By Carson Cobb. Data checked November 2026.
People don’t ask AI the way they search Google. They describe their situation. “My skin gets red and stings with most moisturizers.” “My small dog coughs on a collar.” “I sleep hot.” Then they ask what to buy.
The AI answers by looking for products whose pages, reviews and coverage match that situation. When a shopper clicks through, your product page has to confirm the AI’s answer in seconds. If the page doesn’t say who the product is for, how it fits or what it’s tested for, the AI has less to go on and the shopper has less reason to stay.
So I checked 94 product pages from US Shopify stores for the content that answers those questions.
The short version
- Most pages cover the basics. 79% mention ingredients or materials, 79% returns or a guarantee, 74% shipping or delivery.
- Few explain how to use the product. 23% mention how to use, directions or care instructions.
- Few say who it’s for. 40% use language like “made for”, “designed for” or “best for”.
- Comparisons are rare. 16% compare the product to anything, even another product of their own.
- Sizing help is rare outside apparel. 63% of apparel pages link a size or fit guide. Across all pages it’s 16%.
- These are upper bounds. I counted any mention on the page, including menus and footers, so the share that actually answers each question on the product itself is lower.
What product pages cover
| Content on the product page | Pages |
|---|---|
| Ingredients, materials or fabric | 79% |
| Returns, refunds or a guarantee | 79% |
| Shipping or delivery | 74% |
| FAQ | 71% |
| A visible review count | 67% |
| Subscription option | 67% |
| Who it’s for (“made for”, “best for”) | 40% |
| Certifications or testing | 38% |
| How to use, directions or care | 23% |
| Size or fit guide | 16% |
| A comparison | 16% |
The median page covered 6 of these 11. Only 4 of 94 covered 9 or more.
The pattern is clear. Pages cover the questions a store expects: what’s in it, what it costs to ship, can I return it. They skip the questions a shopper asks AI: is it right for me, how do I use it, how does it compare, will it fit.
Why those gaps matter for AI
In my testing, the way a question is worded changes the recommendation more than anything else. When I rewrote short questions like “best weighted blanket” into the long, specific questions shoppers actually type, ChatGPT kept only half of the brands it had named before. See how question wording changes AI recommendations.
The long questions I wrote, modeled on how customers describe themselves in reviews, asked about:
- Their situation and who the product suits: sensitive skin, a sensitive stomach, sleeping hot, wide feet, a nervous rescue dog.
- Proof: third party testing, heavy metal testing, evidence that it works.
- Fit and size: fits a car cup holder, a small dog that backs out of harnesses, wide feet.
- Diet and ingredients: gluten free, nut free for school, low sugar.
- Comparisons: whether an eco friendly product cleans as well as the big name brand.
Those map directly to the content that is rarest on product pages: who it’s for (40%), testing (38%), how to use (23%), fit guides (16%) and comparisons (16%). When AI can’t find that answer on your page, it relies on reviews, Reddit threads and other sites to decide whether your product fits the shopper, or picks a competitor whose page says it plainly.
By product type
| Product type | Who it’s for | How to use | Size or fit guide | Testing or certification | Comparison |
|---|---|---|---|---|---|
| Apparel | 38% | 13% | 63% | 13% | 6% |
| Beauty and skincare | 38% | 63% | 0% | 38% | 19% |
| Food and beverage | 25% | 0% | 0% | 56% | 6% |
| Home | 56% | 19% | 13% | 31% | 25% |
| Pet | 40% | 13% | 20% | 13% | 20% |
| Supplements | 47% | 33% | 0% | 80% | 20% |
Supplements led on testing (80%), which fits a product where shoppers ask about third party testing constantly. Beauty led on how to use (63%). Food and beverage pages said the least about who the product suits (25%) and none explained how to use or serve the product, even though food questions to AI are often about diets, allergies and occasions. Pet pages rarely explained fit or use, despite questions like harness fit and calming chew timing being exactly what dog owners ask.
Reviews on the page
67% of pages showed a review count. Reviews matter twice: they’re the social proof a shopper checks, and they’re one of the main places AI learns who a product suits, because customers describe their own situation in them. When a page has no reviews visible, AI leans on what other sites say. See star ratings and reviews in AI answers.
What to do with this
- Collect your customers’ real questions. Pull them from reviews, support tickets, chat logs and returns reasons. Write them the way customers say them.
- Answer them on the product page, in plain sentences. Who it’s for and who it isn’t for, how to use it, how it fits, what it’s tested for and how it compares. Put the answers on the product page itself, not only on a separate FAQ page.
- Add one honest comparison. Against the obvious alternative, or between your own products. Shoppers ask AI to compare, and AI prefers to quote a clear source.
- Show the proof. Testing, certifications and review counts, visible on the page, with the details a skeptical shopper would look for.
- Check it from the shopper’s side. Ask ChatGPT, Gemini and Google AI Mode the long version of a question your customers ask, then click through to your page and see if it confirms the answer within a few seconds.
Building the buyer questions for each product, and finding where the page falls short, is the core of my AI Search and Conversion Audit. For a quick read on one page, run the free AI Visibility Check.
How this was done
- Pages: one product page from each of the 94 US Shopify stores in my AI Readiness Index, across apparel, beauty and skincare, food and beverage, home, pet and supplements.
- Method: each page was loaded in a browser on November 5, 2026, and its visible text was checked for words and phrases tied to each content type, such as “how to use”, “directions”, “size guide”, “made for”, “third party”, “certified”, “compare” and “returns”.
- Limits: the check counts a mention anywhere on the page, including menus, footers and pop ups, so these figures are upper bounds. It can’t judge whether the content is good, only whether it’s there. One product page per store.
- Long questions: the 24 long questions are described in how question wording changes AI recommendations.
No store or brand is named in this report.