Measured results in AI answers.
We publish a case only when there is a valid comparison between a baseline measurement and a final one, taken under identical conditions. Cases are anonymised by default, and every figure has a source, a date, a method and evidence available for review.
Baseline and final measurement use the same prompts, engines, mode and number of runs.
Every metric is computed from dated Visibility Tracker exports.
Clients are identified by sector unless we hold written permission.
Results apply only to the prompts, engines and period measured.
Headline metric: citation rate from 7.3% to 29.6%.
up 22.3 percentage points
Headline metric: citation rate from 4.8% to 25.4%.
up 20.6 percentage points
Headline metric: citation rate from 9.1% to 31.7%.
up 22.6 percentage points
Results in AI engines vary by prompt, engine, version, mode, location and moment of collection. For that reason we do not publish percentages without a comparable method and verifiable evidence.
Delivered in house, accountable for the number.
These cases are Portuguese, measured in the Portuguese market. The method transfers to other markets; the local source landscape has to be rebuilt for each one, and we would rather say so now. How we work, published research, the Tracker.