A shopper who asks an assistant for the best running shoe for wide feet, or a gift under a set budget, gets a short list. The list can include product listings and product details, and Google says Merchant Center feeds can help products appear in AI responses. A store that is not in that list does not lose a click. It loses the chance to be considered.
This page explains how we handle AI SEO for ecommerce: the data on clicks, the product fields engines can use, the six jobs we do, two worked examples and a five-minute test you can run today. The wider topic sits on the AI visibility hub.
01What is different about AI SEO for ecommerce
A store has thousands of facts that must be right at once: price, stock, sizes, materials, shipping, returns, reviews. Search engines could show an old snippet and shoppers would click through to check. An AI answer states the fact itself, and the shopper may never visit. So accuracy and completeness of product facts matter more than they did for ranking, and shopper questions are longer and more specific: best for a use, within a budget, compared with another product.
The good news is that Google's guidance for AI features does not ask for anything exotic. It says there is no special schema.org markup to add, no AI text file or Markdown to create, and no need to break content into tiny pieces. Pages need to be indexed and eligible to show with a snippet, and Google points to Merchant Center and Business Profile as ways products and services can appear in AI responses and other results.
02What the click data says, and what it does not
A Pew Research Center study that tracked 900 US adults and 68,879 Google searches found people clicked a traditional result in 8% of visits when an AI summary appeared, against 15% when it did not. They clicked a link inside the summary in 1% of visits, and ended their browsing session in 26% of visits with a summary against 16% without one.
Ahrefs found AI Overviews cut clicks to the top result by about 58% in a 300,000 keyword comparison. Seer Interactive's 53-brand data saw organic click-through on AI Overview queries rise from 1.3% to 2.4% over two months, so the picture moves. Both are correlations, and both measure informational searches more than shopping. The takeaway for a store is narrow: where an AI answer appears, being inside it is worth more than being one link below it.
03The product facts AI engines can use
Engines read your page text, your structured data and your feed, then cross-check them against what reviews and other sites say. When those disagree, the answer gets vague or wrong. Put every fact in plain text on the page, and make the structured data and the Merchant Center feed say the same thing.
| Field | Why it matters | Plain-text example |
|---|---|---|
| Name and brand | Lets the engine tell variants and rivals apart | Trailhead Wide-Fit Walking Shoe, by Northfield |
| Price and availability | The two facts most often quoted wrongly | Price shown in text, stock status shown in text |
| Size, material, dimensions | Answers fit and use-case questions | Wide fit, waterproof mesh, 310 g per shoe |
| Who it is for | Matches the long buyer prompt | For walkers with wide feet and flat arches |
| Shipping and returns | Shoppers ask before they buy | Free returns, delivery times by region |
| Identifiers such as GTIN | Ties the page to the feed and to listings elsewhere | Barcode number in the product details |
| Reviews with substance | Gives the engine something to quote and trust | Verified reviews that mention fit and comfort |
04The six jobs in AI SEO for ecommerce
Open the crawlers
Check robots.txt, the CDN and the firewall for OAI-SearchBot, PerplexityBot and Googlebot. OpenAI says OAI-SearchBot controls ChatGPT search and a robots.txt change takes about 24 hours to apply. Check the page source too: if key facts appear only after scripts run, some crawlers may not see them.
Fix the product data
Make page text, structured data and the feed agree on price, stock, size and identifiers. Mismatches are the commonest cause of wrong answers about a store.
Write category and best-for pages
Shoppers ask best for a use, within a budget, compared with another product. Build pages that answer those questions with a real table, honest limits and links to the products.
Build reviews and reputation
Collect detailed reviews where your buyers read them, and respond to complaints in public. Never buy or invent reviews.
Earn mentions beyond the store
Roundups, editorial coverage, creator reviews and forums feed the answers. In an Ahrefs study of 75,000 brands, branded web mentions correlated 0.664 with AI Overview visibility, the strongest factor measured.
Measure a fixed prompt set
Test the same 20 to 40 shopper prompts every month across the engines, three runs each, and log which products and pages are named. The measurement guide shows the method.
05Product feeds, Merchant Center and structured data
Google says Merchant Center, with its feeds, can help products and services be visible in AI responses and other Search results. Treat the feed as a second copy of your product page that must never drift from the first. Google also says structured data is not required for AI features but remains useful for rich results, so it is worth doing for that reason. Do not expect markup alone to get you named. Hunting for inauthentic mentions across the web is, in Google's words, less helpful than it might seem, so the work stays on real pages and real coverage.
A five-minute self-check you can run today
0 of 7 done06A decision rule by store size
Under about 100 products, fix every product page by hand and write a category page for each real use. From 100 to a few thousand, fix the top 20% of products by revenue first, then template the fields for the rest and spot-check. Above that, fix the feed and the product template before any content, because one template error is repeated across the whole catalogue. In every case, do the five-minute check first, so you spend effort where the answers are actually wrong.
What a typical ecommerce SEO retainer gives you
- Keyword reports aimed only at the Google results page
- Product descriptions rewritten by volume with an AI tool
- Schema added to every page and counted as done
- No view of what ChatGPT or Perplexity say about your products
What we build
- A shopper prompt set, measured across five engines
- Product data that agrees across page, schema and feed
- Category and best-for pages with honest comparison tables
- A plan for reviews and mentions on sites shoppers read
This suits you if
- Your products are chosen on fit, features or use, not only price
- Shoppers in your category already ask assistants for recommendations
- You can keep product data accurate and consistent
- You want shopper prompts measured, not only rankings
Skip it if
- You only sell commodity items where price is the only difference
- You cannot publish prices, stock or return terms in text
- You need a guaranteed citation, which nobody can offer
07What it costs and how to begin
Our programme is $2,000 a month, described on the AI SEO pricing page. To see where you stand first, use the free AI visibility checker or ask for an AI visibility audit. For the engine-by-engine detail see ChatGPT SEO services, AI Overview optimization services and generative engine optimization services. The numbers behind this page are collected on AI search statistics.

