Skip to content
bikma docs
Documentation index

Which tools measure AI share of voice?

Metric

AI Share of Voice is the brand's share of presence in generative answers relative to the competitive set; in Bikma it is measured as frequency, position in the narrative and weight, on the same prompts.

How it is calculated in Bikma

Bikma runs the same prompt set against the plan's models every day and records, answer by answer, whether and how the brand appears next to its competitors. Signals are weighted with the Prompt Probability Index, so the questions that count more carry more, and normalised to make brands comparable. The value reads as a view over time and as a brand-by-cluster matrix.

Mentions and citations kept apart

In the calculation the mention (the brand named) and the citation (the source drawn on) stay distinct: they are two different signals and must be read separately, so that awareness is not confused with influence over sources.

How to read it

The trend over time counts for more than the absolute value: a good benchmark is relative to direct competitors on the same prompt set, not a universal number.