AI assistant can currently read your catalog at all — and most owners have no idea, because there is no dashboard that reports it.

We will tell you, free. Book a 15-minute consultation: no pitch, no deck. Bring your store URL and we will tell you what we see, including if the answer is that you are already in decent shape.

If a quarter of your online revenue is going to come from somewhere new in the next eighteen months, the cheapest possible time to find out whether you are eligible is now.

*Sources: Adobe Analytics — AI traffic and retail machine-readability; Adobe Q1 2026 AI traffic report via TechCrunch; Digital Commerce 360 — AI-referred retail traffic doubles. Brownlee figures are from the store's own Shopify Agentic Storefronts reporting and an independent first-touch trace of every order, published with permission.*