GEO for online stores: how can customers find you in AI search?
Why does AI search mention one store but not another? We explain where to start with technology, product data and buying advice.
Quick answer
GEO for online stores improves technical access, consistent product information and answers to specific customer questions. Start with important category and product pages, and monitor source links, visits and orders separately.
GEO for online stores starts in three places: search systems must be able to reach the pages, product information must be accurate and content must answer specific customer questions. These foundations matter when you want ChatGPT or Google AI search to consider your store as a source.
At sjmedia GmbH, we connect this work with SEO and conversion optimisation. An AI answer is just one possible starting point. Once in the store, customers still need to understand which product suits them and whether they want to buy it on the terms offered.
Why does my store not appear in ChatGPT or other AI search results?
Possible reasons include technical access problems, missing information or other sources that answer the question more precisely. A single AI answer does not establish which cause applies.
Start with a specific question about your range. Is your company mentioned? Is there a link to your store? Or does the answer simply mention a product brand sold by you and many other retailers? These are different outcomes.
For a GEO audit, we record the questions tested, the search system and the sources found. This reveals where a relevant answer is missing and which pages need closer examination. A mention in one query is not enough to demonstrate success.
For an introduction to the terminology, our article on SEO, GEO and AEO explains their different priorities.
Can search engines and AI crawlers read your store?
First check important category, product and advice pages. Are they publicly accessible? Does an indexing restriction prevent inclusion in search engines? Does essential product information only load after an interaction? Resolve these questions before commissioning new copy.
Internal linking matters too. Google recommends making products accessible through categories and subcategories. Products available only through internal search may be missed during crawling. Google on online store navigation
OAI-SearchBot is relevant to ChatGPT Search. It serves search, while GPTBot supports potential model training. OpenAI describes these controls as independent. A store can allow the search crawler while blocking the training crawler. Check firewall and bot protection access as well as robots.txt. OpenAI explains its crawlers
Permission enables access. The search system then decides whether to select a page as a source.
Which product data matters for GEO?
Choose ten products for an initial review, including variants and products that need buying advice. Compare the product page, its structured data and the information reaching the product feed.
Even inconsistent variant names can make matching difficult. Incorrect prices, outdated stock or a link to a general listing instead of the advertised product are more serious.
Structured product data makes information such as offers, prices and availability machine-readable. Google describes these fields in its Merchant Listings documentation. They must match the actual offer. Google on structured data for retailers
Find the cause of conflicting information. Does delivery time come from the inventory system, a store field or an extra feed rule? If several places maintain the same information, correcting one isolated error is rarely enough.
What makes a store useful as a source?
Useful content helps customers decide what to buy. Simply repeating a manufacturer’s description does not necessarily answer the questions your team receives.
Consider a retailer selling height-adjustable desks. A customer wants to know whether a monitor arm fits the desktop, but its thickness and mounting details are missing from the product page. A general guide to ergonomics does not solve that problem. First, the product page needs the missing information.
Other questions belong elsewhere:
| Customer question | Suitable content |
|---|---|
| Which desk size fits a small study? | Selection advice on a category or guide page |
| How do two available frames differ? | A comparison using specific criteria |
| Can I order spare parts later? | Product-specific service information |
| Can the desk be delivered to Switzerland? | Clear delivery terms linked from the product |
Answers should state conditions and limitations. “Suitable for every home office” is less useful than clear dimensions, load capacity and space requirements.
Your team can contribute knowledge absent from standard copy: assembly tips, experience with accessories or common reasons for unsuitable purchases. Use your own photographs and documented measurements where available. Make the source clear when using manufacturer information.
Our recommendation: start with a question your customer service team answers repeatedly. The finished page should be useful enough that the team would send customers a link to it.
What matters for Germany and Switzerland?
A German-language text does not automatically describe the right offer for both countries. Customers in Switzerland need to know whether you deliver to their address, which currency applies and what costs and delivery times to expect.
Check the destination pages for each country’s offers. Do the product, variant and terms match the market where they appear? Is information about returns and possible additional charges easy to reach and understand?
French- or Italian-language offers need their own content review. A translation can be linguistically correct while missing local product terminology or customer questions.
Does GEO work with PrestaShop, Shopify and Shopware?
Yes. These tasks can generally be addressed with PrestaShop, Shopify and Shopware. Whether a change uses existing features, the theme, an extension or custom development must be checked in the individual store.
We look at the output: what appears on the product page, what crawlers receive and which data is exported. Where several modules or apps are installed, establish which component generates structured data and feeds.
Changing ecommerce platforms solely for GEO would be an excessive first step. Identify the specific obstacles in your current store first.
Does an online store need llms.txt for better AI visibility?
Google does not require llms.txt. Google says this file does not affect visibility or ranking in its search. No special GEO schema is required either. Whether the file benefits another service depends on that service actually using it. Google’s guide to generative search optimisation
With limited time and budget, we would first address missing product information, blocked pages or conflicting feeds. These are concrete problems that need solving regardless of an additional file.
The same principle applies to FAQs. A dedicated question section can help resolve outstanding points. It does not need an artificially long list of almost identical questions.
Seven checks to start GEO optimisation
One important category and selected products are enough to begin. Record the questions and pages you will assess before making changes, so you can track what improved.
- Check access: Review indexing, crawler permissions, internal links and technical restrictions.
- Compare product data: Check variants, attributes, prices and availability across store and feed.
- Collect customer questions: Match recurring sales, support and search questions to relevant pages.
- Add answers: Explain selection criteria, compatibility, limitations and required accessories clearly.
- Demonstrate expertise: Include available original images, measurements and meaningful comparisons.
- Review country-specific offers: Check terms and destination pages for Germany and Switzerland separately.
- Measure changes: Monitor source links, brand mentions, visits and orders over appropriate periods.
The review should produce a concrete task list. Technical obstacles and incorrect information take priority. Next comes content customers need for important buying decisions.
How can you tell whether the work is helping?
First verify that the planned changes were implemented. Is product data now consistent? Can customers find the previously missing information? Are important pages accessible and indexed?
Then examine changes in the selected questions and available search and store data. A brand mention, source link and order are different outcomes. Record them separately.
Our article “Google AI Mode and AI Overviews: what changes for online stores?” explains how to analyse Google’s AI impressions. It does not establish a fixed date when a search system will start using your content.
GEO advice for your online store
At sjmedia, we review technical access, product data and content together. Our aim is a priority list your team can work with: what must development change, which data is missing and where do customers need better advice?
To find out where your store stands, talk to us about GEO and AI SEO.