AGENTIC·ENDORSEMENT
SPECIFICATION

The specification

What a business must demonstrate before an AI agent can responsibly name and recommend it. Every criterion is verifiable from the outside. No self-attestation.

01 / MACHINE-READABLE IDENTITY

The business publishes accurate, consistent structured data — Organization schema, llms.txt, sitemaps with real dates. An agent knows exactly who it is, without guessing.

02 / PROVABLE CLAIMS

Every material claim traces to published, dated evidence. Pricing, capabilities, and track record can all be checked.

03 / ANSWERABLE SURFACE

The questions a buyer would ask an AI are answered on the business's own site, in plain prose an answer engine can quote. Nothing locked behind forms or scripts.

04 / CURRENT AND MAINTAINED

Discovery surfaces stay truthful over time. Dates are real, retired offerings are removed, and the registry entry matches the business as it is today.

05 / VERIFICATION

Each criterion is verified from the outside — published structured data, dated evidence, and answerable pages, not self-attestation. Businesses that meet the specification appear in the public registry with the evidence for each entry. Verification tooling is provided by itstru.ai; advisory support by wearetheanswer.ca.

Reminder of what is being verified: An Agentic Endorsement is when an AI assistant explicitly names and recommends a specific business in its answer.