Ulysse HC

I check what AI assistants tell buyers about Shopify products, and fix the data behind it. Co-founder @GEOAngelsAI

[Ulysse HC @UlysseHC31]

I ask ChatGPT, Perplexity and Gemini about Shopify products all day, then find out where their wrong answers come from.

Usually it's the store itself.

Sharing what I find here. Building @GEOAngelsAI with @PascalZeMarshal

[Ulysse HC @UlysseHC31]

@VisionCourse Eligibility gets products into the Catalog. What agents then say about them is a separate check: ChatGPT often recommends the right product and quotes a spec that contradicts the store's own page.

[Ulysse HC @UlysseHC31]

@harleyf In a shared cart, shoppers compare brands side by side, so the product data does more of the selling. On one store we audited, the Shop app showed a different review count than the product page.

[Ulysse HC @UlysseHC31]

@kurtinc Which count will it show, though? On that store the review app, the Shop app and the product page each showed a different number for the same product.

[Ulysse HC @UlysseHC31]

@mcfazeli @Shopify @Muse @instinct Will do. I'll pull the Agentic report again after Black Friday and post what the Meta AI and Muse line shows.

[Ulysse HC @UlysseHC31]

@trevorgrogers @Shopify Recommendations from the storefront agent will lean on product data. On one store we audited, a shared template had copied the same specs onto several different products. Curious whether the agent reads metafields or only the standard category fields.

[Ulysse HC @UlysseHC31]

@tobi The merchant side of https://t.co/kKqcvWDCxg: whatever an agent pulls from Shopify Catalog, it repeats to the shopper. On one store we audited, the standard category attributes were empty on every product. Worth checking before more agents plug in.

[Ulysse HC @UlysseHC31]

Black Friday is a month away. Shoppers will ask ChatGPT for the best gift under $100 and get an answer built from the prices and stock it last read.

Ask it about your best seller today and check the price it quotes.

[Ulysse HC @UlysseHC31]

@glenngabe Matches what we see on a single Shopify store: ChatGPT went from naming it in 0 to 1 of 4 shopper questions, which alone could be noise. So we read mentions as a trend and check the facts it quotes when it does name the product.

[Ulysse HC @UlysseHC31]

@samvanderlee While you're in Review product data mapping, check the category attributes too. On one store we audited they were empty on every product, and Shopify Catalog categorizes from those fields.

[Ulysse HC @UlysseHC31]

@Kevin_Indig The product searches are the part to watch. ChatGPT usually names the right product, then quotes a spec that contradicts the store selling it. In 97% of the products we've checked, at least one fact was off.

[Ulysse HC @UlysseHC31]

@mrfeldmanSEM Worth running the same check on Shopify's category attributes. Shopify Catalog, which feeds ChatGPT and Copilot, categorizes products from those standard fields, and on one store we audited they were empty on every product.

[Ulysse HC @UlysseHC31]

@parker_brydon @Shopify The merchant side of this: on one store we audited, the same products were published as several listings sharing one template, and reviews were split across them. Curious how clustering decides which listing's specs win.

[Ulysse HC @UlysseHC31]

Walmart now chooses which shopping agents get into its store, and on what terms.

On Shopify that choice is already made: stores are agent-ready by default, and the Catalog goes out to ChatGPT, Copilot and Meta AI.

Worth knowing what it says about your products.

[Ulysse HC @UlysseHC31]

Meta says shopping is the top use case for Muse, its new agent.

On a Shopify store we track, the Meta AI and Muse channel brought in $0 over 30 days. ChatGPT and Shop each brought five figures, from the same catalog.

Your numbers are in Sales channels > Agentic.

[Pascal Le Marechal @PascalZeMarshal]

How are we ranking Novaalab 1st in our niche for AI visibility in our niche? (data from TripleWhale and Ahrefs).
=> 3 months of GEO optimization, following our 10-sections GEO Angels strategy.
Dozens of hours of work to make our store from 34% to 93% GEO and AI crawlers optimized.
The difference has been MASSIVE. AI crawlers understand better our products, specs, and they quote us much more often.

Is YOUR store AI-optimized?
Find out with our free check tool on https://t.co/q3XLcnHwCb
Need help getting your store AI-ready? Book a quick call on https://t.co/uUXlk6Yv0v

[Ulysse HC @UlysseHC31]

Same product, three listings, reviews split between them.

To an assistant, each listing looks less proven than a rival with all its reviews on one page. Search your admin for products that share a title.

[Ulysse HC @UlysseHC31]

@AnthonyEclipse On a store we audited, the specs sat in the descriptions and the standard category fields Shopify Catalog uses were empty on every product. Worth checking what the AI put in those fields.

[Ulysse HC @UlysseHC31]

@lilyraynyc Seen on the product side too: ChatGPT recommends the right product, then quotes specs it picked up from a review site or an old FAQ. By the time the shopper searches the brand, they're checking the wrong facts.