What broke
In classic SEO the shopper typed “buy running shoes” — short, utilitarian, measurable. The tool showed 50,000 searches a month at difficulty 70, and you fought for a top-10 place on that query.
In the AI era the same shopper asks ChatGPT: “which running shoes are best for trails up to 30 km if I have flat feet and run on gravel in the rain”. A long, specific, human query — the way you would talk to a real salesperson. And the trick is that the AI fans it out into 5–15 sub-queries through query fan-out — “trail shoes for flat feet”, “footwear for running in the rain”, “how to protect your ankle running on gravel” — and looks for an answer to each one separately.
Your “keyword” no longer exists as a unit. Its place has been taken by question clusters — dozens of related questions around a single buying topic.
What works now
Question intent clusters. Instead of a keyword list with volumes, a map of the questions people ask around a product. One product = 30–80 questions. Each question is its own page or section.
Deep pages. Francine Monahan of iPullRank offers the figures: 82% of citations in Google AI Overviews come from deep pages — not the homepage, not a category, but a specific product, FAQ or blog page. If your answers sit deep in the catalogue, that is a good thing.
Fresh sources of tooling. AlsoAsked, AnswerThePublic, the Perplexity API (you can simply ask Perplexity “what are common questions about X”), ChatGPT with the instruction “expand this query into related questions”. Old SEO tools built on keyword volume do not work for AEO.
Shopify specifics
Collection pages become question hubs. Previously they were a list of products with filters. Now they are a page with a “Frequently asked questions” section at the bottom, where 8–12 shopper questions are answered in 100–150 words each. Those answers become fraggles — fragmented citations an AI pulls into its answer.
Product descriptions are written for sub-queries. If the product is running shoes, the description should answer “for what pace”, “for what runner weight”, “how to wash them”, “when to replace them”. Each sub-paragraph is one fan-out sub-query you have won.
The blog becomes a hub for conversational queries. Not “10 tips for choosing running shoes” (the old SEO formula), but “running shoes for flat feet — what actually matters and what is a myth”. A specific question in the headline means a higher chance of citation.
What to do this week
Take your top three products by sales. For each:
1. Open Perplexity and ask “common questions buyers ask about [product]”. You will get a list of 15–30 questions.
2. Open ChatGPT and ask it to “expand each question into 2–3 related queries”. That gives you another 30–60 sub-queries.
3. Compare them against your product descriptions and FAQs. How many of those questions are answered on the site? For most Shopify stores it is 20–30%.
4. Add the missing answers — 80–150 words each, wrapped in FAQ schema.
After 30 days, check the same queries in ChatGPT and Perplexity. If your store is being cited, scale the formula across the rest of the catalogue.