Our own numbers

We measure ourselves with the instrument we point at your brand.

Every figure below is read from our own tracking engine when this page is built, and it rebuilds every hour. The prompt set is published in full. So is the competitor list the maths depends on. So are the questions where we do not get named at all.

Last read 10 September 2026, 19:55 UTC.

2.9%

RankZero was named in 88 of the 3,072 AI answers our own prompt set produced in the 30 days to 10 September 2026.

31 published prompts, run against ChatGPT, Google AI Overviews, Google Gemini and Perplexity. The denominator is every one of those 3,072 answers, not a subset we picked.

Why this denominator, and not our own headline metric.

Share of voice, the figure our product leads with, is measured against a list of competitors we choose. Change the list and the percentage changes, backwards as well as forwards, without a single answer being different. So this page divides by the prompt set instead. That set is published below, it is the same set every day, and nothing we do to our own settings can move it. Share of voice is still on this page, further down, with that warning attached to it.

1.4%
The same figure over 7 days
6 of 436 answers. A week is a few hundred runs, so it swings. It is here because one window is the easiest number to choose flatteringly, and the two are directly comparable because they divide by the same thing.

This is not a traffic number.

It counts how often an AI answer names us. It says nothing about how many people then visited this site, and we are not going to imply otherwise. Those are different measurements and only one of them is on this page.

Secondary

Share of voice, and why it sits below the number above.

4.87%
Share of voice
#9
Rank, out of the 19 brands we track

Share of voice divides by a smaller number than the headline does. Its denominator is the 1,808 answers that named at least one brand on our tracked list, not all 3,072 answers. Ranking position works the same way, against the 19 brands listed further down.

That list is one we chose, which means the percentage moves when we change it. Stop tracking the three brands named most often (Ahrefs, SE Ranking and Profound) and the very same 88 answers read 6.59% instead of 4.87%, without a single extra answer naming us.

It moves the other way too, and it just did. 5 of the 19 brands below (Semrush, Omnia, Rank Prompt, ZipTie AI and Radyant) were added to the list inside the window shown here. Adding a competitor pulls more answers into the denominator, so every other percentage falls, ours included, with nothing having changed in what the models actually said.

A number our own settings can move by points is not the one to lead a page like this with. That is the whole reason the figure at the top divides by every answer instead.

Checked on this render: our share of voice recomputed from the counts on this page matches what the tracking engine reports, to within a tenth of a point.

The prompt set

20 of these 31 questions did not name us once.

This is the whole set, not the flattering part of it. Every prompt we track is here with the number of answers it produced, how many named us, and the brand named most often in the ones that did not. Nobody in this category publishes the second table.

PromptAnswersNamed usShareNamed most often
white label AI search tracking tool for SEO agenciesUnited Kingdom832833.7%SE Ranking (70)
what is ai search visibility?United Kingdom1022423.5%Mentions (43)
why is my SaaS brand not showing up in ChatGPT resultsUnited Kingdom1002323.0%Mentions (14)
what is LLMO and how does it work for SaaS companiesUnited Kingdom10144.0%Mentions (16)
AI SEO service for SaaS brandsUnited Kingdom10232.9%Ahrefs (44)
how to get your SaaS product mentioned in AI search answersUnited Kingdom8411.2%Mentions (21)
how do we track competitor visibility in ai search results?United States9611.0%SE Ranking (60)
best AI search optimization service for B2B software companiesUnited Kingdom10011.0%Profound (44)
ai seo agencyUnited States10111.0%ZipTie (36)
How do I track my competitors' AI visibility and ranking?United Kingdom10111.0%SE Ranking (67)
how to add AI search monitoring to an SEO agency's service offeringUnited Kingdom10211.0%SE Ranking (53)
agency to help my SaaS company rank in ChatGPT and AI searchUnited Kingdom840noneZipTie (10)
AI SEO strategy for B2B SaaS companiesUnited Kingdom1000noneAhrefs (44)
best AI SEO software for marketing agenciesUnited Kingdom970noneSurfer SEO (77)
best platform to manage AI SEO across multiple clientsUnited Kingdom1020noneSE Ranking (67)
best tools for SEO agencies to monitor AI search rankingsUnited Kingdom1020noneAhrefs (92)
How can I improve my website’s SEO to appear in AI answer snippets?United Kingdom1000noneSE Ranking (33)
how do AI models like ChatGPT and Perplexity decide which brands to mentionUnited Kingdom1020noneMentions (36)
how do AI search engines decide which SaaS tools to recommendUnited Kingdom1000noneMentions (23)
how is AI search changing the way people find brands onlineUnited Kingdom990noneZipTie (7)
how to optimize a SaaS brand for Google AI Overviews and PerplexityUnited Kingdom1000noneZipTie (21)
How to rank in chatgptUnited Kingdom1030noneMentions (34)
how to rank in google ai overviewUnited Kingdom1030noneSurfer SEO (29)
how to report on AI search performance to clientsUnited Kingdom1000noneSE Ranking (21)
how to track a client's brand visibility in ChatGPT and PerplexityUnited Kingdom1020noneAhrefs (28)
What are the best strategies to optimize my content for AI-powered search results?United Kingdom1000noneSemrush (15)
What does visibility mean in SEO?Germany1000noneAhrefs (45)
what is AI SEO and how is it different from traditional SEOUnited Kingdom1010noneSemrush (25)
what is share of voice in AI search and how do you measure itUnited Kingdom1010noneMentions (49)
what is the best ai seo analytics toolUnited Kingdom1020noneAhrefs (82)
What steps can I take to ensure my brand is recommended by AI assistants like ChatGPT or Perplexity?United Kingdom1020noneMentions (24)

