Nobody Talks About Your Brand Yet. AI Can Still Recommend Your Products

The short answer

Yes, for product questions. When a shopper asks ChatGPT or Google AI Mode for a waterproof hiking boot under $200, the assistant is not looking for who has the most press. It is assembling a shortlist from product sources it can read: merchant feeds, product pages and the structured facts on them. A brand nobody has reviewed yet can appear on that shortlist if its product information is where the assistant looks and says what the shopper asked about.

The honest limit: for "which brands are best" questions, mentions and reviews matter more, and a new brand will trail there for a while. Most shopping conversations are not that question. They are specific, and specific is where a cold-start brand can compete now.

What assistants actually read for a product answer

Two mechanisms matter, and neither depends on brand fame.

The first is feeds. Google's AI performance insights report in Merchant Center measures how a merchant's products surface for conversational shopping queries on AI Mode and AI Overviews, and its traffic filter is limited to organic AI traffic from free listings. In plain terms, products in Merchant Center are eligible for Google's AI shopping answers without an ad budget. On the OpenAI side, ChatGPT's product feed specification is how merchants put their catalog in front of ChatGPT's shopping features directly.

The second is retrieval of web pages. Ahrefs studied 1.4 million ChatGPT prompts and found that the assistant retrieves dozens of URLs per answer and cites about half of them. Before it reads any page, it screens on title, snippet and URL, and titles that match the sub-questions it generates get picked more often. The average cited page was around 500 days old. None of those gates is "has this brand been written about."

Where a cold-start brand loses

The pattern we saw in September conversations with ecommerce leaders was consistent. A brand with a small catalog and no coverage asks the assistant for the best option in its category, sees three competitors, and concludes the game is rigged toward big names. When we looked at the product pages behind those examples, the problem was usually narrower: the page did not say the thing the shopper asked about.

Google's report now shows this directly. Its popular attributes scorecard lists the structured specifications shoppers ask for in a category, such as size, material or dimensions, that may be missing from the merchant's product data, and its top terms scorecard lists the words shoppers use, with examples like "maximum cushioning" or "arch support". Google's own recommendation is to add those terms to titles and descriptions and populate the missing attributes, starting with the most frequent ones.

A brand with a thousand reviews can be sloppy about this and still appear. A brand with none cannot.

Three moves that do not require anyone to mention you

1. Be in the feeds, with complete products

Merchant Center with free listings enabled, and the ChatGPT product feed if you sell in a supported market. Check the basics the assistant reads before it reads anything else: price and availability that match the page, a description that is not the title repeated, a return policy, GTINs where they exist. Then go through the attribute list for your category and fill the ones the Google report flags as popular and missing.

This is the single highest-leverage task for a brand with no reputation, because it is the part of the shortlist that is decided on data rather than on authority.

2. Title pages the way shoppers ask

The Ahrefs finding about titles applies to product and collection pages, not only to blog posts. A collection page called "Boots" competes badly for "wide-fit waterproof hiking boots for men". A page that says exactly that, with the fit, waterproofing and weight details in the first screen, gives the assistant a reason to open it. Keep one page per real shopper question, and do not spin out dozens of near-duplicate pages. Google's scaled content abuse policy covers pages produced at scale to manipulate rankings, however they are made.

3. State the facts reviews would otherwise carry

When a brand has no reviews, the assistant cannot borrow "runs small" or "the zipper failed after a month" from shoppers. Put the facts on the page yourself: fit guidance with measurements, materials with the actual composition, care, what is in the box, warranty terms. These are the details the assistant uses to decide whether your product answers the question, and they are the details a competitor with a famous name often leaves out.

What to measure while you wait for reputation

Use the free sources first. Merchant Center's AI performance report shows share of voice against the competitor set Google assigns, by shopping stage, with products showing per term. A zero means the products did not appear, and a dash means no impression data yet, which is the honest starting point for a new brand.

Track the same tracked questions over time rather than spot checks, because answers change between runs. Semrush's 2026 AI Visibility Index reported that 45% of marketing leaders cannot measure their brand's visibility in AI answers. A new brand does not need a tool to beat that. It needs the Google report, a fixed list of questions, and a monthly comparison.

Reputation will come, and when it does it compounds. Until then, the shortlist is decided on whether your product information answers the question, and that part is entirely in your hands.

This is the work Paz does for lean ecommerce teams. A Paz expert tracks the shopping questions that matter for your buyers, finds the products that are missing from the answers, fixes the product information and feed fields behind each gap, and reports appearances over time. You approve the guidelines, the platform publishes and tracks. If your brand is new and you want to know where you stand, talk to us.