Writing Product Pages That AI Shopping Assistants Can Read
A shopper who wants new running shoes used to open three or four tabs, read reviews, compare sizes and prices, and then decide. Now they can ask an AI assistant: “Which running shoes under $200 are best for flat feet, and where can I buy them in Australia?” The assistant does the comparing and comes back with a short list.
For online stores, that short list is the new shop window. If an AI assistant can’t read your product pages clearly, or doesn’t trust what it reads, your products won’t make it onto the list, however good they are.
AI shopping assistants compare products on details such as materials, sizing and price, so those details need to be easy to find.
Alt text: A pair of minimal leather sneakers on a clean studio background next to a small swing tag
For smaller retailers, most of what helps is within your control and doesn’t depend on a big advertising budget. It comes down to how clearly your product pages describe what you sell.
How AI Assistants Compare Products
When a shopper asks for product advice, AI assistants pull information from product pages, retailer feeds, reviews and comparison content. They then try to match the details against what the shopper asked for.
That matching is literal. If a shopper asks for a waterproof jacket that packs down small, the assistant looks for pages that state the waterproof rating and the packed size. A page that says “perfect for any adventure” and leaves the specifications in an image gives it very little to match.
What they look for
| Detail | Why it matters to an AI assistant | Common gap on product pages |
| Specific attributes | Matches shopper requirements such as size, material or capacity | Specs shown only in images or PDFs |
| Price and availability | Lets the assistant recommend products the shopper can actually buy | Out-of-date stock status, price only visible at checkout |
| Use cases | Connects the product to the shopper’s situation | Generic copy copied from the manufacturer |
| Reviews | Adds evidence beyond the retailer’s own claims | Reviews loaded by a script the assistant can’t read |
| Shipping and returns | Answers “can I get it by Friday?” style questions | Policies buried in a footer link |
The details AI shopping assistants rely on, and where product pages often fall short.
Write the Page for the Question, Not the Brochure
Manufacturer descriptions are written to sell. AI assistants are trying to answer. Rewriting even your top 20 products with that in mind can make a noticeable difference, and the same changes tend to lift conversion rates for human shoppers too, because they answer questions that would otherwise go to live chat or end in an abandoned cart.
Put specifications in text
Dimensions, weight, materials, compatibility and care instructions should sit in plain text on the page, ideally in a simple table. Assistants can’t reliably read text inside product images or downloadable spec sheets.
Describe who it’s for
A sentence such as “suits wide feet and long-distance walking” answers the kind of question shoppers ask AI tools. Be specific and honest. If a product isn’t suitable for something, saying so builds trust with shoppers and gives the assistant a clear signal.
Answer the common questions on the page
Most retailers already know the questions customers ask, from emails, live chat and returns data. Add the top five as a short Q&A section on the product page. “Does this fit a 15-inch laptop?” is exactly the kind of detail AI assistants look for.
Keep it original
If fifty stores use the same manufacturer description, AI tools have no particular reason to prefer yours. Adding your own photos, sizing notes or a short comparison with similar products gives them a reason, and it also protects you if the manufacturer changes its copy.
Specifications in plain text and a short Q&A section give AI assistants details they can match against a shopper’s request.
Alt text: A laptop showing an abstract online store product page with an image, text and a specification table
Get the Structured Data Right
Structured data is code on the page that labels each detail for machines. Google’s documentation on product structured data explains how to mark up a product’s name, image, price, availability, reviews and more, and which details qualify a page for richer shopping results.
Two points trip up many stores.
Price and availability need to be accurate
Structured data that says “in stock” when the item sold out last week is worse than having no markup at all. Make sure your platform updates the markup whenever stock or price changes. If you sell through Google, the Merchant Center product data specification sets out the attributes it expects, and keeping your feed and your pages consistent avoids mismatches.
Add shipping and returns details
Newer markup options let you describe shipping costs, delivery times and return policies. These answer the practical questions shoppers often ask an AI assistant right before they buy.
Make Reviews Readable and Specific
Reviews are one of the few sources of evidence an AI assistant has that doesn’t come from you. Two things decide how useful yours are.
Can the assistant see them?
Many review apps load reviews with a script after the page opens. Some crawlers don’t run those scripts, so they miss the reviews. Check your product page’s source and see whether the review text is there. If it isn’t, ask your review provider about a server-rendered option.
Do they say anything specific?
“Love it!” helps your star rating but tells an assistant little. Prompt customers with a question in your review request email, such as “How did the sizing compare with other brands you’ve worn?”. Answers like that are the details shoppers ask AI tools about.
Don’t Forget Images and Alt Text
Some AI assistants can now read images, but they still rely heavily on the text around them. Give each product photo a descriptive file name and alt text that says what the image shows, such as “navy merino crew-neck jumper, front view, folded”. It helps shoppers using screen readers and gives machines another clear description of the product.
Price Displays Still Have to Follow Australian Law
As you tidy up prices for AI assistants, check they meet Australian requirements too. The ACCC’s guidance on price displays explains that businesses need to show a single total price including unavoidable extra fees, and that part prices can’t be shown more prominently than the total. A price in your structured data that leaves out a compulsory fee could mislead shoppers and the tools recommending you.
Where to Start if You Have Thousands of Products
Rewriting a whole catalogue isn’t realistic for most stores. A sensible order is:
- Your top sellers. The products that already earn most of your revenue.
- Your highest-margin products. Where a small lift in visibility pays off most.
- Products that attract questions. Anything that generates regular emails or returns.
- Category pages. A short, useful guide at the top of each category helps assistants understand your range.
Many stores tackle the first 50 products in-house over a few weeks. Larger catalogues often need help with templates, structured data and feed management, and specialist AI SEO services tend to earn their keep there: building page templates and markup once, so each new product added to the store inherits them.
Clear shipping and returns details answer the last questions shoppers ask before they buy.
Alt text: Cardboard shipping boxes, tissue paper and tape on a packing table
How to Check Your Progress
Pick ten of your products and ask ChatGPT, Gemini and Copilot the questions a shopper would ask about that type of product, adding “in Australia”. Note whether your store appears, which competitors do, and what details the assistant quotes.
Repeat the same questions once a month. Keep a note of the exact wording so the results are comparable, and watch your analytics for visits referred by chatgpt.com, gemini.google.com and copilot.microsoft.com. If assistants start quoting your specifications, sizing notes or Q&A answers, the rewrites are working.
The Takeaway for Online Retailers
AI shopping assistants reward product pages that state clear facts, answer real questions and back them up with accurate structured data and genuine reviews. Retailers who make their products easy to understand stand a better chance of being recommended than those relying on brand-style copy.
Start with your top ten sellers this month. Move the specifications into text, add five common questions and answers to each page, and check that price and stock data update automatically.
