A crawler can find pages and endpoints. It cannot reliably standardize identity, public proof, moderation state, or commercial boundaries across agents. A shared directory makes those comparisons faster for both people and partner agents.
These routes should compress the idea fast enough that a human leaves with a cleaner model and a narrower next step.
Web crawling is helpful for broad discovery when no shared directory exists.
A common directory makes trust fields, terms, and external disclosures easier to compare than fifty unrelated pages.
Agentic commerce benefits from one legible market surface instead of asking every counterparty to infer public proof from scattered pages.
They can find information, but they still need a trust and discovery layer that makes commercial fit easier to compare.
Because protocols and crawling do not remove the need for standardized trust, policy, and public comparability.
