More and more shoppers are asking ChatGPT or Google’s AI Mode what to buy, and the products that get recommended tend to be the ones whose details are easiest for those tools to read and trust. Most online shops already hold the information AI needs but don’t present it clearly, so this guide looks at how to fix that before the Christmas rush.

AI is now part of how people shop
AI search has moved from novelty to habit remarkably quickly. A few figures show the scale of the change:
- ChatGPT now has 1.2 billion weekly users, according to OpenAI’s DevDay announcements in September 2026
- Google’s AI Mode passed 1 billion monthly users within a year of launch, and AI Overviews now reach more than 2.5 billion people a month, Google said at I/O 2026
- In the UK, referral traffic from AI platforms to retail sites grew by 1,439% between October 2024 and August 2026, according to Adobe Digital Insights
- The same Adobe research found that 37% of UK consumers use AI assistants to shop, mainly to compare prices, find deals and assess product features
Crucially, these visitors buy. Adobe found that in May 2026, AI-referred traffic to UK retail sites converted better than traditional web search for the first time, and by August it held a 20% lead over non-AI traffic. In the US, traffic from generative AI tools to retail sites rose by 693% year on year over the 2025 holiday season. With Black Friday and Christmas approaching, AI search is a channel worth being ready for.
If you’d like a refresher on the terminology, our guide to the difference between AEO and GEO explains both.
How AI tools decide which products to recommend
When someone asks an AI assistant for “a quiet dishwasher under £500 that fits a small kitchen”, it doesn’t simply match keywords. It breaks the question into smaller searches, gathers product information from several sources and weighs up which options best fit the request.
Microsoft’s guide for retailers, From discovery to influence, describes three sources of product data that AI systems draw on:
- Crawled data: what AI tools learned in training and retrieve from indexed pages, which shapes how they see your brand, categories and reputation
- Product feeds: the structured data you send to platforms such as Google Merchant Center and Microsoft Merchant Center, which supplies prices, stock levels and specifications
- Your live website: what AI browsers and agents see when they visit, including reviews, delivery estimates, promotions and the checkout itself
These sources are closely tied to traditional search. Research by Peec AI, published by Search Engine Land, analysed more than 43,000 products in ChatGPT’s shopping carousels and found that 83% matched products in Google Shopping’s organic results. Products ranking higher in Google Shopping also tended to appear earlier in ChatGPT’s carousel. In other words, strong product data on Google is one of the most direct routes into AI shopping results.
Get your product schema right
Product schema (structured data) is code on your product pages that describes each product in a format machines can read without guessing. It’s the clearest way to tell Google, ChatGPT and other AI tools what you sell, what it costs and whether it’s in stock.
Most online shops have some product markup, often added automatically by WooCommerce, Magento or Shopify, but it’s frequently incomplete or duplicated by competing plugins. We cover the common problems in the hidden SEO mistakes on most product pages. For AI visibility, you should aim to include:
- Core identifiers: product name, brand, SKU and GTIN (barcode number), so AI tools can match your product to the same item elsewhere.
- Offer details: price, currency, availability and condition, kept in step with your stock system.
- Reviews: AggregateRating and Review markup based on genuine customer reviews.
- Delivery and returns: shipping costs, delivery areas and your returns policy.
- Variants: size, colour and material for products that come in options.
- Collections: ItemList markup on category pages, so AI understands how your products are grouped.
Here’s a simplified example of complete product schema in JSON-LD, the format Google recommends:
{
"@context": "https://schema.org",
"@type": "Product",
"name": "Harlow Oak Bedside Table, 3 Soft-Close Drawers",
"description": "Solid oak bedside table with three soft-close drawers, sized for smaller bedrooms.",
"sku": "HAR-BST-03",
"gtin13": "5012345678900",
"brand": { "@type": "Brand", "name": "Harlow Home" },
"image": "https://www.example.co.uk/images/harlow-bedside-oak.jpg",
"aggregateRating": {
"@type": "AggregateRating",
"ratingValue": "4.7",
"reviewCount": "212"
},
"offers": {
"@type": "Offer",
"url": "https://www.example.co.uk/harlow-oak-bedside-table/",
"priceCurrency": "GBP",
"price": "189.00",
"availability": "https://schema.org/InStock",
"itemCondition": "https://schema.org/NewCondition",
"shippingDetails": {
"@type": "OfferShippingDetails",
"shippingRate": { "@type": "MonetaryAmount", "value": "0", "currency": "GBP" },
"shippingDestination": { "@type": "DefinedRegion", "addressCountry": "GB" }
},
"hasMerchantReturnPolicy": {
"@type": "MerchantReturnPolicy",
"applicableCountry": "GB",
"returnPolicyCategory": "https://schema.org/MerchantReturnFiniteReturnWindow",
"merchantReturnDays": 30,
"returnFees": "https://schema.org/FreeReturn"
}
}
}Two rules matter as much as the markup itself. First, the facts in your schema must match what shoppers see on the page and what’s in your product feed. Second, never show search engines or AI crawlers different content from what customers see; Microsoft’s guide is explicit that the rendered page should contain the same facts consumers see. You can check your markup with Google’s Rich Results Test.
