- 66 public ChatGPT and Gemini shopping conversations.
- Conversation link to full Markdown to compact evidence trail.
- Temperature-zero extraction for brand, merchant, and routing outcome.
- Deterministic aggregation into winner, leak, loser, and site-health layers.
Back to research
AI shopping visibility: the buy path problem.
Genesis 03 studies how ChatGPT and Gemini route shopper demand after recommending ecommerce brands.
A recommendation is only commercially useful if it can turn into a clean route to buy. This field study tracks which recommendations become official product-page wins, which leak to third-party merchants, and which end without a usable purchase path.
What the study measures.
The study treats AI recommendation as a routing system. Once a model names a brand, the question becomes whether the shopper is sent somewhere that the brand owns, somewhere it does not, or nowhere commercially useful.
Official wins send demand to the brand's own page. Third-party leaks route the shopper to retailers or marketplaces. No-PDP losses recommend a brand without giving the shopper a usable product-detail path.
36.8%
Official winners
17.6%
Third-party leaks
45.5%
No-PDP losers
AI assistants can recommend a brand and still fail to create a buy path.
That makes visibility a two-step problem: getting named by the model, then making sure the system can route the shopper to the right commercial surface.
Genesis 03 also scans canonical brand domains as a separate diagnostic layer. The paper finds site health is weakly correlated with AI routing, so it should be treated as a related but distinct signal.
257
Domains scanned
231
Accessible sites
41/100
Median health score
The full PDF includes the 66-link source matrix so each result can be traced back to a public AI conversation and the extraction evidence used for scoring.
Brands are already being recommended inside AI shopping experiences. Genesis 03 shows that the next question is whether those recommendations send demand to the right place.