The tools AI assistants cite for AI-visibility tracking do not necessarily have large Google search footprints.
In Brandvane’s sample on 2026-08-28, several domains appeared repeatedly in answers while Ahrefs reported only small numbers of ranking keywords or estimated organic visits for some of them. That is not a judgment about product quality. Several are legitimate products. It is an observation that recommendation visibility and conventional organic-search footprint did not move together neatly in this category and sample.
This was one 23-answer sample on one day. It is not a market census, assistants vary between runs, and the pattern does not establish causation.
See where your brand stands When our own measurement was wrong
Read the transcript
We asked ChatGPT, Claude and Perplexity who tracks AI visibility. 27 planned samples, 23 came back.
Eight domains came up again and again. The most-cited one appeared in 8 of the 23 answers.
Then we looked them up in Ahrefs. That most-cited domain had 8 ranking keywords. Another one, cited three times, had two.
But not all of them. One had 532 keywords. So the headline isn’t that these tools don’t rank.
The real finding is narrower. Getting cited over and over didn’t require a big search footprint. The two just didn’t move together.
We appeared in 0 of 23. We were two days old. See where your brand stands, free, at brandvane.ai.
What we sampled
We sampled 5 buyer-intent prompts about AI-visibility tracking across OpenAI, Anthropic, and Perplexity on 2026-08-28. Of 27 planned samples, 23 returned answers.
Brandvane appeared in 0 of 23 returned answers. We were two days old. That absence is the honest baseline for a new domain, not a result to explain away.
The figures in this article are answer-level counts from that collection and Ahrefs footprint data retrieved on the same date. We are not filling in prompt-level or engine-level splits that were not part of the supplied result set.
Which domains the answers cited
The cited-domain count below is the number of returned answers containing the domain, out of the same 23 returned answers.
| Domain | Answers citing domain |
|---|---|
| siftly.ai | 8 of 23 |
| workduo.ai | 7 of 23 |
| dageno.ai | 6 of 23 |
| trysight.ai | 4 of 23 |
| aeovision.ai | 4 of 23 |
| ziptie.dev | 3 of 23 |
| chatobserver.com | 3 of 23 |
| genrank.io | 3 of 23 |
These are citation-incidence counts, not a permanent ranking of the companies. One answer can cite more than one domain, and another run can return a different source set.
The table also does not say that a higher count means a better product. A citation is evidence that a source appeared in a sampled answer. Product quality, customer fit, accuracy, and commercial maturity require different evidence.
Their Ahrefs search footprints
We retrieved the following Ahrefs footprints on 2026-08-28 using US organic data. The available fields are not identical for every domain, so the table retains the supplied metric instead of treating an unreported field as zero.
| Domain | Domain Rating | Supplied US organic footprint |
|---|---|---|
| siftly.ai | DR 35 | 8 ranking keywords |
| genrank.io | DR 39 | 7 estimated monthly organic traffic |
| chatobserver.com | DR 28 | 2 ranking keywords |
| dageno.ai | DR 32 | 1,300 estimated monthly organic traffic; 174 ranking keywords |
| trysight.ai | DR 55 | 2,063 estimated monthly organic traffic; 532 ranking keywords |
| ziptie.dev | DR 71 | 1,833 referring domains |
The differences matter. Ahrefs reported 174 ranking keywords for Dageno and 532 for Trysight, compared with 8 for Siftly and 2 for ChatObserver. It reported 1,300 estimated monthly organic traffic for Dageno and 2,063 for Trysight, compared with 7 for GenRank. Ziptie’s supplied metric was 1,833 referring domains, which is not the same measure as traffic or ranking keywords.
That is why “these tools do not rank on Google” is a useful provocation, not a literal claim that every cited domain had no Google presence. The measured pattern is that repeated AI citation did not require a uniformly large US organic footprint in this category.
What appears decoupled
In this sample, AI recommendation appears decoupled from Google ranking in the AI-visibility-tool category. The frequently cited set included domains with very different Domain Ratings and very different supplied organic footprints.
That observation challenges a simple assumption: that the products assistants surface must be the products with the largest conventional search presence. The sample did not show such a clean ordering.
It does not show why a domain was cited. The answers may have drawn on product pages, third-party sources, link authority, topical specificity, freshness, retrieval behavior, or other signals. The supplied figures do not isolate those mechanisms, so we do not assign a cause.
It also does not show that Google visibility is irrelevant. Dageno and Trysight combined repeated citation with the larger supplied traffic and ranking-keyword figures in this result set. Ziptie’s referring-domain count is a different measurement, so we do not substitute it for traffic or keyword data that were not supplied.
The defensible conclusion is narrower: large Google organic footprints were not a prerequisite for repeated citation in these 23 returned answers.
What the sample says about Brandvane
Brandvane’s 0 of 23 result belongs in the report. Being two days old provides context, but it does not turn absence into success. We have no Brandvane appearance in this collection to narrate as visibility.
That baseline is useful. Later collections can ask whether Brandvane appears in a disclosed number of returned answers, whether its domain is cited, and which sources recur when it is absent. The later result must keep its own denominator and collection conditions; it should not be described as a fixed rank.
Publishing the zero also prevents a conflict common in vendor research: using a market observation to imply that the vendor conducting the study already leads the market. We do not.
What this study does not establish
This research does not establish:
- what every user saw;
- how often buyers ask the sampled prompts;
- whether the same domains would appear in another run;
- an engine-specific difference;
- product quality or customer satisfaction;
- a universal relationship between Ahrefs metrics and AI citations; or
- whether search footprint caused, prevented, or had no effect on citation.
Assistants vary between runs. One day of sampling cannot separate ordinary answer variability from a durable pattern. Correlation here is not causation.
The named companies should therefore be read as observed sources, not winners or losers. Several are legitimate products that simply have small search footprints in the supplied Ahrefs data. This study is about the mismatch between two measurement surfaces, not a disparagement of their businesses.
How we would read a later sample
A later collection should preserve the same evidence boundaries: planned samples, returned answers, exact date, tested engines, and domain citations out of returned answers. Any prompt, model, or collection change should be disclosed rather than hidden inside a trend.
If the cited-domain pattern repeats, the result becomes a recurring observation across disclosed samples. If it changes, that variability is part of the finding. Neither outcome creates a permanent AI rank.
For the evidence framework, see how Brandvane measures AI visibility. Automated weekly AI reports for ChatGPT and Google AI Overviews are coming soon; the planned reports will keep the two surfaces separate and show completed-check counts.
The takeaway
Across 23 returned answers from 27 planned samples on 2026-08-28, assistants repeatedly cited several AI-visibility-tool domains with small supplied Google organic footprints. Brandvane appeared in 0 of 23 while it was two days old.
The result supports one bounded conclusion: AI recommendation appeared decoupled from Google ranking in this category and sample. It does not support a causal story, a universal rule, or a claim about the next answer.