Connect Google Merchant Center and search console
Schema tells Google about your products when it crawls your pages. A Google Merchant Center feed sends the same information directly, so Google doesn’t have to wait for its next crawl to learn about a price change or a product going out of stock. Google’s own product structured data documentation says that providing both structured data and a Merchant Center feed maximises your eligibility for shopping experiences and helps Google understand and verify your data.
Google Search Console then becomes your source of truth for how product pages perform in search. Alongside the standard Performance report, three reports are worth checking regularly:
- Merchant listings: flags structured data errors that stop product pages appearing in Google’s free shopping listings.
- Product snippets: shows issues with the product details, such as ratings and prices, that appear in standard search results.
- Merchant opportunities: available once you link Search Console to your Merchant Center account, this highlights ways to improve how your shop appears on Google, such as adding shipping and returns information.
Google is also introducing AI performance insights in Merchant Center, which show how products are discovered in AI Mode, AI Overviews and the Gemini app, including share of voice against similar brands and which product attributes are missing. It was announced in May 2026 for the US, Canada, Australia, India and New Zealand first, so UK retailers who get their feeds in order now will be well placed if and when it arrives here.
If you also want visibility in Copilot and Bing, the same principles apply to a Microsoft Merchant Center feed.
Write product content that answers real questions
AI assistants interpret what people mean rather than just the words they type. Someone might ask for “a waterproof jacket for a three-day hike that packs down small”, and the products that get recommended are the ones whose content clearly matches that situation.
- Descriptive titles: pair the product name with its key differentiator, for example “Harlow Oak Bedside Table, 3 Soft-Close Drawers” rather than just “Harlow Bedside”
- Front-loaded benefits: open descriptions with who the product is for, what problem it solves and what makes it better than the alternatives
- Use cases: state when and where the product works best, such as “ideal for small bedrooms” or “suitable for outdoor use all year round”
- Specifications as data: show dimensions, materials and compatibility as clear lists or tables rather than burying them in paragraphs
- Comparisons: include simple tables comparing related models, so AI can explain the difference between them
- Images and video: write alt text that actually means something and provide transcripts for product videos
Avoid copying manufacturer descriptions word for word. If hundreds of shops use the same text, AI has no reason to single yours out. Much of this comes down to structure rather than volume, which we explore in most businesses don’t have a content problem, they have a structure problem.
Add well-structured FAQs (and FAQ schema)
FAQs are one of the easiest ways to give AI tools self-contained answers they can lift and cite. Questions such as “Which size should I choose?”, “Is it suitable for sensitive skin?” or “How long does delivery take to Scotland?” mirror what shoppers actually ask AI assistants.
The quality of those answers matters. A list of generic questions with one-line answers rarely gets used. In our guide to writing better FAQs for LLMs, we set out a three-layer approach: a short, extractable answer first, then evidence and specifics, then a clear next step. For product and category pages, that means:
- Writing questions in full, the way customers would phrase them
- Giving each question its own heading and keeping answers in the page’s HTML, so they can still be read inside an accordion
- Placing FAQs on the product or category page where the question naturally comes up, rather than on one catch-all FAQ page
- Keeping facts such as delivery times, returns windows and warranties identical everywhere they appear
- Adding FAQPage schema that matches the visible questions and answers exactly
It’s worth being realistic about FAQ schema. Since 2023, Google has largely stopped showing FAQ rich results for commercial websites, so the markup is unlikely to change how your page looks in search. Its value now lies in making your questions and answers unambiguous to the systems reading your page, and Microsoft’s guide still lists FAQ among the schema types retailers should use. The visible content does the heavy lifting; the schema confirms it.
Build trust signals AI can check
AI tools favour products and retailers they can verify. Microsoft’s guide warns that exaggerated or unverifiable claims count against you, while factual, consistent information builds credibility.
