Getting found in ChatGPT shopping and AI assistants: what fashion brands can control
Nobody can buy a guaranteed place in an AI assistant's answer. Fashion brands can, however, control crawler access, product data and policy clarity, and those are the levers that matter.
KEY TAKEAWAYS Summary by the editors
- OpenAI states that placement in ChatGPT search is not guaranteed, so brands should manage what they can control rather than chase a ranking.
- To be eligible for ChatGPT search, a site must allow OAI-SearchBot to crawl it, and its host or CDN must not block OpenAI's published IP addresses.
- OpenAI says ranking in product results considers availability, price, quality, primary-seller status and whether Instant Checkout is enabled.
- Google says there are no additional requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible to show a snippet in Search.
- Accurate, current product feeds with variants, identifiers and return terms are the most concrete brand-controlled input to AI shopping surfaces.
Can a fashion brand buy or guarantee a place in ChatGPT shopping?
No. OpenAI's help documentation says that ChatGPT ranks search results using multiple factors and that placement is not guaranteed. Google makes a similar point about its own AI features: meeting the technical requirements does not guarantee crawling, indexing or serving. A brand's realistic goal is therefore eligibility and accuracy, not a promised position.
That framing matters for budgets. Agencies and tools that promise to engineer a ranking inside an AI assistant are selling something no publisher has documented. What can be documented is a short list of inputs that a brand owns, and the rest of this analysis sets those out.
What does ChatGPT use to decide what to show?
OpenAI describes the signals only at a high level. For product results it says ranking considers availability, price, quality, whether the seller is the primary seller and whether Instant Checkout is enabled, and it adds that checkout-enabled items receive no ranking preference. For general answers that use web search, the help page says responses may include citations, and that results can be incomplete, outdated or incorrect.
Two consequences follow. First, a brand sold through many resellers competes with itself, and being recognised as the primary seller is a stated factor. Second, ordinary commercial basics such as price and stock accuracy are part of the stated criteria, which is a useful reminder that these systems reward operational quality as well as content.
Which levers do brands actually control?
| Lever | What the publisher documents | Practical action |
|---|---|---|
| Crawler access | OAI-SearchBot must be allowed for ChatGPT search eligibility | Check robots.txt and CDN or firewall rules |
| Product feed | Required fields include ID, title, description, URL, brand, seller, image, availability and price | Supply complete, current data with one row per variant |
| Search eligibility | Google needs pages indexed and eligible for a snippet | Keep pages indexable and avoid blanket nosnippet rules |
| Return and policy data | Feed fields exist for return acceptance, window and policy URL | Publish accurate terms in structured form |
| Seller status | Primary-seller status is a stated ranking factor | Make brand-owned pages the clearest source of truth |
How do crawler settings affect visibility?
OpenAI separates its crawlers by purpose. OAI-SearchBot surfaces sites in ChatGPT search features, GPTBot collects content that may be used to train models, and ChatGPT-User fetches pages when a person asks a question. A site can allow the first and disallow the second, because the settings are independent. OpenAI notes that a change to robots.txt takes about 24 hours to be reflected in search results.
Many fashion sites run bot-management rules at the CDN level that were written to stop scrapers. These can block legitimate search crawlers by accident. OpenAI specifically tells site owners to confirm that their host or CDN allows traffic from its published searchbot IP addresses. This is a quick check with a high payoff, and it belongs with whoever manages the site's security configuration, not with the marketing team alone.
Why does product data matter more than copywriting?
OpenAI's feed specification is blunt about what a usable record looks like: a stable identifier, a factual plain-text description, an absolute public URL, a direct image link, an availability value and a price with currency. It rejects placeholders such as null or n/a, and expects variants to appear as separate rows grouped to a parent. Identifiers such as GTIN and MPN are optional but help matching.
For fashion, this is where many catalogues struggle. Colour names that are marketing terms rather than plain colours, size charts that differ between systems and stale stock data all make machine matching unreliable. A tidy catalogue also serves Google Shopping, marketplaces and the brand's own site search, so the work is not wasted if an AI channel disappoints.
- Write titles that state the product type, fit and colour in plain words.
- Keep descriptions factual: material, construction, care and measurements.
- Match prices and availability across site, feed and marketplaces.
- Group variants consistently and keep image URLs permanent.
Resellers and marketplaces add complexity. If a wholesale partner publishes a thinner or older description than the brand's own, an assistant may rely on it. Brands that supply retailers with images, copy and data in a consistent, current form reduce that risk. It is worth checking, product by product, which version of a style's description is most visible, and whether it is the one the brand would choose.
What about Google's AI Overviews and AI Mode?
Google's documentation says there are no additional requirements to appear in AI Overviews or AI Mode, that no special optimisation is necessary and that no new machine-readable files or special markup are needed. A page must be indexed and eligible to appear in Search with a snippet. Brands that use nosnippet or noindex controls on large parts of their site should check that they are not excluding themselves by design.
For shopping surfaces, the practical lever on the Google side is the quality of the product data a brand supplies through its merchant tools, which is the same discipline as the feed work described above.
How should a brand measure progress?
Because rankings are not published, measurement is indirect. Track whether crawlers can reach the site, whether feed validation passes, and whether the brand and its products appear and are described correctly in a fixed set of prompts that the team re-runs on a schedule. Record inaccuracies and fix their source, usually a stale page or a third-party listing. Treat the results as directional, not as a league table, since answers vary between sessions.
It also helps to separate visibility from sales. Being named in an answer is not the same as being bought, and OpenAI's product ranking factors suggest that practical matters such as stock and price will influence which items surface. Keep a log of cases where an assistant described the brand incorrectly, because these point to specific pages or listings that need fixing. Over several months that log becomes the most useful evidence a brand has about its standing in these systems, and it costs little to maintain.
Finally, set expectations internally. These surfaces are changing quickly, publishers revise their documentation, and what is true in October 2026 may not hold next year. Revisit the lever list every quarter, and avoid long contracts with vendors that promise rankings no publisher has documented.
Frequently asked questions
How do I get my fashion brand into ChatGPT shopping results?
Allow OAI-SearchBot, keep a complete and current product feed, and apply through OpenAI's merchant process if you want checkout. OpenAI says placement is not guaranteed, so focus on eligibility and data accuracy.
Does blocking GPTBot remove my site from ChatGPT search?
Not according to OpenAI. GPTBot relates to training use, while OAI-SearchBot governs inclusion in ChatGPT search, and the two settings are independent.
Do I need special markup or an llms.txt file for Google AI Overviews?
Google says no additional requirements apply beyond being indexed and eligible for a snippet, and that no new machine-readable files or special markup are needed.
What data does ChatGPT want in a product feed?
OpenAI's specification asks for an item ID, title, description, URL, brand, seller name, image URL, availability and price, with variants as separate rows. Return fields and identifiers such as GTIN are optional extras.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
