● Find out for free what AI says about your business
Saykora
GEO guide by sector · E-commerce

E-commerce: how to get your products recommended by ChatGPT, Gemini and Claude

Updated 27 September 2026

“Which brand of hard-wearing walking boots should I buy?”: AI assistants now act as shopping advisers. To be part of their answers, a shop has to provide complete product information, useful comparisons and reliable reviews. This guide sets out the levers specific to online retail.

What AIs check before recommending a product

  • Product pages on shops and marketplaces: specifications, price, availability.
  • Reviews and comparisons published by the press and specialist sites.
  • Review platforms and the reviews shown on product pages.
  • Forums and discussions between users.
  • Product feeds sent to search engines, such as Google Merchant Center.

Pages to create or improve

  • Complete, original product pages: dimensions, materials, compatibility, care, recommended use, rather than a copy of the manufacturer’s description.
  • Buying guides that answer real questions (“which size should I choose?”, “what is it for?”).
  • Honest comparison pages between your own ranges, and with other products as long as they stay objective and verifiable.
  • Clear delivery, returns and warranty pages: timescales, costs, cancellation rights.
  • An “About us” page: who is selling, where products ship from, customer service.
  • Product and Offer structured data (schema.org), with price, availability and, where they exist, genuine reviews.

Customer reviews: a strategy that plays by the rules

  • Send a review request to every buyer once their order has arrived, without filtering out unhappy customers.
  • Publish negative reviews as well as positive ones: a uniformly perfect profile isn’t credible.
  • State on your website whether and how reviews are checked: it’s a legal requirement in France.
  • If you use a third-party review platform, follow its rules on incentives.

Mistakes to avoid

  • Product information shown only via JavaScript or in images: some crawlers read it poorly.
  • Pages for out-of-stock products deleted without a redirect, or left with no availability information.
  • Different prices on your website, your product feed and the marketplaces.
  • Check your robots.txt file: blocking OAI-SearchBot, Claude-SearchBot or PerplexityBot can keep you out of answers that use web search. You can refuse the training crawlers (GPTBot, ClaudeBot) without blocking the search ones.
  • Reviews imported from another product or another shop.

E-commerce checklist

  • Original, complete product pages for your best sellers.
  • Buying guides that answer customers’ questions.
  • Clear delivery and returns pages.
  • Valid, up-to-date Product and Offer structured data.
  • Product feed consistent with your website.
  • Reviews collected from every buyer, with the checking method shown.
  • AI search crawlers allowed in robots.txt.

Measure your visibility in AI answers

Ask shoppers’ questions, by category and by budget, and note the brands and shops cited as well as the sources. Saykora’s free test asks ChatGPT, Gemini and Claude with web search on, identifies the competitors named instead of you and the sources used, and suggests an action plan. To track several shops, see our pricing.

FAQ

Do I need to be on the marketplaces?

Not necessarily, but they are often cited. If you’re not on them, your own product pages need to be at least as complete.

Is Product markup enough?

It clarifies price and availability, but assistants mainly recommend based on explicit content: specifications, comparisons, reviews.

Do AIs take special offers into account?

They may reuse a price they found online, sometimes an old one. Keep prices and availability in sync everywhere.

Sources consulted (7)

Other sectors: Restaurants · Hotels · Clinics and medical practices · Law firms · Accountants · Estate agents · B2B SaaS

Try it free →