GEO for webshops: how AI decides whether your products get found

For twenty years, being found online meant one thing: ranking in Google. That assumption is quietly expiring. Research by Annalect found that 33% of Dutch consumers already use AI tools such as ChatGPT to find information about products and services. Two years earlier, that was 2%. Ruigrok measured this summer that the share of Dutch people using AI at least weekly doubled in a single year, from 19% to 40%. And in the US, Pew Research found that when Google shows an AI-generated summary above the results, users click a regular result on only 8% of visits, against 15% without the summary.

The marketing world's answer to this shift is generative engine optimization, or GEO: making sure AI engines mention, cite and recommend you. By now there is no shortage of Dutch articles explaining what GEO is. What almost all of them have in common: they are written for websites that publish content, not for webshops that sell products. And for a webshop, GEO is a fundamentally different job.

This article explains that difference, shows what AI engines actually look at when they answer a product question, and gives you a concrete plan, including the part your Shopware shop already handles for you.

GEO in one paragraph, and why most advice doesn't fit a webshop

GEO is optimizing for the engines that generate answers, such as ChatGPT, Gemini and Google's AI Overviews, rather than for a list of ten blue links. The standard advice follows from how these engines work: publish authoritative content, answer real questions, structure your pages clearly, and be the source the engine wants to cite. All true, and for your blog and category texts it applies to a webshop too.

But a webshop's most important pages are not articles. They are product pages, and an AI engine does not read a product page the way it reads an essay. When someone asks an AI assistant for "a quiet dishwasher under €600 that fits a 45cm cabinet", the engine is not looking for beautiful prose. It is looking for structured facts: noise level, price, dimensions, availability, reviews. Content-GEO is writing work. Product-GEO is data work. That is why generic GEO advice leaves webshop owners with a blog strategy and an invisible catalogue.

How AI engines actually see your products

Three things determine whether your products surface in AI answers.

Structured data. Schema.org markup on your product pages, such as Product, Offer and AggregateRating, is the difference between an engine knowing your price, stock status and review score, and an engine guessing. This has been SEO hygiene for years. AI answers raise the stakes, because an engine composing a shortlist simply skips the product whose attributes it cannot read.

Machine-readable, current facts. AI engines increasingly check prices and availability at answer time. A feed or page that still shows last week's price, or an out-of-stock product presented as available, does not just cost that one sale. It teaches the engine that your data cannot be trusted.

Complete, consistent product information. The attributes the engine filters on, such as dimensions, materials, compatibility, energy label and warranty, have to exist as data, not as a line buried in a marketing text. And they have to be consistent everywhere the engine might look: your shop, your feeds, your marketplace listings.

Notice what is not on this list: tricks. There is no meta tag that makes an AI recommend you. The engines reward shops whose product information is complete, structured and true. That is either bad news or excellent news, depending on the state of your product data.

The foundation you have (hopefully) already laid

If you have followed the compliance calendar this year, you have been doing GEO groundwork without calling it that. The EU guarantee labels that became mandatory in September forced warranty data onto every product. The Digital Product Passport will do the same for sustainability data from 2027. And agentic commerce, where AI agents do not just recommend but actually buy, demands exactly the same machine-readable catalogue.

That is the pattern worth seeing: regulation, AI findability and agentic commerce all reward the same discipline. Product data maintained once, in a PIM as the single source of truth, distributed through an integration layer to your shop, your feeds and your marketplaces. Shops that treat each of these as a separate project will do the same data work three times. Shops that fix the foundation do it once.

What Shopware gives you out of the box

If you run Shopware, you are not starting from zero. The standard Shopware storefront already includes basic Schema.org markup on product detail pages. Shopware also offers a free Agentic Readiness Scanner: enter your store URL and it rates your storefront on product data, discoverability and emerging agent standards, with prioritized recommendations. In practice, it is a report card on exactly the machine-readability this article is about. And release 6.7.15 added sitemap and SEO URL support for headless sales channels, closing a findability gap for composable setups.

The honest caveat: out of the box covers the mechanism, not the content. The standard markup is a starting point. Details that AI shopping engines increasingly use, such as shipping costs and return policy, are not part of Shopware's default output and need to be added through your theme or an extension. And Shopware can only mark up the attributes you maintain. It cannot invent the dishwasher's noise level you never entered. The scanner tells you where you stand. The work it points to is data work.

A five-step GEO plan for your webshop

Step one: measure. Run the Agentic Readiness Scanner, and check a handful of your top products in Google's Rich Results Test to see what an engine can actually read.

Step two: audit your structured data. Are price, availability, reviews and key attributes present as data on every product, or only on the products someone once got around to?

Step three: fix the product data at the source. Complete the attributes that matter for your category, in your PIM or ERP, so every channel inherits the fix.

Step four: make the time-sensitive facts reliable. Price and stock must be correct everywhere an engine might read them, every day. This is an integration question more than a content question.

Step five: measure AI traffic separately. Referrals from ChatGPT, Perplexity and similar engines are still small for most shops, but they are the trend line that tells you whether this work pays off. Set up the segmentation now, while the baseline is low.

Frequently asked questions

What is the difference between SEO and GEO?

SEO optimizes for a ranked list of links; GEO optimizes for being used in a generated answer. They overlap heavily: clean structure, real expertise and solid technique help both. But GEO punishes missing structured data harder, because an engine cannot cite a fact it cannot read.

Can I buy my way into AI answers with ads?

Not meaningfully, today. AI engines compose answers from what they can read and verify. That is precisely why complete, structured and current product data is the lever, and why early movers are hard to displace later.

Is this relevant for a B2B webshop?

Yes, arguably more. B2B buyers ask exactly the specification-heavy questions AI engines are good at ("which supplier stocks X in quantity Y"), and B2B catalogues tend to have the richest attribute data, provided it is maintained and machine-readable.

My organic traffic is still fine. Why act now?

Because the shift is compounding, not sudden: a third of Dutch consumers already use AI to find products, and weekly AI use doubled in a year. Product data takes months to fix properly, and the engines are forming their picture of your shop now.

Learn more about making your webshop findable for AI

Being found is moving from a ranking game to a data game, and for webshops that is the real meaning of GEO: not a new bag of content tricks, but the moment your product data quality becomes your visibility. The good news is that everything pushing you in this direction, from guarantee labels and product passports to AI agents and AI answers, rewards the same single investment.

As a Shopware Silver Partner and Alumio partner since 2020, Memo builds the product data foundation and the integrations that keep it accurate in every channel. Want to know how visible your products are to AI engines today? Request an AI findability scan or book an introduction call, and get a prioritized to-do list for your shop.