Research
We out-cite a $1B company 26x in AI search
Profound has raised $154M at a $1B valuation. 22 days of publishing put us ahead of them on every AI engine that reads the web. Perplexity names us in 3,947 answers to their 152. Here is the whole measurement, 32,908 answers on the same 235 questions, including the one engine where nobody wins.
Where we rank now#
Same 235 questions, every vendor in the category, 32,908 measured answers. Times each domain was named as a source:
On Google AI Overviews we sit 12 citations behind Semrush, a public company, and 9.6x ahead of Profound. On Gemini we lead every funded platform in the category.
The entire gap was built in 22 days.
Key takeaways
- Perplexity now cites us in 2 of every 3 answers, up from 1 in 2 three weeks earlier. It rose every single week on roughly 1,700 polls each: 50.4%, 60.7%, 66.3%.
- We out-cite Profound 26x on Perplexity, 3,947 against 152.
- First in the category on three engines. Perplexity by 6x, Google AI Overviews ahead of Ahrefs and 9.6x ahead of Profound, Gemini ahead of every funded platform.
- All of it from 74 pages in 22 days. No ads, no backlink buying, no PR. Pages written to match how buyers ask, sourced, dated, served as static HTML.
- ChatGPT cited every funded platform in the category zero times, us included. It returns 2.79 sources per answer and fills them with the engine vendors' own docs. Nobody wins it, so nobody should sell it to you.
In this article
How fast it moved#
One site, one fixed set of 122 buyer questions, polled every day for three weeks. 21,454 scored answers.
Four independent weekly samples, all pointing the same way. Nothing here rests on a lucky pair of dates.
of Perplexity's answers cited us by week three, up from 50.4%
Gemini and Google AI Mode moved against us over the same weeks, and we publish that too. Gemini cut how often it cites YouTube from 47.3% of answers to 13.2% inside three weeks. Every vendor we track lost ground there, and we lost more, because our gain came from publishing pages and Gemini moved further away from reading pages. That is the useful part of measuring daily: a brand without this data finds out an engine changed its mind about a quarter late.
The work behind it#
No tricks, nothing bought. In 22 days: 74 new answer pages, 10 research posts, 216 commits, 53,389 lines. Every page written to match how a buyer phrases the question, every claim sourced and dated, all of it served as static HTML because AI crawlers do not run JavaScript.
That is the whole intervention. It is what we do for clients, run on ourselves first.
The one engine nobody wins#
ChatGPT is the exception, and it is one for the whole industry. Out of 8,266 ChatGPT answers on these questions, here is what every AI visibility platform scored:
The reason is structural and it has nothing to do with how good anyone's content is. ChatGPT names far fewer sources than any other engine:
On questions about how AI engines work it fills those few slots with the engine vendors' own documentation: developers.google.com, support.microsoft.com, help.openai.com. Only two independent publishers break in, Ahrefs and Search Engine Land, each carrying a decade of authority.
The door is shut, for everyone. Worth knowing before you pay someone to walk through it.
What this means for you#
Perplexity is winnable this quarter. It reads the live web, so a page published Monday can be cited by Friday. We went from half its answers to two thirds in 22 days, and the best funded platform in the category is still on 152 citations against our 3,947. It is the cheapest ground in AI search and almost nobody is taking it.
Do not pay anyone to put you in ChatGPT this month. Every funded platform in this category scores zero there. Ask any vendor who promises it what their own domain scores.
A single visibility score hides the useful part. Blended across engines our own three weeks read as a decline. Per engine, Perplexity gained 29% and AI Overviews 2.8x while Gemini moved against us. Three engines, three different instructions, and the blended number carries none of them.
We measured ourselves in public. We will measure your brand the same way, on your buyers' questions.
Method#
235 questions, five engines, polled daily from to . 32,908 answers collected, 30,786 scored: Perplexity 8,565, ChatGPT 8,266, Gemini 7,552, Google AI Overviews and AI Mode 6,403 combined. A citation means the domain appeared in the answer's source list, resolved through any redirect. Rates are computed per engine and never blended.
For the before and after we used a fixed cohort: only the 122 questions tracked every day from first to last, so adding questions over time cannot move the rate. Google AI Overviews and AI Mode often return a successful poll with no AI surface rendered; those are dropped from the denominator rather than scored as an absence.
This is one measured period on one query set, not a market ranking. It says how often a domain appeared on the questions we track, nothing more. There is no control group, so the changes are what happened over three weeks, not proof of what caused them. The Google AI Overviews figure moves on a first week of only 100 scored polls and should be read as directional. Gemini masks its sources behind redirect URLs; we resolve them, and unresolved cases were zero across this period.
Sources
- Noetio tracking database, 14 July 2026 to 5 August 2026. 32,908 answers across five engines, 30,786 scored. Method at noetio.com/methodology.
- Profound funding and valuation: Profound Series C announcement, February 2026. $154.5M raised across four rounds at a $1B valuation.
- Ahrefs and Search Engine Land, the two independent publishers that appear in ChatGPT answers on this question set.
The measurement is first party, read from our own polling database on . Competitor funding is from public rounds recorded in our competitor dossiers.