How shopping questions are answered
When a shopper asks an assistant for a recommendation, the reply is a short piece of writing, not a page of results. It might name a few products or shops, explain what to look for, and sometimes link to sources.
To write that, the assistant draws on pages that explain choices clearly: buying guides, comparison pages, well-described categories and product pages with plain specifications.
Many shops have the opposite:
- Category pages are a grid of tiles with no text.
- Product pages repeat the manufacturer's description.
- Delivery and returns sit in a footer link.
None of that is wrong, but it gives an assistant very little to quote.
What tends to help a shop
- Buying guides that actually guide. "How to choose" pages that set out the trade-offs honestly, link to the relevant categories and are kept up to date.
- A short introduction on each main category. Two or three paragraphs saying who the range is for and how the options differ.
- Product facts in plain text. Size, materials, compatibility, care, delivery times and returns written on the page, not only in images or tabs that do not load without a click.
- Accurate structured data. Product, offer and review markup that matches what the page shows. Our answer on schema markup and AI answers explains what it does and does not do.
- Mentions elsewhere. Independent reviews and guides that name your shop. You cannot buy these into an AI answer, but you can earn them.
Where to start with GEO Agency
- Run a site audit on Pro or Agency (10 credits). It crawls up to 100 public pages and ranks the weakest, which on a shop usually means thin category pages and guides that never answer the question in their title.
- Score the handful of pages you most want quoted with page analysis (1 credit each) and work through the suggested fixes.
- Try answer simulation. It shows how an assistant might answer a shopping question and whether your page is likely to be used.
If a competitor keeps appearing where you do not, a competitor comparison (1 credit) lists the questions their page answers and yours does not. Our page on why AI recommends a competitor walks through that comparison.
Before any of that, the two-minute AI-readiness self-check is a quick way to spot obvious gaps on one page.
Measuring it sensibly
Pick ten shopping questions, ask each several times, and record whether the shop is named or linked. Repeat monthly. Sales and traffic from assistants are hard to attribute today, so treat this as an early signal alongside the numbers you already track, not a replacement for them.
Keep it current
Shops change faster than most sites: ranges rotate, prices move and delivery promises shift with the season. An assistant that read your page in spring may still describe the old range.
- Put a visible "last updated" date on buying guides.
- Retire pages for products you no longer sell rather than leaving them half-empty.
- Re-check your main category pages after each range change.
Small, regular updates do more than one big rewrite a year.
Honest limits
The tools run on Google Gemini, so scores and simulations are estimates of how an assistant might respond, not a record of what shoppers were told. We do not promise rankings, product placement or a place in any AI answer. Prices and plan details are on Pricing, VAT included.
Common questions
Which pages matter most for a shop?
Usually buying guides and category pages, because they answer "which one should I get" questions. Product pages matter for the facts: specifications, sizes, materials, delivery and returns.
Does product schema markup help?
It helps machines read your prices, availability and product details correctly. It does not, by itself, get you named. Treat it as necessary housekeeping rather than a lever.
Can I check thousands of product pages?
You do not need to. A site audit crawls up to 100 public pages and ranks the weakest; start with your main categories and guides. Each page analysis is 1 credit, so focus beats volume.
Will this show me what ChatGPT says about my products?
No. GEO Agency runs on Google Gemini and estimates how an assistant might answer and how usable your pages are as a source. It is a guide, not a log of real shopper conversations.
Which plan should a shop use?
Pro at £29 a month (VAT included, 50 credits) suits one shop. Agencies looking after several shops usually choose Agency at £129 a month for 500 credits and a team workspace.