Let’s talk

Marketing Insights

How should a product page read for AI shoppers?

Sam Hill
·
·
6
MIN READ
A grid of 100 dots with 32 filled, showing 32% of shoppers use AI for shopping advice

Write each product page so a machine can lift every fact without guessing: name, price in AUD, availability, delivery time, returns terms and who the product is for, all in plain text near the top, backed by Product structured data. Australia Post found 32% of shoppers already use AI for shopping advice.

Write each product page so a machine can lift every fact without guessing: name, price in AUD, availability, delivery time, returns terms and who the product is for, all in plain text near the top, backed by Product structured data. Australia Post found 32% of shoppers already use AI for shopping advice.

Black Friday is on 27 November, so this is the week to make the change. The work is a template edit, not a rebuild.

What do AI shoppers need from a product page?

They need the same facts a careful human needs, stated plainly. Nobody outside the AI companies can say exactly how each assistant reads a page, so we stick to what is documented and what is sensible.

The demand is real. The Australia Post eCommerce Report 2026 found that 6 in 10 shoppers are comfortable using AI and 32% already use it for shopping advice. We unpacked what that means for visibility in A Third of Shoppers Ask AI What to Buy.

Shoppers already asking AI
A grid of 100 dots with 32 filled, showing 32% of shoppers use AI for shopping advice.

Which facts belong in plain text?

Anything a shopper would ask about before buying. If it only appears inside an image, a PDF or a pop-up, an assistant may never see it.

Put these on the page as readable text:

  • Product name, model and variant, spelled the same way everywhere.
  • Price in Australian dollars, and the sale end date if there is one.
  • Whether it is in stock, and when it ships.
  • Delivery cost and time to a Melbourne postcode and to a regional one.
  • The returns window and who pays for return postage.
  • Size, materials, ingredients or specifications, with units.
  • Who it suits and who it does not, in one or two sentences.

The last point does the most work. An assistant answering “best moisturiser for sensitive skin under $60” needs a page that says what the product is for. Vague copy gets skipped.

Facts to put in plain text
Four rows listing price, stock, delivery and returns facts that belong in plain text.

What markup should the page carry?

Product structured data, written to match what the shopper can see. Google says that adding Product markup can make a page eligible for merchant listing experiences that show price, availability, shipping and return information.

Some technical rules from Google’s documentation are worth knowing. Product markup supports pages that focus on a single product or its variants, not category pages. For merchant listings the price currency is required, and a different currency needs a different URL. Google does not guarantee that any rich result will appear.

Google also has separate markup for merchant return policies and shipping policies. Our view is that the markup must never say something the page does not. If the visible text says 30-day returns, the data says 30-day returns.

How should sale prices be marked up?

Mark the current price as the active price and the original price as a strikethrough price. Google’s documentation says the strikethrough price is set with the priceType property, that a current price must also be given, and that validFrom and validThrough can carry the sale dates.

For Black Friday, that means three jobs. Set the sale price and its end date. Change both back when the sale ends. Check that the number in the markup matches the number on screen, because a mismatch defeats the point of the markup.

Marking up a Black Friday price
Marking up a Black Friday price: set the active and strikethrough prices, add dates, switch back, check the match.

What about delivery information?

Put cut-off dates in text on the product page, not only at checkout. Australia Post’s October update says 71% of shoppers are less likely to buy if delivery information at checkout is unclear. A page that states the last order date for Christmas delivery gives an assistant a fact to quote and gives the shopper certainty.

How do you test it?

Ask ChatGPT, Gemini and Perplexity about your product in the words a customer would use, then ask each to describe your product directly. Check whether the price, availability and returns terms are right. Fix the page, not the prompt.

Our guide to answer-shaped content shows how to structure the copy, and FAQ schema and AI citations covers the question-and-answer layer. If you want a template change scoped for a Shopify or Webflow store, talk to us.

FAQ

Do I need structured data for AI assistants to read my product page?

Not necessarily, and no one outside the assistants can promise it helps with each one. It is documented for Google, it forces you to state the facts cleanly, and the cost is low.

Should I put specs in tabs or accordions?

Put the core facts in the main page text. Tabs and accordions can be fine, but a fact that matters to a purchase should be visible without a click.

Does this matter if my products are sold on a marketplace too?

Yes. Your own page is the one place where you control the facts. Keep the name, price and returns terms consistent with your marketplace listings.

How often should the markup be updated?

Whenever price, stock or policy changes. If your platform generates it from live product data, check the output after each sale change.