Who Writes the AI Visibility Tool Comparisons? We Counted

If you are choosing an AI visibility tool, you will read a comparison article. Probably several. So we asked a narrow question with a checkable answer: who writes them?

We pulled the two search results pages a buyer actually lands on, counted every comparison article, and looked up who published each one and where that publisher placed itself.

Of 27 comparison articles, 22 were published by a company selling a tool in the list. Most of those put themselves in first place.

We are one of them. EchoWi publishes a tool comparison and thirteen vendor reviews. This article is partly about our own category and partly about us, and the method below is stated so you can run it on us too.

The short version

1. 22 of 27 comparison articles were published by a vendor competing in the list.
2. On “best geo tools”, exactly one of fourteen comparison articles came from someone with nothing to sell.
3. At least eleven placed themselves first or second, in their own ranking.
4. One is owned by a company it ranks first, which no reader would guess from the domain.
5. The pattern is not fraud. It is the predictable result of who has a reason to write these.

Method

A limitation, stated plainly: for several vendor domains we inferred that the publisher sells a competing product from the domain and the article’s framing, without buying the product. For the eleven cases where we name a self-placement, the position comes from the ranking visible in the result itself. Where we could not see the position, we counted the article as vendor-published but did not claim a rank.

A second limitation: two queries and one country. This is a snapshot of a specific search result page on a specific day, not a longitudinal study. Repeat it and expect movement.

What we found

”ai visibility tools”

Thirteen of the twenty results were comparison articles. Nine came from companies selling in the category.

”best geo tools”

Fourteen of the twenty results were comparison articles. Thirteen came from companies selling in the category.

One independent article in fourteen. A consultant’s blog.

The Backlinko case, which is the instructive one

Most of the list above is obvious once you look. A tool’s own domain publishing a ranking with itself on top declares its interest in the URL.

One does not. Backlinko reads as an independent SEO education site, and for a decade that is exactly what it was. Its article names the Semrush AI Visibility Toolkit first.

Semrush acquired Backlinko in January 2022, as announced in Semrush’s own investor materials. So an article ranking Semrush first is published by Semrush, on a domain most readers will not connect to Semrush.

We are not alleging anything improper. Acquisitions are public, Backlinko’s ownership is not concealed, and the recommendation may be perfectly sincere. The point is narrower and more useful: domain independence is not the same as editorial independence, and the check takes one search.

That is the most transferable lesson in this article. When a comparison looks independent, look up who owns the publisher, not just who wrote the post.

Why this happens, without the outrage

It would be easy to write this up as an exposé. It is not one, and pretending otherwise would be its own kind of dishonesty.

Comparison articles are expensive to write and almost nobody has a reason to. Doing it properly means loading thirteen pricing pages, checking which vendors still exist, and understanding what a credit or a per-domain unit actually costs. That is a week of work with no revenue attached unless you sell something at the end.

Vendors have the reason. They know the category, they employ people who can write, and the article pays for itself in demo requests. So the supply of comparison content in any young B2B category is overwhelmingly vendor-produced, and the ranking inside it drifts toward the publisher.

The retrieval layer rewards it further. These articles answer the question directly, carry structured comparisons and get updated, which is precisely what both search and AI retrieval favour. The incentive and the algorithm point the same way.

So the honest conclusion is not “vendors are lying.” It is that a category this young has almost no independent press, and the buyer is reading marketing that has been formatted as journalism. That is a structural problem, not a moral one, and it will persist until the category is big enough to be worth covering by people with nothing to sell.

The lists do not agree, which is the other half of the story

If these articles were converging on the truth, they would name roughly the same tools. They do not.

Across the 27 comparisons we counted, the named products include Frase, Profound, Peec AI, Semrush, Athena, SE Ranking, Surfer, Writesonic, Otterly, ZipTie, AthenaHQ, Am I on AI, Rankshift, XFunnel, Knowatoa, Waikay, Evertune, Superlines, Rankscale, Scrunch, SISTRIX, Rankability, Hall, LLMrefs, Conductor, Quattr, HubSpot AEO, Yotpo Discover, Bluefish, Geoptie, Brandi AI, GrowthOS, AIclicks, Limy, Nightwatch, Cognizo, Trakkr, Goodie, GetCito and Respona.

That is more than forty distinct products across articles that each claim to rank the best five to twenty-two. Several appear exactly once, in the article published by the company that makes them.

Two things follow. The first is that “best” is doing no work in these titles: a category with forty contenders and no agreed shortlist has not been surveyed, it has been advertised. The second is more practical. If you read three of these articles you will meet three largely different sets of products, which feels like diligence and is closer to noise.

