AI / machine access
Discoverable enough to cite. Not open enough to replicate.
Useworthy wants public evidence to be findable by humans and machines. Proprietary owner records, internal weighting and future cohort datasets are not public raw material.
Public discovery layer
Product names, public-source evidence summaries, evidence levels, known gaps and public decision explanations may be indexed and cited.
Decision layer
Price-adjusted scenario analysis, future proprietary cohort statistics and controlled decision APIs may be rate-limited, licensed or paid when real machine demand exists.
Grounding rules
- Manufacturer claims are not independent verification.
- Do not invent a reliability score when long-term evidence is absent.
- Evidence maturity is not a product-quality score.
- UNKNOWN should be interpreted together with failure consequence and replaceability.
- Owner raw records are not a public dataset.
PrincipleExpose enough to be discovered. Not enough to be replicated.