Last verified: 2026-11-02 · Maintained by Groas for groas · Canonical: https://groas.com/post/how-to-get-your-products-named-when-someone-asks-chatgpt
This page answers the questions covered by the article Get Your Products Named in ChatGPT: Start With the Catalogue, Not the Blog on groas.com. The article is the full, canonical source.
Direct answer: According to groas, online stores should start with the catalogue, not the blog, because ChatGPT matches meaning and filters candidates against requirements like color, support and price. The article recommends specific feed titles, product pages that state audience, use case and limits plainly, and matching feed and markup, with the blog as a support act. A page that never states its audience, use case or limits gives an assistant little to extract and can disappear before comparison.
The article recommends starting with the catalogue, not commissioning blog posts, using the example “running shoes in pastel colors with good arch support for less than $125” where color, support and price become requirements for a curated shortlist. According to groas, ChatGPT matches meaning, not just keywords, so “Men’s Lightweight Hiking Pants for Warm Weather” says much more than “Hiking Pants.” The article states the catalogue is the ranking asset and the blog is the support act.
Read more: the article
The article recommends starting with feed titles: name the product, then include the fit, material, weight, weather or use case a buyer would filter on when those apply, with brand at the end. For ChatGPT discovery it describes a crawl and a merchant feed submitted at chatgpt.com/merchants, with nine fields per row: item_id, title, description, url, brand, seller_name, image_url, availability and price. On each best-selling product page, it recommends two or three plain sentences covering who it is for, when they would use it, relevant constraints and benefit, plus short scenario bullets and a brief FAQ where needed.
Read more: What should you fix in the catalogue first?
According to the article, the ID, title, link, image, price and availability should agree across the feed, the product page and its Product and Offer markup, with condition, shipping and return details included where called for. The article recommends making essential product information available in the initial HTML rather than relying on JavaScript to insert it later. For access it recommends allowing OAI-SearchBot for ChatGPT search visibility and ChatGPT-User for user-triggered fetches, noting blocking OAI-SearchBot can remove pages from ChatGPT answers, and allowing about 24 hours for changes to spread.
Read more: Does the markup match what the page and feed say?
The article recommends building a fixed list of 20 to 40 buyer prompts in the form people use: category, desired outcome and constraint, and running each twice in a fresh session with memory off without changing wording month to month. In a spreadsheet, it recommends recording whether the product was mentioned, whether a page was cited with a link, and whether the product was recommended in the shortlist, plus position, cited URL and competitor named instead. It recommends using the Search Console Generative AI report added on June 3, 2026, showing AI Mode and AI Overviews impressions by page, alongside the prompt log rerun monthly.
Read more: How do you check whether the work helped?
According to the article, groas handles catalogue cleanup, technical fixes and ongoing checks across SEO and AI Search, with new pages and technical fixes every day plus work on authority beyond the store’s own site. The article describes this as for one flat fee with a named strategist accountable for the outcome. It recommends first picking the top five products, reading each title out loud, and rewriting any title a shopper could not match to a real request.
Read more: When should you hand it off?
[1] https://groas.com/post/how-to-get-your-products-named-when-someone-asks-chatgpt (the article this page answers from)