"Is 8% good?" is the most common question we get about AI visibility, and it has no answer without a distribution to compare against. So here is one, measured across 608 answers and 66 brands in a single competitive category.
The distribution
| Percentile | Share of answers naming the brand |
|---|---|
| Top brand | 24.2% |
| 5th | 10.2% |
| 10th | 5.8% |
| 20th | 3.6% |
| 30th | 1.8% |
| 50th | 0.8% |
| Bottom quartile | under 0.5% |
Sixty-six brands, 608 answers, five engines, three months.
How to read your own number against it
Above 15%: you are one of the handful of names an assistant reaches for in this category. Three brands in sixty-six are here.
5% to 15%: you are in the answer set and appear regularly. Realistically a strong position, and about eight brands occupy it.
2% to 5%: you appear sometimes. A buyer running several queries might see you once. This is where most brands that are doing anything at all sit.
Under 2%: you are effectively absent from the buying conversation. Half the measured brands are here, which is worth knowing before treating it as a personal failure.
Zero: you are not in the answer set at all, and repetition will not surface you. Six of the 66 brands never appeared once.
The distribution is more concentrated than search
The top brand takes roughly a quarter of all answers. The tenth takes under 6%. The thirtieth takes under 2%.
On a search results page, ranking ninth still gets you seen. In an answer, ninth means being named in 6% of answers, and the person asking usually stops reading after three names. There is no page two to be on.
That concentration is the most important structural fact about AI search and it is the opposite of what "the field is new so it is wide open" implies.
Why a category benchmark beats a universal one
These numbers are from one category, marketing and AI search agencies, which happens to be unusually contested. In a quieter category the top brand might take 40% and the tenth might take 15%, because there are fewer names competing for the same three slots.
So the useful benchmark is not this table. It is the same table built for your category, which takes the same method: fix a prompt set, run it across engines over weeks, count who gets named.
What generalises is the shape. A steep head, a long thin tail, and a large fraction of brands at effectively zero.
The number that matters more than your share
Your ratio of cited to named.
A brand read six times more often than it is recommended has a different diagnosis from one with a low share because nobody reads it at all, and the two point at completely different work.
A brand at 3% that is barely cited has a content and retrieval problem. A brand at 3% that is cited constantly has a third-party presence problem. The share alone does not distinguish them.
Why we publish the method with it
Every number on this page is reproducible. The prompt count, the period, the engine list and the counting rule are all stated, which means anyone can disagree with the result specifically rather than generally.
That matters more than usual in this category, because most published visibility figures come with no method attached at all, and a benchmark you cannot reproduce is not a benchmark. It is a claim.
If you think a figure here is wrong, run the same prompts under the same conditions and tell us what you get.
Method
39 buyer prompts, five engines, 608 answers on a rolling schedule between 24 May and 24 August 2026. A mention is the brand named in the answer body, once per answer. Treat a gap of about two percentage points between adjacent brands as noise rather than a difference.
What to do once you know your number
A benchmark is only useful if it changes what you do next. Here is the decision it should drive.
Below 1%. You are in the long tail, which is where 26 of the 66 brands in this dataset sit. The first job is not content. It is finding out whether you are cited but not named, because that decides everything else and takes twenty minutes to check.
Between 1% and 5%. You are around the median and you are being named occasionally, which means the models know you exist. This is the band where third-party presence has the highest marginal return, because the recognition is already there and it needs reinforcing rather than creating.
Between 5% and 10%. You are ahead of most of the field and the remaining gap to the leaders is large. Getting past this usually means becoming a source other people cite rather than a business other people list.
Above 10%. Five brands out of 66 are here. At this point the work changes from getting named to staying named, which means keeping the sources that mention you current and accurate.
The one thing not to do at any level is compare yourself to a universal benchmark. Category concentration varies enormously, and a 3% share in a tight category can be a stronger position than 8% in a fragmented one.
The measurement behind the benchmark
The figures above are why the research on this page exists. We run this measurement continuously across every account rather than sampling it, which is the only way to reach a number like 82,000 measured answers and the only way to see a change in the week it happens rather than the month after.
A worked example of using the benchmark
Two businesses with the same visibility figure, and two completely different situations.
Business A: 2.1% share, cited on eleven of thirty questions. They are being read a great deal and named almost never. Their content is not the problem; it is demonstrably good enough that models build answers from it. The remaining work is external, slow, and the highest-value thing available to them.
Business B: 2.1% share, cited on two of thirty questions. They are barely being retrieved at all. The fix starts on their own pages and their own listings, and it is the faster of the two situations to move.
Same number, opposite prescriptions. This is why a single visibility figure is close to useless as a starting point and why the two-column split matters more than the benchmark itself.
What the benchmark is actually for. Deciding how much room there is. In a category where five brands hold 40% of namings and half the field sits below 1.5%, a move from 2% to 6% is realistic within a year and would put a business ahead of most of the field. In a flatter category the same move would be worth much less.
Read the benchmark for headroom. Read the two columns for direction.
See which questions name you
A free AI visibility audit on your own buying questions, across all five answer engines.
Book your audit
Ashur Homa
Built and scaled a digital brand to $100M+ in sales with zero ad spend. Has helped businesses generate millions through AI go-to-market strategy. Leads growth at Omni Eclipse.
Connect on LinkedIn
