Every few weeks, someone in the trade gets pitched a new service. AIO, short for AI optimization. GEO. AEO. A separate budget line, a separate team, a separate set of tricks to make sure ChatGPT and Google's AI answers mention your store. It sounds urgent, and it sounds expensive.

Here is the short version, and it will save some of you real money. SEO and AIO are not two disciplines. In May 2026, Google published its first official guidance on the question and stated plainly that optimizing for its AI features, AI Overviews and AI Mode, is still SEO. Those features draw from the same search index and the same ranking systems that decide who ranks in an ordinary result. For a jewelry retailer or supplier, that means one workstream, not two budgets.

There is a catch, and it matters. It is just not the catch the people selling AIO packages want you thinking about.

What Google Settled in May 2026

The guidance runs long, but the spine of it is one idea. Google's AI answers are built on retrieval augmented generation, which is a technical way of saying the model reads Google's existing search index to ground what it writes. If a page is not indexed and eligible to show in Google Search with a snippet, it is not eligible to appear as a supporting link in an AI Overview or AI Mode either. The front door is the same door.

That single fact collapses most of the AIO hype. Google went further and named tactics you can ignore for its search: llms.txt files, chopping content into tiny fragments so a machine can parse it, bolting on special markup made only for AI, and chasing manufactured mentions across forums. None of it is a ranking input for Google Search. Structured data still earns its place, but not because AI needs a secret format. Clean markup has always helped search understand a page, and that has not changed.

So the work that earns a citation in an AI answer is the work that earned a blue link. Useful content, a crawlable site, clear structure, genuine expertise. The surface changed. The discipline did not.

Where SEO and AIO Stop Overlapping

Now the catch. Google's statement governs Google's own surfaces, AI Overviews and AI Mode, and nothing else. ChatGPT, Perplexity, and Claude run their own retrieval, keep their own freshness windows, and choose their own sources. Some of them read files that Google ignores. So "it is all just SEO" is fully true for Google and only partly true everywhere else.

The gap is wider than it sounds. Ahrefs analyzed roughly 730,000 AI Mode and AI Overview response pairs in December 2025 and found the two reached the same conclusion about 86 percent of the time, yet cited the same web pages only 13.7 percent of the time. Same company, same index, very different citations. Across independent engines the spread is wider still, and it keeps moving. Ahrefs separately found that about 45 percent of AI Overview citations change when the answer regenerates.

Strong fundamentals travel across every engine, because each one is trying to surface the same thing, a clear and credible answer. But no specific citation is guaranteed, and the ones you earn churn. For a jeweler, that means a ranking today is not a durable asset by itself. The durable asset is the structured record behind the product: the stone attributes, the setting options, the certificate, the policies, the availability, and the price logic. That is what every engine has to interpret before it can recommend anything.

For Jewelry, It Comes Back to Product Data

Here is where the trade should actually spend its attention, because our category has a specific version of this problem. Every surface we have mentioned, Google results, AI Overviews, and the agentic shopping channels now reading Shopify's Catalog, is fed by the same raw material: structured, accurate, consistent product data. Google's own guidance makes the commerce version explicit, pointing merchants to clean product feeds as the thing that surfaces products in AI answers.

Jewelry is where product data tends to fall apart. Loose diamond inventory arrives as messy feeds with attributes that change from one supplier to the next. Ring configurations break the moment a shopper wants something past the three options a page was built to show. Certificate details sit in PDFs or images instead of structured fields, so an engine may see a document but still miss the product facts that matter. We see this constantly in the supplier feeds we normalize for the VDB marketplace, and it is the reason two stores selling nearly identical stones can end up with completely different visibility.

Fix the data once and every surface has something cleaner to read at the same time. A diamond listing with complete, consistent attributes (cut, color, clarity, carat, measurements, and a certificate number checked against the grading lab) is legible to Google, to an AI Overview, and to an agent building a cart inside an assistant. This is the same groundwork we covered when Shopify's agentic checkout went live - read here and when we wrote about getting found inside ChatGPT and Copilot - read here. The channel keeps changing. The data discipline underneath does not.

The point is easier to see in a table. The same listing is read by different surfaces in different ways, and the same weakness sinks it everywhere.

What This Changes About How You Spend

The budget implication is the useful part. You do not need a separate AIO team or a package of AI-specific hacks. You need the fundamentals funded properly, then measured across more than one surface.

Three things earn the return. First, content only your team can write, what Google calls non-commodity content, the operator insight a generic summary cannot reproduce. Second, product data clean enough to be machine-legible everywhere, which for jewelry is the hard part and the high-leverage one. Third, monitoring that looks past Google, because Search Console will never tell you whether Perplexity or ChatGPT cited you.

The payoff shows up where it counts. Seer Interactive found that brands cited inside an AI Overview earn about 35 percent higher organic click-through on those queries, so presence in the answer compounds rather than cannibalizes. And the audience is already here. EMARKETER expects 31.3 percent of the US population to use generative AI search in 2026. That is not a forecast to prepare for. It is current traffic.

None of the techniques that work are new, either. The original 2023 research that coined the term generative engine optimization found that visibility in AI answers rose by up to 40 percent through citing sources, adding statistics, quoting experts, and writing with authority. Read that list again. It is a description of good SEO.

The surfaces are going to keep multiplying. AI Mode, shopping agents, whatever ships in the next Editions release. Chasing each one as a separate project is how a marketing budget gets spent on tactics that expire in a quarter.

The discipline underneath is stable, and for the jewelry trade it is refreshingly concrete. Be the clearest, best-attributed, most structured source on the stones and the pieces you sell. That is a data-quality problem before it is a marketing problem, and the shops that treat it that way will keep showing up no matter which surface a buyer happens to be standing in front of.