Recommendation Authority, measured in outcomes.
Illustrative diagnostic examples showing how Salienci measurement is reported. We do not publish claims we cannot substantiate.
Aggregate movement across diagnostic engagements over a 90-day period.
Net new AI-surface citations captured per week at steady state.
These are illustrative diagnostic examples of how Salienci reports movement, not a single client result and not a forecast. Salienci does not publish invented case studies.
Verified engagements coming soon.
We only publish named engagements once outcomes are verified and the client has approved disclosure. Until then, the platform benchmark above is the honest picture.
How to read these results
Read every figure on this page as an illustrative diagnostic example of how Salienci reports movement, not as a promise or a customer outcome. Salienci publishes three distinct kinds of number, and they are never mixed.
Platform benchmark
Aggregated movement across Salienci diagnostics. It describes a dataset, never one client, and its sample provenance is documented before publication.
Illustrative example
Interface values used to show what a report looks like. They are labelled wherever they appear and are not observed client results.
Client-specific outcome
A measured result for one named engagement. Salienci publishes these only when the outcome is verified and the client has approved disclosure.
How this is measured: Salienci Score methodology · Data integrity commitments
Why attribution has limits
AI answer engines are non-deterministic and rarely expose a referrer, so no measurement can claim complete credit for a lead. Salienci reports every attribution claim with a confidence tier and reports insufficient evidence as insufficient rather than estimating it.
- The same question asked twice can return different brands, so a single observation is never treated as a rank.
- Many AI surfaces pass no referrer, so downstream conversions can only be attributed at a stated confidence level.
- Score movement and commercial outcome are related but separate measurements; neither is presented as proof of the other.
- Comparisons are only valid where both sides were measured under the same methodology version, engine set and coverage.
Continue reading
These pages give the next layer of detail behind what you have just read.
Published · Last reviewed