The honest way to narrow it is to stop asking which tools are best and start asking which ones your requirement rules out. Engine coverage, price transparency, whether the sampling is published, whether your market is covered. Those are binary and checkable, and they cut forty products to a handful in an afternoon. A ranking cannot do that for you because it does not know your constraints.

We ran the same check on ourselves

It would be cheap to publish this and not apply it, so here is our own entry in the table above.

EchoWi publishes a comparison of every tool we could verify, and we are in it. We place ourselves nowhere in particular, because the article is not a ranking: it is a table sorted by entry price with a decision framework underneath. That was a deliberate change after an earlier version did rank, and the ranking was the part we could not defend.

Where we would still fail a strict reader: we are the only vendor in our own comparison whose Chinese engine coverage is presented as a differentiator, and that is a genuine advantage we happen to sell. A reader is entitled to discount it exactly as much as they discount Profound listing agent analytics first or Otterly listing unlimited seats.

What we think earns a little trust back, and all of it is checkable rather than promised:

• Every price carries the date it was read and a link to the page it came from.
• Every review has a section on where we are the wrong choice, and one on what we could not verify.
• We correct in public. We published Profound’s valuation wrong in five places and said so at the top of the review when we fixed it, and we had Semrush recorded as bundled at the wrong price until we rechecked it.
• We report findings that do not help us: not one funded vendor in this category appeared in six runs of the query that describes the whole product category, which is inconvenient for every company in it, ours included.

None of that makes our ranking trustworthy. It makes our facts checkable, which is a smaller claim and the only one anyone in this category is currently in a position to make.

What this means for how you read them

Four habits, in order of how much time they save.

1. Check position one against the domain. If the publisher is first in its own list, you have learned something about the article, not about the tool. Read the rest for the facts, ignore the ranking.

2. Look up who owns the publisher. Backlinko is the example here, and it will not be the only one. A domain that looks independent may have been acquired.

3. Look for a date on the facts, not on the article. “Updated 2026” at the top means nothing. “Prices read from the vendor’s own page on 6 August 2026” is checkable, and its absence usually means nobody checked. When we started reviewing this category we found a shut-down product still being recommended and four vendors whose obvious domain was not theirs.

4. Prefer articles that say where they are wrong. A comparison with no section on who should not buy the product is a brochure. Ours each carry one, which is a claim you can verify rather than a promise.

What this does not show

• Nothing about product quality. A vendor ranking itself first may still make the best tool in the category. This article measures who is talking, not what is true.
• Nothing about intent. We did not ask any publisher why it ranked itself first, and there are honest answers to that question.
• Nothing durable. Two queries, one country, one day. Search results move.
• Nothing about AI answers. This is Google organic. What assistants recommend is a separate measurement, and we ran that one separately: across six runs, no funded vendor was cited at all.

Common Questions About This Study

Who writes AI visibility tool comparisons?

Overwhelmingly, the vendors themselves. In our count of 27 comparison articles across two queries, 22 were published by a company selling a tool in the list, and at least eleven placed themselves first or second in their own ranking.

Is it wrong for a vendor to publish a comparison?

Not inherently, and we publish one. What separates a usable comparison from a brochure is whether it discloses the conflict, states where its facts came from and when, and says plainly who should not buy the product.

How can I tell if a comparison is independent?

Check whether the publisher appears in its own list, and look up who owns the domain. Backlinko reads as an independent SEO site and has been owned by Semrush since January 2022, which is not obvious from the page.

Why are there so few independent comparisons?

Because doing one properly costs a week of work and produces no revenue unless you are selling something. Young B2B categories have almost no independent press, so the supply is vendor-produced by default.

Does this mean the rankings are wrong?

It means the ranking carries no information. The underlying facts in a vendor comparison are often accurate, because they are checkable and being caught is expensive. The order is the part shaped by the publisher’s interest.

How do I reproduce this?

Search the two queries in the same market, list the comparison articles, and for each one check whether the publisher’s product appears and where. It takes about an hour, and the method is in the table above.

Where this leaves you

The useful finding is not that vendors write comparisons. Anyone would guess that. It is the ratio: on the main GEO tools query, exactly one comparison article in fourteen came from someone with nothing to sell.

That changes how much weight any single article deserves, including this one. Read three, ignore all three rankings, and keep only the facts that carry a date and a source you can open.

And apply it here. We publish a comparison, fourteen reviews, and this study. We compete with almost every company named above. The correct response is not to trust us more than them; it is to check the parts that are checkable, which is why every price we publish carries the date it was read and the page it came from.

Next: what the AI recommends when you ask it, measured six times, the same question in three markets, and our own comparison, with its conflict declared.