Which AI visibility tool is mentioned most by AI engines?

In our 2026-08-09 measurement, Profound was named in 41 of 90 answers, ahead of Otterly (38) and SEMrush (38). Six days earlier the order was different: Otterly led with 42 and Profound had 37. That is a measure of how often each brand appears in the sources the engines draw on, not a quality ranking. A tool can be excellent and rarely named, and a tool can be named constantly because it has been listed in a lot of roundups.

Does being cited more often mean a tool is better?

No, and it is worth being blunt about that. Citation frequency measures presence in the sources an engine reads, which is mostly a function of how many third-party roundups and comparison pages list you. It rewards marketing reach and age. It says nothing about accuracy, engine coverage or price. Read it as "who the engines have heard of", not "who is best".

Why is Cituna not in its own list?

Because we barely register in it, and inflating our own row would make the whole measurement worthless. In the 2026-08-09 reading exactly one of the 15 questions produced a citation for us, and it was the one that names Cituna directly. On the other 14, the questions a buyer asks without knowing we exist, we were cited zero times across all six engines. Claude did not cite us on any question. We checked the 29 third-party roundups the engines drew on and none list Cituna, which is exactly why a brand goes uncited.

How often does this change?

Enough that a single reading is a snapshot, not a fact. Engine answers move as the pages behind them are rewritten and re-indexed. We publish the date and the scan id for that reason, and we re-run the same prompt set daily rather than quoting a number from a post written months ago.

Which AI Visibility Tools Do AI Engines Actually Cite?

Q: How was this measured?

Repeated scans of cituna.com using one fixed prompt set: 15 buyer questions about AI visibility and rank tracking, asked across ChatGPT, Perplexity, Gemini, Claude, Grok and Google AI Overviews. That is 90 answers per reading. We count every brand named in each answer and publish the date and scan id for each one. Two readings so far, 2026-08-03 and 2026-08-09. The same measurement is repeatable on any domain.

What we measured

Almost every “best AI visibility tools” article is an editor’s shortlist. It tells you which tools the writer knew about. That is a legitimate thing to publish, but it answers a different question from the one buyers are actually asking now, which is: when someone asks ChatGPT or Perplexity to recommend a tool, whose name comes out?

So we measured it, and then we measured it again. Each reading is one scan of our own domain: the same 15 buyer questions about AI visibility and rank tracking, put to ChatGPT, Perplexity, Gemini, Claude, Grok and Google AI Overviews. That is 90 answers per reading. We count every brand named in every answer. No weighting, no editorial adjustment, no removing names we would rather not print.

Two readings so far: 3 August 2026 (scan 6a703b8e) and 9 August 2026 (scan 6a78a0b6). We publish the date and the scan id every time, because a number without them is a claim rather than a measurement, and because a second reading is the only thing that turns a list into a finding.

The results

Brands named in the most answers, out of 90, in the latest reading (2026-08-09). The faint bar behind each one is the same brand six days earlier.

Brands named across 90 AI answers. Solid bar 2026-08-09 (scan 6a78a0b6), pale bar 2026-08-03 (scan 6a703b8e). Counts are answers containing the name, not sentiment or recommendation strength. The latest reading reports its top 10 brands, so six brands measured on 2026-08-03 are outside this chart rather than at zero.

What moved in six days

The single most useful thing in this page is not the order. It is that the order changed at all, over six days, with no product launches in between.

SE Ranking: 16 to 24 answers (+8).
AthenaHQ: 13 to 20 answers (+7).
Profound: 37 to 41 answers (+4).
Peec AI: 30 to 34 answers (+4).
Otterly: 42 to 38 answers (-4), losing the top position it held on 2026-08-03.

Nothing about these products changed in six days. What changed is the set of third-party pages the engines drew on, and how those pages were re-crawled and re-weighted. That is the practical case against reading any AI visibility number that does not carry a date, including a monthly refresh: by the time a monthly figure is published, the order it describes has already moved.

One honesty note on this chart:

The 2026-08-09 reading reports its top 10 brands. Six brands measured on 2026-08-03 (LLM Pulse, Writesonic, AIclicks, Frase, Conductor, Rankscale) are not in it. That means they were outside the top 10, not that they were named zero times. Reporting an unmeasured value as a zero is the most common way this kind of index misleads, and we are not going to do it in our own.

What the ranking actually means

It is tempting to read the top of that chart as “the best tools”. It is not, and treating it that way would be the same mistake the roundups make.

Citation frequency measures one thing: how present a brand is in the pages an engine reads before it answers. That is mostly a function of how many third-party comparison posts, alternatives pages and tool roundups list you. It rewards reach and age. It is silent on accuracy, on which engines a tool actually covers, and on price.

Which means a tool can be genuinely good and near-invisible here, and a tool can sit at the top because it has been listed everywhere for two years. Both are true of names on this chart.

We are not on our own list

Cituna was not named in those 90 answers, so it is not in the chart. Adding ourselves to a measurement we did not appear in would make the measurement worth nothing. The absence is the finding: we checked the third-party roundups the engines cited for these questions and none of them list us. That is precisely how a brand ends up uncited, and it is the problem this product exists to make visible.

Our own numbers, from the same scan

We are in this market, so publishing a citation index without our own row in it would be the same editorial selection we just criticised. Here is what the 2026-08-09 scan measured about Cituna, from the identical 90 answers.

Of 15 questions, exactly one produced a citation. That question was “Is Cituna the most affordable tool for AI Visibility?”, which names us, so five engines answering it found us. ChatGPT, Perplexity, Gemini, Grok and Google AI Overviews cited us there. Claude did not cite us on any question at all.

On the other 14 questions, the ones a buyer would actually ask without knowing we exist, we were cited zero times across all six engines. Five cited cells out of 90 is a citation rate of about 5.6%, and essentially all of it comes from a branded query.

That number is bad and it is the point. Our domain authority is 0 and the site is weeks old, so the engines have no reason to reach for us yet. It is also the cleanest demonstration of the mechanic in the next section: our comparison pages are long, structured and fully marked up, and none of that put us in an answer, because the engines were reading third-party roundups instead.

Why we publish this rather than the flattering version:

A citation index whose author always scores well is a marketing page. The test of whether this one is a measurement is whether it prints a number we would rather not print, so it does, every reading.

Where the answers come from

The more useful half of the scan is not the name count, it is the sources. Across those 90 answers the engines cited 195 distinct third-party roundup and alternatives pages: posts on tool blogs, comparison sites and independent newsletters.

We fetched the 29 most-cited of those pages and checked each one. Every single one listed at least one competitor. None mentioned Cituna. Nothing about our own pages caused that; our comparison pages are long, structured and carry full schema. The engines simply were not reading them, because they were reading somebody else’s roundup instead.

That is the mechanic worth taking away, whatever tool you use: AI engines mostly quote third-party pages, not vendor sites. Improving your own page helps you rank on Google. Being present in the pages engines already read is what gets you named in an answer. They are different jobs and only one of them is on-site.

Run it on your own brand

Everything above is one scan of one domain. The same measurement runs on any domain: your tracked questions, the six engines, the brands named, and the exact source pages each engine drew on, so you can see which roundups are answering for your category and whether you appear on them.

Cituna does that daily. There is a 3-day free trial; a card is required to start it and you can cancel any time during the trial without being charged.

Which AI visibility tools do AI engines actually cite?