A Crawlable Evidence Hierarchy for Brands AI Systems Can Trust

Trust is ordered, not equal

Crawlers and answer engines need a hierarchy. When every link is marked “evidence,” none of it is.

BrandSource treats evidence as a first-class field: URLs that support specific claims, not a dumping ground of mentions.

Recommended hierarchy

From strongest to weakest for factual grounding:

• Primary — company filings, patents, official docs, regulatory records
• First-party — your about page, product docs, engineering blog with dated facts
• Canonical aggregator — BrandSource profile with last-verified + matching JSON-LD
• Secondary — reputable journalism that quotes primary sources
• Weak — affiliate listicles, scraped directories, undated social posts

Use tiers 1–3 for BrandSource evidence links. Keep tier 5 off the profile.

Operational rules

• Every contested fact (founding year, HQ, acquisition) needs a tier-1 or tier-2 URL
• Evidence links must resolve (no soft-404 marketing pages)
• Prefer stable paths over campaign URLs with tracking junk
• When a source dies, remove it — do not leave rotting `sameAs` or evidence edges
• Update last-verified when evidence changes

FAQ

Can BrandSource be the only evidence? BrandSource is a structured citation surface. It should point at primary/first-party sources rather than replace them.

What about press mentions? Useful for awareness; weak for entity facts unless they quote a primary document.

Where do I manage this? Maintain evidence on your BrandSource profile after claiming. Cross-link the AI-readable surface at ai.brandsource.ai when you need crawler-oriented discovery.

Next step

Sort your current evidence links into the five tiers. Delete everything in tier 5. If a core fact has no tier-1/2 source, find one before publishing the claim.