AI SRE Watchlist
  • Tools
  • Observability
  • Resources
  • Updates
Search Watchlist

Trust model

Evidence before confidence

Profiles are research aids, not rankings. Every material claim should expose its source, evidence state, and review date.

Documented
Supported by a linked primary source and checked by the Watchlist.
Vendor-claimed
The company explicitly states this, but the Watchlist has not independently reproduced it.
Observed
Recorded during a disclosed hands-on test with scope and date.
Unknown
No adequate reviewed source is attached. Absence is never converted into a negative claim.
Source priority
Sources are not interchangeable.
  1. Official documentation, repositories, security pages, and release notes.
  2. First-party announcements, with promotional language labeled as a vendor claim.
  3. Approved practitioner interviews and disclosed Watchlist testing.
  4. Independent reporting used only when a primary source cannot answer the question.
What the Watchlist does not infer
Pricing, integrations, security posture, product availability, customer outcomes, and “verified” status remain unknown unless a reviewed source supports them. A company profile mapping is identity work, not endorsement.
Corrections and conflicts

Corrections require a source and enter a private review queue. Company submissions never publish directly.

Conflicting sources remain visible as a conflict until the discrepancy is resolved. Review dates change only after a real review.

No paid rankingsNo inferred verificationNo hidden public evidence

AI SRE Watchlist

Evidence-led research and private evaluation workflows for reliability teams.

MethodologyEditorial policySubmit a correctionPrivacyTerms