Evidence-led notes for independent Shopify founders and ecommerce operators investigating how one product is found, shortlisted, compared and ordered in AI shopping answers.
FEATURED RESEARCH · 03 SEP 2026
The latest product-level observation.
Each note turns captured evidence into a clearer view of retrieval, inclusion, comparison and ordering.
Known-catalog search and external AI product discovery are different evidence surfaces. Anthropic's reference architecture helps make the boundary—and the merchant testing problem—concrete.
A product can be found, understood and fairly compared without appearing first. Separating retrieval, inclusion, comparison and ordering reveals whether there is a useful problem to fix.