Surfaces
Every assistant reads differently — and keeps changing.
We measure each surface on its own terms, through developer APIs, and track it over time.
OpenAI
Answers are grounded in live web search. It rewards pages it can actually read: allowed crawlers, complete structured data, and identifiers it can confirm. Vague titles get skipped for a competitor with a cleaner listing.
Grounds answers in search and product feeds. Category and attribute discipline decide who surfaces — the same signals that have always powered Shopping, now read by the assistant.
Microsoft
Rides emerging agentic-commerce rails rather than a public query API today. We track readiness — feeds, identifiers, structured data — so you're prepared as the protocol lands.
Per-surface deep dives are on the way. Meanwhile, the audit shows exactly how each one reads your store today.