In short
- The basis for visibility is complete, consistent product data in the shop, structured data and product feed: title, description, price, availability, brand and GTIN.
- OpenAI’s standard format for product feeds currently targets the US, according to its specification. In Europe, Google Merchant Center is the established route. Clean data helps with both.
- EU obligations: online offers need manufacturer details and warnings (Art. 19 GPSR). For price reductions, the lowest price of the previous 30 days must be stated.
- Status questions, returns, feed checks and drafts of product texts can be automated, with approval for refunds and legally relevant information.
Ecommerce SEO starts with product data
Search engines, price comparison sites and AI assistants read the same basic data. The more complete and consistent it is, the better.
| Data field | Why it matters | What to watch |
|---|---|---|
| Title, description, image | Required fields at Google and OpenAI, and the first thing customers and AI see | Unambiguous and factual, each variant with its own ID |
| Price and availability | Required fields in both specifications, must match the page | Feed the feed and shop from one source, up to date daily |
| Brand and GTIN | Google requires the brand for new products and strongly recommends the GTIN. The MPN is enough only if there is no GTIN | Give only identifiers you are sure are correct |
| Manufacturer details and warnings | Art. 19 GPSR: manufacturer’s name and contact details, product identification with picture, warnings | Clearly visible in every online offer |
| Price reductions | Art. 6a Price Indication Directive: the prior price is the lowest price of the previous 30 days | Store the price history and derive the statement from it |
Structured data and feeds
Google distinguishes two routes. Product snippets suit pages where people cannot buy directly. Merchant listings are meant for pages where customers can buy, and they support details such as shipping and returns. Both rely on product data in the page markup, see Schema markup. Added to that is the product feed in Google Merchant Center.
OpenAI describes a feed format of its own for ChatGPT. Each row is one purchasable item or variant, and nine fields are required, including ID, title, description, brand, seller name, image, availability and price. According to the specification, the standard format currently targets the US, with further markets only after OpenAI confirms. For Europe it therefore pays to build the data foundation for Google first. It can be reused for other channels later.
Where automation helps in a shop
The examples are typical processes, not client projects.
- Status questions and returns: the AI looks up the shop and the carrier, replies and issues the returns slip. Refunds above a fixed amount and special cases go to people. More at Automate customer enquiries.
- Feed checks: every day, missing identifiers, price differences between shop and feed, and broken image links are looked for and reported.
- Product texts: drafts arise from manufacturer data. A person approves them, and mandatory details are added.
- Reviews: draft replies and an analysis of recurring criticism.
- Advice in the shop: a chatbot answers questions from your own product data, see Website chatbot.
How shops get found: Google and AI assistants
- Advice instead of repeated text: category pages with buying guides, comparisons and care instructions answer the questions customers actually ask.
- Genuine reviews: visible, not bought, with factual replies.
- Clear shipping and returns information: on pages of their own and in the markup of the offers.
- Do not lock crawlers out: check whether your robots.txt blocks AI crawlers by mistake, see AI crawlers and robots.txt.
- Measure: ask questions such as “Where do I buy …” and “Which shop for …” and see who is named.
For a chatbot in the shop, the AI disclosure under Art. 50 of the AI Act applies; for customer data, data protection with a data processing agreement. This overview is general information, not legal advice.
How to start
- Check your data: take a sample of products. What is missing in identifiers, brands, images?
- Add mandatory details: manufacturer details, warnings, price history.
- One source for everything: shop, markup and feed are generated from the same data.
- Choose one automation: usually status questions or returns.
- Measure: enquiries per day, response time, share of cases solved, and mentions in AI assistants.
Sources
- Google: Product structured data
- Google Merchant Center: Product data specification
- OpenAI: Feed reference for products
- Regulation (EU) 2023/988 on general product safety, Art. 19: EUR-Lex
- Directive 98/6/EC (price indication), Art. 6a, inserted by Directive (EU) 2019/2161: EUR-Lex
- Regulation (EU) 2024/1689 on artificial intelligence, Art. 50: EUR-Lex