Articles — Product data for shops
Product data for shops
Structured data publishes facts machines can read
Schema publishes facts machines can read, including name, brand, price, and stock, so they do not have to guess from marketing copy. It does not by itself make ChatGPT name you.
If you care about showing up in ChatGPT, Claude, Perplexity, or Gemini, adding a structured block can sound like the step that gets you named. Getting named still depends on trust, reviews, and corroboration. See How AI chatbots build a shortlist. When the price or the shop is wrong, see Wrong price, wrong shop, fake discount.
JSON-LD is a structured block on the page. Your product lives in three places defines that block, schema, the feed, and the product codes.
Why machines need clearer facts than humans do
Humans can read a messy product page and still understand the offer. Machines do better with separate fields.
Illustrative example, not a logged test. The Northloop Trail Watch is a fictional product at €300.
- name = Northloop Trail Watch
- brand = Northloop
- price = 300 EUR
- availability = InStock
“From €249” and a crossed-out old price are easy for a person to skip and easy for a machine to treat as the price. The number in the structured data should be what the customer pays now.
If those facts live only inside marketing copy, a shopping system may miss the product, get the price wrong, confuse it with a similar model, or skip it when the shopper asked for a hard limit such as “under €300”.
Schema and feeds help with an accurate description. They do not invent brand authority you do not have. Agentic commerce, checkout or stock lookup inside a chat, still reads those same fields.
What structured data does not do
JSON-LD alone will not make ChatGPT suddenly recommend your brand.
Google’s guide to generative AI features in Search says there is no special Schema.org markup you need to add to appear in AI Overviews or AI Mode. Markup can still make a page eligible for a richer Google listing (price, stars) and it labels facts machines can read. Microsoft’s October 2025 note on AI search answers says schema helps search engines and AI systems understand the page. Neither source shows that markup gets a shop or a place named in a chatbot answer.
Structured data is still worth doing because it is:
- Concrete. You can validate what you published.
- Foundational for a shop. The same clarity shows up in feeds and catalogs.
- A clearer picture of the job. The assistant consumes data and sources.
- Useful when you are skipped. Unclear or inconsistent product facts are one place to check.
A practical order of work
If you only screenshot chatbot answers, you get interesting examples and still may not know what to fix.
A stronger sequence:
- Make product facts clear for machines, on the page, in the structured block, and in the feed.
- Observe how answers behave, including who gets named, cited, or described wrong.
- Change the weak spots, such as titles, descriptions, sources, and freshness, and check again.
If a chatbot writes the JSON-LD for you, check name, price, and stock against what a person sees on the page, then run Google’s Rich Results Test.
Structured data makes the facts clear enough for a machine to read. It is not the step that gets you named.
Read next if: Wrong price, wrong shop, fake discount · Your product lives in three places