By Steve Merrill, Founder of WRKNG Digital | September 22, 2026
I took one Shopify store and asked three different AI assistants to shop it. ChatGPT, Perplexity, and Google's AI. Same store, same products, three very different results.
The failures weren't dramatic. They were boring. Missing fields. Blocked crawlers. A price that loaded too late. And boring is exactly why most stores never fix them.
What ChatGPT got stuck on
ChatGPT found the store but couldn't confirm stock on half the products. The availability data wasn't in the page it read, so it hedged. "You may want to check the site for current availability."
That hedge is a lost sale. A shopper reading that moves on. The fix was putting Offer availability into the raw HTML. Not glamorous. Just missing.
What Perplexity got stuck on
Perplexity is aggressive about citing sources, and it wanted specs it couldn't find. Material, dimensions, care instructions. The store had them in an image. The machine can't read an image of a spec table.
So Perplexity cited a competitor who wrote the specs as text. Same product category. The competitor just made their facts readable.
What Google's AI got stuck on
Google's AI leaned on the product feed and structured data. Where the feed and the live page disagreed on price, it went quiet on the product entirely.
One stale feed price. That's all it took to drop a best seller out of the answer. The data argued with itself and the machine walked away.
The pattern across all three
None of them failed on branding or copy. They failed on facts. Missing availability, specs trapped in images, a feed fighting the site.
Here's the bottom line: these are hygiene problems, not strategy problems. Every one was invisible to a human and fatal to a machine.
What I'd fix first
Availability in raw HTML. Specs as text, not images. Feed and site reconciled. In that order, on your ten best sellers.
I ignored a platform shift once and it cost me years. This is a cheaper lesson. Run the test yourself this week and write down where each assistant hedges. That list is your roadmap.
Further reading
Frequently Asked Questions
How do you test whether AI shopping agents can read your store?
Ask each assistant to shop your category and recommend products, then ask about specific products from your store and note where it hedges or gets facts wrong. Test ChatGPT, Perplexity, and Google's AI separately, because they weight sources differently. The spots where an assistant says "check the site" are usually missing structured data or specs it couldn't read.
Why did Perplexity cite a competitor instead of my store?
Often because the competitor made their facts machine-readable and you didn't. Specs trapped in an image, a PDF, or a graphic can't be read by an assistant, so it cites the store that wrote the same information as plain text. Convert spec tables and key details into readable text on the page to compete for the citation.
Can one wrong price in my feed hurt AI visibility?
Yes. When your product feed and your live page disagree on price or availability, an AI assistant sees conflicting facts and may drop the product from its answer rather than risk being wrong. A single stale feed price can quietly remove a best seller from AI recommendations, so keep your feed and live pages in sync.
Want to know if AI assistants can actually find and recommend your store? Get a free AI-visibility read on your Shopify store at WRKNG Digital. We show you exactly what ChatGPT, Perplexity, and Google AI see when they look at your products.
