Category Benchmarks: Average AI Share of Voice by Industry
Within any category, a handful of domains absorb most AI citations and a long tail gets almost none. Knowing your category's concentration tells you how hard the climb really is.
AI share of voice is the percentage of citations within a defined set of category queries that a single domain captures. The defining finding across MentionRadar’s index is concentration: in most categories a small group of domains absorbs the majority of citations, while a long tail of sites is cited rarely or never — the same power-law shape as organic search. That makes a universal “good” number meaningless; the realistic benchmark is relative. In concentrated categories (mature SaaS niches, regulated industries) a few incumbents dominate and even single-digit share is a strong position. In fragmented categories (local services, emerging tools) citations spread across many domains, so the leader’s share is lower and entry is easier. Treat any specific percentage as a directional estimate from a sample. The actionable read: measure your own share against your category’s concentration, not an absolute target.
What is AI share of voice?
AI share of voice (SOV) is the share of citations a domain wins across a defined set of queries that represent a category. If a category is described by 100 representative buyer questions and your domain is cited in the answers to 12 of them, your SOV is loosely 12% on a query-coverage basis. (You can also weight by number of models or by how prominently you’re named.) It is the answer-layer analogue of organic share of voice — and like that metric, it only means something relative to the category around it.
How is the benchmark calculated?
From the index, we take a set of queries that define a category, collect the cited domains across ChatGPT, Gemini and Grok, and compute how citations distribute across domains. Two things fall out: the leaders’ combined share (how concentrated the category is) and the shape of the tail (how many domains get cited at all). Both are directional — sensitive to which queries you choose to represent the category and how you canonicalise domains — so we publish them as reference ranges, not precise constants. When we publish reference ranges, each will be sampled, dated and sourced per our methodology — we will not post a benchmark number we cannot stand behind.
Why benchmarks differ so much by industry
Concentrated categories
Mature software niches, regulated industries, and topics with a few authoritative sources tend to be highly concentrated. A handful of well-known domains and review platforms soak up most citations. Here, getting cited at all already beats most competitors, and displacing an incumbent is a multi-query campaign — expect a slower climb.
Fragmented categories
Local services, newer tool categories, and broad consumer topics tend to be fragmented: citations spread thinly across many domains, no single site dominates, and the leader’s share is modest. These are easier to enter — a well-structured, extractable page can win queries quickly — but also easier to lose, because the field is crowded and volatile.
Why concentration matters more than the raw number
A 5% share means very different things in a concentrated versus a fragmented category. In a winner-take-most niche, 5% might put you in the top handful of cited domains; in a fragmented one it might be middle-of-the-pack. So the benchmark to chase is not a fixed percentage — it’s your position relative to the category’s leaders and tail. Always read SOV next to concentration.
How to lift your category share of voice
- Target the queries you’re missing. Find the answers where a competitor is cited and you aren’t — your AI keyword gap.
- Publish citable formats. Comparisons, original data and clean docs win disproportionately — see which content types get cited most.
- Earn review-platform presence for evaluative queries — see the review-platform effect.
- Cover all three models. Concentration differs by model, so a query lost in one may be winnable in another.
Measure your own share of voice
To benchmark yourself, you first need your real citation list. The free Domain Check returns the queries the three models already cite your domain on; run it on the category leaders too and you can estimate concentration and your position in it directly. For the full set of structural findings, return to the State of AI Citations hub.
Frequently asked questions
What is category AI share of voice?
It is the proportion of a category’s AI citations that a single domain captures. In practice it is concentrated: a handful of domains absorb most citations while a long tail is rarely cited.
Do you publish exact benchmark numbers?
Not yet. We describe concentration qualitatively and will only publish reference ranges once they are sampled, dated and sourced per our methodology. We do not invent percentages.
Is being cited at all worth anything in a concentrated category?
Yes. Where citations concentrate on a few domains, being cited at all often puts you ahead of most competitors. Displacing an incumbent, though, is a multi-query campaign rather than a single content fix.
How do I find my own share of voice?
Start with the free Domain Check to see which queries already cite you, then run competitors in the same category to gauge concentration first-hand.