Looking for an oak dining table delivered to my door. Any recommendations?
An AI tool listed:
- 1Depo
- 2Rūme
- 31a.lv
Shoppers already ask ChatGPT, Google AI and Perplexity where to buy. The answer usually names two to five stores.
Example
Looking for an oak dining table delivered to my door. Any recommendations?
An AI tool listed:
I want decent cotton bed linen, and soon. Where do I order it?
An AI tool listed:
Asked about a service, AI tools often cite directories: zl.lv, firmas.lv, viss.lv. Asked about a product, they cite the stores’ own product pages.
On 4 September 2026, asked “Where in Latvia can I buy bed linen online”, Perplexity cited product pages on blues.lv, 220.lv and saulesgultasvela.lv. Asked about a dentist and an accountant, it named directories only.
So for a store its own pages decide: whether an AI tool can reach them, what it reads there, and whether the product data is complete.
The product data was invented for this page.
Illustrative example
The title says what is sold
€14In stock. Delivery in 2–3 working days.
Price and availability are visible in the text
Black fabric bag in 100% cotton, with no fastening and no pockets. Handle length 70 cm.
Facts, not adjectives: material, dimensions, care
Structured data: Product, Offer
Structured data matches the visible text
How to draft a description like this with AI, and what to check before publishing: How to use AI to draft product descriptions for an online store
Some store platforms and security tools block GPTBot, OAI-SearchBot or PerplexityBot. If the crawler cannot open the page, the store does not appear in the answer.
A shopper asks for “cotton bed linen 200 × 220”, not “home textiles”. A category page has to say what is in it, in which sizes and at what price.
Material, dimensions, what is included, care, price, availability. “Perfect” and “premium quality” tell an AI tool nothing.
Product and Offer data must match what the shopper sees: the same price, the same availability.
+60%higher conversion on visits from AI tools
In US online stores, visitors who arrived from AI tools bought 60% more often than the rest in July 2026.
No such measurement exists for Latvia. So we measure what can be measured: how often AI names your store, and which pages it cites.
Adobe, 19 August 2026Prices exclude VAT. The full price list and the request form are on the Pricing page.
Free
We put eight shopper questions about your product group and show which stores the AI names.
€290
30 questions in four AI tools, three runs each, an AI-crawler access check and a list of what to fix first.
€690
We fix access, put the structured data in order and rewrite the 6–10 category and product pages that matter most.
€290–790 a month
We measure every two weeks, show what moved, and write answer pages.
€290 then €200 a month
Your products advertised in ChatGPT conversations. The budget stays in your account.
If you have hundreds of products, AI can draft their descriptions and attributes from the supplier’s data.
Yes. In a check on 4 September 2026, asked about bed linen, Perplexity cited product pages from three Latvian stores. Which stores the AI names for your product group is what the free audit shows.
No. We start with the category pages and the products shoppers ask about most. In the Foundation sprint that is 6–10 pages.
It does. The platform decides how fixes are made, not what the AI reads. First we check that neither the platform nor a security tool blocks AI crawlers.
OpenAI has published a product feed specification with nine required fields. Who can use the feed today is something we check before we promise to include it in the work.
With repeated measurement: the same questions, the same tools, three runs, dated. One screenshot proves nothing, because the next day the answer can differ.
We use cookies to improve analytics and understand site use
Could not save your choice. Please try again.