By Carson Cobb. Data checked November 2026.
Most brands ask the same question first: “Does ChatGPT recommend us?” The more useful question is what it recommends in your category, how many brands it names, and how fixed those picks are. So I asked ChatGPT 48 buyer questions across six categories, the way shoppers actually ask them, and recorded every brand it named.
The short version
- ChatGPT names about 4 brands per answer. Median 4, up to 8. Being in the answer means being one of a handful.
- Smaller brands do get picked, often first. 27 of 48 answers named at least one growing Shopify brand from my dataset, and 14 of 44 top picks were one of them.
- Not every question gets brands. 4 of 48 answers named none, giving general advice or asking a follow up question. Health questions did this most.
- The top pick is stable. The rest of the list isn’t. In 11 repeated questions, the first pick stayed the same 10 times, while about 1 in 8 of the other brands changed.
- Store readiness barely separated who got named. Stores ChatGPT named scored only slightly higher on the AI Readiness Index than stores it didn’t.
What ChatGPT named, by category
| Category | Questions | Answers naming brands | Brands named | Answers naming a brand from my Shopify dataset |
|---|---|---|---|---|
| Food and beverage | 8 | 8 | 39 | 6 |
| Apparel | 8 | 8 | 36 | 4 |
| Home | 8 | 8 | 35 | 5 |
| Pet | 8 | 8 | 29 | 5 |
| Supplements | 8 | 6 | 23 | 6 |
| Beauty and skincare | 8 | 6 | 20 | 1 |
| All | 48 | 44 | 182 | 27 |
Food, supplements and pet were the most open to growing brands. Questions about spices, coffee, hot sauce, training treats, kids vitamins, electrolytes and mushroom coffee all put a growing direct to consumer brand at or near the top.
Beauty and mass categories leaned on household names. Foundation, lip stain, cereal and running shoes went mostly to brands sold everywhere. When a category is dominated by big retail brands, the editorial lists ChatGPT reads are too.
Some questions got no brands at all. Ingrown hair treatment, acne cleansers, perimenopause supplements and “a drink for focus” got advice about ingredients or a follow up question instead of products. For health adjacent categories, ChatGPT often answers the condition first and the product second.
Broad versus specific questions
I wrote 20 broad questions (“best creatine brands”) and 28 specific ones (“best multivitamin for picky eater kids”).
| Question type | Answers naming brands | Median brands named | Answers naming a brand from my dataset |
|---|---|---|---|
| Broad | 20 of 20 | 5 | 15 of 20 |
| Specific | 24 of 28 | 4 | 12 of 28 |
Broad questions produced longer lists, which gave more room to smaller brands. Specific questions produced shorter lists and more product level answers, and they were the ones most likely to return advice instead of brands. The common belief that smaller brands win mostly on niche questions didn’t hold in this sample. They won where their category’s review sources already featured them.
How stable are the picks?
I asked 11 of the questions a second time in a fresh temporary chat.
- The top pick matched in 10 of 11.
- About 87% of the brands from the first answer appeared again in the second.
- The changes happened in the middle and bottom of the list. In one category a growing brand from my dataset dropped out of the second answer, and in another a different one appeared.
If your brand sits in positions 3 to 6, you’re in the zone that changes from one conversation to the next. That’s normal, and it’s why one screenshot proves very little either way.
Does an AI ready store get recommended more?
18 of the 94 stores in my AI Readiness Index were named somewhere in these answers.
| Group | Stores | Median readiness score | Average score |
|---|---|---|---|
| Named by ChatGPT | 18 | 72.5 | 69 |
| Not named | 76 | 70 | 63 |
The named stores scored a little higher on average, but the overlap is large. Named stores ranged from 40 to 90. A third of them had duplicate product data, about the same rate as everyone else. A small sample and questions chosen around the stores’ categories mean this shouldn’t be read as cause and effect.
What it does suggest: your own site’s readiness is not what gets you into the answer. What gets you in is being featured by the sources ChatGPT reads for that category. Your site’s readiness decides whether AI describes you correctly once you’re there. See the sources ChatGPT cites.
What to do with this
- Ask ChatGPT your customers’ real questions. Use five to ten, broad and specific, and note who gets named and which sources it shows.
- Ask twice. If you’re named once and not the second time, you’re in the moving middle of the list.
- Find the sources behind the answers. The lists, reviews and experts ChatGPT cites in your category are where to earn a mention.
- Then make sure AI describes you correctly. Clean product data and clear product pages decide what it says once you’re in the answer.
How this was done
- Questions: 48 buyer questions, 8 per category, in apparel, beauty and skincare, food and beverage, home, pet and supplements. 20 broad and 28 specific, written in everyday language. Questions were chosen around product types sold by stores in my 94 store sample, which makes those stores more likely to appear than in a random set of questions.
- How they were asked: in ChatGPT on a paid plan, with the default model and default thinking setting, each in a new temporary chat, in November 2026. ChatGPT can still use approximate location and account settings in temporary chats. A few answers included brands local to where the test was run, and those were left out of the dataset comparisons.
- Brands: recorded from each answer’s ranked list or product cards. “Brands from my dataset” means brands with a store among the 606 US Shopify stores behind my other reports.
- Stability: 11 questions asked again in a fresh chat on the same day. A 12th repeat didn’t finish in time and is left out.
- Readiness comparison: stores from the 94 store AI Readiness Index named anywhere in the 48 answers, compared with the rest.
No store or brand from the dataset is named in this report. Run the free AI Visibility Check to see whether ChatGPT names your brand.