Measuring share of voice in AI answers comes down to a repeatable test: ask the assistants a fixed set of category questions, then count how often your brand appears in the answers relative to your competitors. That proportion, tracked over time and broken down by assistant, is your share of voice. It is the AI-era version of the old media metric, adapted to a place where the result is a generated answer rather than a ranked list, and it tells you how present your brand is at the point where buyers increasingly start their decisions. The detail that trips people up is that the answer differs sharply by assistant, so a single blended number hides more than it reveals.
What AI share of voice actually measures
AI share of voice answers a simple question: when a buyer asks an assistant about your category, how often does your brand come up, and how does that compare with competitors? It is the AI-era equivalent of share of voice in traditional media, adapted to a medium where the answer is generated rather than listed. It is distinct from related metrics such as mention rate, which is how often you appear in answers to category questions in absolute terms, and citation rate, which is whether a model attributes a specific claim to your content. Share of voice is explicitly comparative: your mentions as a proportion of all brand mentions for a query set.
How to calculate share of voice
The most common method is citation or mention based. Define your prompt set, run it, and for each answer count the brand mentions. Your citation-based share of voice is your brand's mentions divided by the total brand mentions across the answer set, expressed as a percentage. A more granular variant is word-count based, where you measure the share of an answer's words devoted to your brand, which captures prominence rather than just presence. Both are valid; mention-based is simpler and word-count based is more sensitive to how much attention each brand receives within an answer.
Why you must measure each assistant separately
The same brand can have very different visibility across assistants, because each one draws on different sources and cites differently. One assistant may lean on encyclopaedic and major-news sources and cite only a few times per answer, while another may pull from community forums and review sites and cite many sources per answer. As a result, a brand can show a strong share of voice on one assistant and a weak one on another for the identical query. Averaging across assistants without looking at each hides exactly the information you need to act on, so measure each, then compare.
A practical measurement process
Begin by building a representative prompt set that reflects how real buyers ask about your category, covering awareness, consideration, and decision-stage questions. Run the set across each assistant on a consistent schedule, because answers change over time and a single snapshot is unreliable. For each answer, record which brands are mentioned, which appears first, which sources are cited, and whether the mention of your brand is positive, neutral, or negative. Aggregate into share of voice per assistant and overall, and track the trend rather than a single reading. A number of independent tools automate this by running prompt sets continuously and reporting mention trends, share of voice, and sentiment, which lets you validate any campaign with third-party data.
Turning measurement into action
Measurement is only useful if it changes what you do. Low share of voice in high-intent decision-stage prompts is the most commercially urgent gap, because those are the questions closest to a purchase. Address visibility gaps through the two routes that put a brand into AI answers: organic generative engine optimisation, which earns durable citations over time, and paid content-layer placement, which puts your brand into the publisher content assistants read at the moment a relevant question is asked. blankspace supports the paid route and lets advertisers track presence in AI responses before and during a campaign, validated by any third-party measurement tool, so share of voice becomes a metric you can move rather than only observe. For publishers, the same data viewed from the other side shows which brands their content is already surfacing in AI answers, and what that audience is worth to the advertisers trying to reach it.
Frequently asked questions
What is a good AI share of voice?
There is no universal benchmark, because it depends on your category and competitors. What matters is your share relative to direct competitors and the trend over time, especially in high-intent decision-stage prompts. Rising share against competitors is the signal to watch.
How often should I measure share of voice?
On a regular, consistent schedule rather than as a one-off, because AI answers change frequently. Continuous or weekly measurement reveals trends and the impact of campaigns, while a single snapshot can be misleading.
Which tools measure AI share of voice?
Several independent platforms run fixed prompt sets across ChatGPT, Perplexity, Gemini, and others and report mention rate, share of voice, and sentiment. Using a third-party tool lets you validate results independently of any single vendor.
Why does my share of voice differ between ChatGPT and Perplexity?
Because each assistant draws on different sources and cites differently, both in which sources it trusts and how many it cites per answer. This is why each assistant must be measured separately rather than averaged.
How do I improve my share of voice in AI answers?
Through organic generative engine optimisation to earn durable citations, and paid content-layer placement to put your brand into the publisher content assistants read at the moment of a relevant query. Target the high-intent prompts where visibility matters most.