Ask an assistant which agency can get your business recommended by AI and it will name someone. Ask a hundred times and it names a fairly consistent set of someones. We wanted to know who.
Over three months we ran 39 buyer questions against five answer engines and collected 608 answers, then counted how often each of 66 agencies and tools was named.
The top of the table
| # | Brand | Named in | Share |
|---|---|---|---|
| 1 | First Page Sage | 147 of 608 | 24.2% |
| 2 | iPullRank | 120 | 19.7% |
| 3 | NoGood | 115 | 18.9% |
| 4 | Siege Media | 80 | 13.2% |
| 5 | Omniscient Digital | 62 | 10.2% |
| 6 | Optimist | 58 | 9.5% |
| 7 | Profound | 43 | 7.1% |
| 8 | WebFX | 39 | 6.4% |
| 9 | AEO Engine | 38 | 6.3% |
| 10 | Prime Avenue Group | 35 | 5.8% |
The full 66 and the raw data are published, linked at the end.
The distribution is more concentrated than Google
First Page Sage appears in roughly one answer in four. The brand ranked twentieth appears in one in twenty-eight. Below about rank thirty, brands appear in under 2% of answers, which means a buyer running a handful of queries never sees them.
This is not how a search results page distributes attention. A page ranked ninth on Google still gets seen. A brand ranked ninth here is named in 6% of answers, and the person asking usually stops reading after three names.
If you have been told AI search is a level playing field because it is new, this data does not support it. It is more concentrated than the results page it is displacing.
Being read is not being recommended
The most interesting number in the dataset is not in the ranking.
The pattern repeats across the whole dataset: brands the models read often, and then do not name. Being read more than a rival is no guarantee of being recommended more than them.
First Page Sage is close to the reverse: cited 59 times, named 147. They are being named inside other people's articles rather than their own.
That gap is the whole game. Being a good source gets you read. Being named in someone else's source gets you recommended. Most agencies, ours included, have been optimizing the first one.
This is our own list, so weigh it accordingly. Run the prompts yourself in a logged-out session and compare.
The most-cited source is not an agency
We also counted which domains the models pulled from.
| Domain | Answers citing it |
|---|---|
| reddit.com | 225 |
| aeoagency.org | 148 |
| youtube.com | 145 |
| linkedin.com | 134 |
| aeoengine.ai | 130 |
Reddit is cited 52% more often than the next domain and more often than every agency website in the category. YouTube and LinkedIn both beat every vendor site too.
The implication is awkward for anyone selling content strategy: a well-argued Reddit comment in the right subreddit has a better chance of ending up inside an AI answer than a well-optimized page on your own domain.
Method
39 prompts, written the way buyers ask rather than the way keywords are written, fixed for the whole period.
Five engines: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Gemini.
608 answers, collected on a rolling schedule between 24 May and 24 August 2026 rather than in one sitting, because these systems answer differently day to day.
A mention is the brand named in the answer body, counted once per answer so a verbose answer cannot inflate a score. Aliases matched. A domain in a citation footnote counts as a source, not a mention.
Where the method is weak
Answer engines vary run to run, so treat a two-point gap between adjacent brands as noise.
The 66 brands are a large sample of the category, not a census. Absence means we did not track that brand.
Prompt selection is a judgment. A different reasonable set of questions produces somewhat different numbers, which is why the prompt list is published.
And it measures being named, which is not the same as being good. No index can tell you which agency does better work.
What to do with it
If you are in the table you already know whether you are happy.
If you are not, the useful question is not how to rank higher, because there is no ranking to climb. It is which sources the model reads when it answers your buyers, and whether you are in them. For this category that was Reddit, YouTube, LinkedIn and a handful of agency blogs, in that order. For yours it is a different list, and it is knowable.
What a buyer should take from this
Three things, and none of them is "hire whoever is at the top".
Concentration means the list is shorter than it looks. Sixty-six brands appear, five of them hold 40% of all namings, and half the field sits below 1.5%. When an assistant gives you three names, it is drawing from a much smaller pool than the number of agencies that exist. Being named at all is the achievement; the order within the top is noise.
Being named is not being good. This measures which brands the models have learned to mention, which is a function of how much has been written about them and where. It is not a quality ranking and no measurement of this kind can be. Use it to build a shortlist, then evaluate the shortlist properly.
The mechanism is visible in the source data. The brands at the top are named inside other people's articles, not their own. That is the finding worth acting on whatever category you are in, and it is the reason a listicle you are absent from costs you more than a page you have not written.
What the same measurement does for a client
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.
What the source data shows about how they got there
The rankings are the least interesting part of this dataset. The citations are the interesting part.
We traced every one of the 608 answers back to what was cited to produce it. The brands named most often are not the brands with the best websites. They are the brands that appear inside other people's content.
The most-cited source in the entire category is a community site, ahead of every agency's own domain, including the domains of the agencies that get named most.
That is the mechanism, stated plainly: getting named is largely a function of how often you appear in sources other people control. A brand can publish continuously on its own domain and stay invisible in the answer, and this dataset contains several that do.
Why this is good news rather than bad. Publishing on your own site is competitive, slow and expensive. Being present in the right third-party sources is unglamorous, cheap, and something almost nobody in most categories has bothered to do properly.
What to take from the table. Not who to hire. Which behavior correlates with being named, which is: be somewhere other than your own website.
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