Country is the locale the prompt runs in, shown next to each one. 28 in United Kingdom, 2 in United States and 1 in Germany. The set is weighted the way it is because it grew out of the questions our own buyers ask, not out of a country plan, and that is a limit of this measurement rather than a feature of it.

The other half of the method

The 19 brands the ranking is measured against.

A share of voice figure means nothing without the set it is a share of. This is ours, as it stands at the timestamp on this page, with the date each brand was added.

#BrandShare of voiceTracked since
1Ahrefs43.03%11 September 2025
2SE Ranking40.10%29 July 2025
3Profound25.88%28 July 2025
4Mentions25.17%28 July 2025
5Semrush22.46%31 August 2026added inside this window
6Surfer SEO19.80%11 September 2025
7ZipTie10.56%28 July 2025
8Peec9.73%28 July 2025
9RankZerous4.87%28 July 2025
10Rankscale4.15%28 July 2025
11AEO Engine1.88%23 June 2026
12BuzzSumo0.83%28 July 2025
13SpyFu0.77%28 July 2025
14Omnia0.72%31 August 2026added inside this window
15Rank Prompt0.28%31 August 2026added inside this window
16ZipTie AI0.22%31 August 2026added inside this window
17SmartSites0.06%28 July 2025
18Sembly AI0.00%29 July 2025
19Radyant0.00%31 August 2026added inside this window
Method

How this is measured, and what it cannot tell you.

How it is measured

Each of the 31 prompts is sent to ChatGPT, Google AI Overviews, Google Gemini and Perplexity on a daily cycle, in the country listed beside it. That produced 3,072 answers in the 30 days covered here, the first on 11 August 2026 and the last on 10 September 2026.

A mention counts once per answer. An answer that names us three times is one answer, not three. We read the answer text the model returned, not a search result page.

The figures come from the same tables and the same query that serve our customers' dashboards and our public API. There is no separate reporting path for this page.

What it cannot tell you

The brand list is closed. An answer that names a company nobody has added to a tracked list is recorded as naming no one, so "answers that named someone" is a floor and not a count of the whole field.

The window stops at 90 days. Anything older than that is not reachable through this measurement, so this page cannot show you a year of history and does not pretend to.

And the number moves. This page reports the window it is looking at right now, not the best window we have ever had, which is why the seven day figure sits next to the thirty day one at the top.

There is no trend line on this page, on purpose. When a brand joins the tracked list, only the most recent few days of answers are re-read against it, so a share of voice chart spanning that moment would slope gently downwards when what actually happened was a settings change on one day. We would rather show you two windows that divide by the same fixed prompt set than a curve that cannot tell those two things apart.

We will run this on your prompts and show you the same table.

The audit call maps the questions your buyers actually ask, runs them against the same models, and gives you your own version of the page you just read. Including the rows where you lose, because those are the ones worth working on. See what the program covers.