- Genuine reviews: collect reviews consistently and mark them up. Review volume and recurring themes, such as “true to size” or “easy to assemble”, give AI natural language it can repeat. Our post on online reviews covers why they matter
- Clear policies: make delivery, returns and warranty information easy to find on every product page
- Certifications and awards: state them as plain facts, such as “B Corp certified” or “FSC-certified timber”, and link to the awarding body where possible
- Third-party coverage: link to independent reviews and press features that mention your product
- Consistency: make sure prices, brand details and product information match across your site, feeds, marketplaces and social profiles
Keep this information current, too. Outdated prices or discontinued products undermine trust with shoppers and machines alike, as we explain in outdated website content is costing you search rankings right now.
Make sure your live site works for people and AI agents
AI tools are starting to do more than recommend. Google, OpenAI and Microsoft are all building agents that can visit shop websites, compare options and help complete purchases. Microsoft’s guide makes the point plainly: if your live site doesn’t work properly, the sale fails, however good your product data is.
Pages that are fast, accessible and logically structured are easier for both shoppers and machines to use. Accessible websites are easier for AI to read, and the same issues that make ecommerce customers abandon their cart, such as hidden costs, slow pages or a confusing checkout, will also trip up an AI agent. A user experience and conversion review is a good way to find and remove that friction.
Where to start
You don’t need to rebuild your shop to improve AI visibility. Most retailers already hold the data AI tools need; it just isn’t presented clearly. A sensible order of work is:
- Audit your product schema and fix the errors flagged in Search Console’s Merchant listings and Product snippets reports
- Set up or clean up your Google Merchant Center feed and link it to Search Console
- Rewrite titles and descriptions for your best-selling products around real customer questions
- Add well-structured FAQs to key product and category pages
- Check that delivery, returns and review information is complete and consistent everywhere
- Run realistic product queries in ChatGPT, Gemini and Copilot to see where you appear, using the approach in how to tell if your organisation is visible in AI search
At We Create Digital, our SEO, AEO and GEO team audits product data, schema and feeds, and our ecommerce website team builds WooCommerce, Magento and Laravel shops with clean, AI-readable structure from day one.
Product visibility in AI search FAQs
Does product schema help my products appear in ChatGPT and Google AI Mode?
Product schema helps your products appear in AI search because it gives Google, ChatGPT and other AI tools a clear, machine-readable record of each product’s name, price, availability, brand and reviews. It doesn’t guarantee a recommendation, but products without it are harder for AI systems to understand, verify and compare.
Google combines product schema with Merchant Center data to power its shopping results, and those results feed into other AI tools. A Peec AI study found that 83% of products in ChatGPT’s shopping carousels matched Google Shopping’s organic results, so markup that improves your Google Shopping visibility can also improve your chances in ChatGPT. Microsoft’s guide for retailers recommends Product, Offer, AggregateRating, Review, Brand, ItemList and FAQ schema alongside accurate feeds.
Start by testing a few product pages in Google’s Rich Results Test and checking the Merchant listings report in Search Console for errors. If your markup is missing key fields or conflicting between plugins, We Create Digital’s technical SEO team can audit and fix it across your catalogue.
Do I need a Google Merchant Center account if my product pages already have schema?
You don’t strictly need a Google Merchant Center account, as product schema alone can make pages eligible for merchant listings in Google Search. However, Google says that providing both structured data and a Merchant Center feed maximises eligibility, so most online shops will benefit from running the two side by side.
A Merchant Center feed sends prices, stock levels and product details to Google directly rather than waiting for your pages to be crawled, which helps keep information accurate in AI Mode and AI Overviews. Linking it to Search Console unlocks the Merchant opportunities report, and Google’s new AI performance insights, announced in May 2026, report on share of voice across Google’s AI shopping surfaces, starting outside the UK.
Whichever route you take, your feed, schema and on-page content must show the same prices and availability. We Create Digital can help connect and check your feed as part of our ecommerce website and SEO support.
How can I tell if AI search is sending sales to my online shop?
You can see whether AI search is sending sales to your online shop by tracking referral traffic from AI platforms such as chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com in Google Analytics 4, then reporting on their conversions and revenue alongside your other channels.
Set up a custom channel group in GA4 so AI referrals are reported as their own channel rather than mixed into general referral traffic. Bear in mind that clicks from Google’s AI Overviews and AI Mode arrive as standard Google organic traffic, and some AI visits come through with no referrer at all, so GA4 will undercount AI’s true influence. Adobe’s UK data, which shows AI-referred visitors converting 20% better than other traffic, is a good reason to measure it properly.
Pair your analytics with simple checks: ask new customers how they found you and run regular product searches in AI tools to see where you appear. We Create Digital can support with GA4 reporting that separates out AI traffic, so you can see exactly what it’s worth to your shop.





