Share of Voice (AI-SOV)
The percentage of AI responses in a given category that recommend a specific brand compared to its competitors.
What is Share of Voice?
AI share of voice adapts a familiar media metric to answer engines. Take a defined set of prompts for a category, run them across one or more AI engines, count every brand mention in the answers, and express each brand's mentions as a percentage of the total. A brand with 40 percent AI-SOV in "email marketing platforms" is named in answers 40 percent as often as all tracked brands combined for that prompt set.
The number is only meaningful with a stated competitor set and prompt set, and it is usually broken down further: by engine, since Perplexity and ChatGPT often favour different brands; by topic cluster, since a brand can dominate one sub-category and be absent from another; by intent, since being named in "what is" answers is different from being named in "which should I buy" answers; and over time, so that the effect of content, PR or model updates can be seen.
Many teams weight the raw count. A first-position mention in a ranked list, a mention with a citation to the brand's own site, or a mention with positive sentiment can each be scored higher than a passing reference, producing a weighted share of voice that better reflects influence on the buyer.
Why it matters for AI search
Share of voice is the metric executives recognise. It converts hundreds of individual AI answers into a single competitive number that can be reported alongside search and paid media share, and it makes the trade-off visible: if a competitor is gaining AI-SOV, the brand is losing it. Tracking it per engine and per topic shows exactly where the competitive gap is, which is where GEO effort should go first.
Related terms
Mention Frequency
A metric tracking how often a brand is mentioned in AI responses relative to a total set of industry-relevant prompts.
Citation Intelligence
The analysis of which specific sources and pages an AI model chooses to reference when answering a query.
Prompt Intelligence
The study of which user prompts most frequently trigger a brand recommendation or citation in AI responses.
AEO (Answer Engine Optimization)
The practice of optimizing content to be cited and recommended by AI-driven search engines and conversational assistants like ChatGPT and Perplexity.
Frequently asked questions
How is AI share of voice calculated?+
Brand mentions divided by total mentions of all tracked brands across the same prompt set, engine and period. Weighted versions give extra credit for first position, owned-page citations or positive sentiment.
Why does my AI-SOV differ between ChatGPT and Perplexity?+
Each engine uses a different model, different retrieval sources and different citation habits. Perplexity leans heavily on live web retrieval; ChatGPT blends training knowledge with search results. The same prompt can therefore produce different brand lists.
How often should AI-SOV be measured?+
Frequently enough to catch changes from content updates and model releases, and with repeated runs per prompt so that natural output variance does not masquerade as a trend.