Fill Shopify's category attributes from the data you already have
Every Shopify product has a category, and every category has attributes with standard values: a dress's silhouette, a kettle's capacity. In most stores the category is set and the attributes are blank.
Why the blanks cost you
The category attributes are what Shopify passes to Google, to Shopify Catalog (which ChatGPT reads) and to any channel built on its taxonomy. A product whose attributes are filled travels with its facts. A product whose facts live only in its description travels without them, and an assistant asked for a relaxed-fit linen blazer in sand has nothing to match. Shopify says the same fields make products easier to find on your own site, and you can filter its analytics and reports by them.
Why they stay empty
Filling them by hand means opening every product, reading its description, and picking the taxonomy's value for each attribute, which for a dress runs to a dozen fields. A thousand products is weeks of work, and the values have to be the taxonomy's own (Sheath, not column) or the field is as empty as before.
What Sighted does
Sighted reads the product's own description, options and photos, writes the attributes the category defines, and maps each one to the closest standard value. A column dress becomes Sheath. A real fact with no close value becomes Other. A fact Shopify has no field for goes into a Sighted metafield rather than being dropped. Nothing is invented: a value the product's data does not support is held for your review instead of written. On approval the values go into the product's taxonomy fields in your admin, and a changed product is done again when Shopify's webhook reports it.
See how the assistants read your store.