Mistake 1. Thinking AEO replaces SEO
The most common and the most expensive. AI answers through query fan-out — it expands a query into 15 sub-queries and looks for pages matching each of them in Google and Bing. If you are not cited in ordinary search results for those sub-queries, the AI will not find you. Full stop.
Kyle Risley, senior SEO lead at Shopify, puts it plainly: classic SEO is the foundation of AEO, not its rival. If you dropped technical SEO in order to “do AEO”, you drained the pool and are wondering why you cannot swim.
Mistake 2. Basic Product schema with no extended attributes
A standard Shopify theme gives you the minimum: price, availability, brand. Without material, weight, dimensions, warranty_details and GTIN, your product looks to an AI like “one of dozens of faceless options”. When a shopper asks for a “1.2-litre stainless steel cordless kettle”, the citation goes to the store that spells those attributes out in its JSON-LD.
The check: open the Google Rich Results Test, run your five top products through it and see which fields are empty.
Mistake 3. Descriptions written to advertise, not to answer
“Discover our premium collection of…” gets skipped. “320 g/m² polyester, does not build static at −10°C” gets quoted. If your descriptions open with adjectives, rewrite them to “problem → solution → proof → limitation”. Acknowledging a limitation is the single most quoted element.
Mistake 4. Never opening your llms.txt
Since May 2026 Shopify generates llms.txt and agents.md automatically at the root of every store — a machine-readable index for AI bots and agents. The mistake is not that the file is missing; it is that nobody has ever opened it. The file is assembled from your store data, so vague collection names, empty descriptions and missing policies go into it exactly as they are. Open yourstore.com/llms.txt — five minutes, and you know what AI bots actually read about your store.
Mistake 5. No measurement system
If you do not know which queries cite you and which do not, every change is a guess. The minimum dashboard: 20 key queries, checked manually once a week in ChatGPT and Perplexity, recording mentions and positions. Half an hour a week.
Without that shelf you are optimising into the dark. With it, you can see which change produced a result and scale exactly that one.
A 30-minute audit checklist
1. Type five key queries into ChatGPT and Perplexity. Are you cited? If not, which sites hold your place?
2. Open Google Search Console: is the site indexing properly, are there crawl errors?
3. Run your five top products through the Google Rich Results Test.
4. Open yourstore.com/llms.txt: is it a meaningful index of your store, or an empty stub?
5. Open the description of your best-selling product and read it aloud.
If any of those is a “no” or a “not sure”, you have at least one of the five mistakes. Fix from the bottom up: SEO foundation → schema → descriptions → llms.txt → measurement.