We asked AI which SEO tool to buy, 74 times. Two brands won almost everything.

We asked ChatGPT, Perplexity and Gemini the questions a real buyer asks before choosing an SEO tool. Twenty-five questions, three engines, 74 usable answers.

Two brands took almost everything.

Semrush was named in 78% of answers. Ahrefs in 74%. One of those two was the first brand mentioned in 40 of the 74 answers. Everybody else in a crowded, well-funded category is fighting over what's left.

Here's the whole table.

Which SEO tools do AI engines actually recommend?

74 answers across ChatGPT, Perplexity and Gemini. July 2026.

The gap between second and third place is enormous. Ahrefs at 74%, then Google Search Console at 50% — and Search Console isn't even a competitor, it's a free Google product the models mention out of reflex. The first real challenger, Moz, sits at 36%.

The Moz problem: mentioned constantly, recommended never

Look at row four. Moz appears in 27 answers and was named first in zero of them.

That's a very specific kind of failure, and it's worth understanding because it's probably happening to you. Moz has enormous brand equity. The models clearly know it exists — it turns up in more than a third of all answers. But it always turns up as the third thing in a list, after Semrush and Ahrefs, in the "other options include…" clause.

Being known is not the same as being recommended. A model that lists you third is a model that has learned you belong to the category but hasn't learned you're the answer to the question. If you're measuring your AI visibility with a simple mention count, Moz looks fine. It isn't.

This is why Share of Voice, properly calculated, matters more than raw mentions.

The part that made us laugh, then stop laughing

We slipped five questions into the study about AI visibility tools — our own category. Things like "best tool to track how my brand appears in ChatGPT" and "which tool tracks brand mentions in AI answers".

Here's who the engines recommended when asked, specifically, for an AI-visibility tool:

14 answers across three engines.

Read that again. When you ask an AI engine which tool to use for tracking your brand in AI answers, it recommends Ahrefs and Semrush more often than it recommends any company that actually builds AI visibility software.

Profound has raised $155M and is the best-known name in this space. It was named twice. Peec AI has raised $29M. It was named four times. Trakkr, AthenaHQ, Scrunch, Rankscale, Knowatoa: zero.

And Spottlo — the company that built the tool that ran this study, and wrote this post? Zero mentions out of 74 answers. We are invisible in the exact thing we sell.

We're not going to pretend that's a clever marketing angle. It's just true, and it's the most useful data point in the whole study.

What this actually tells you

Three things, and none of them are about SEO tools specifically.

1. AI answers are far more concentrated than search results

A Google results page gives you ten organic links, plus a knowledge panel, plus a few ads. There's room to be fifth and still get clicked.

An AI answer names two or three brands. That's the whole page. If you are not in that handful, you do not exist — there's no scrolling, no page two, no long tail. Only 8% of users click any result when an AI summary is shown, and just 1% click a link inside the summary itself.

The winner-take-most dynamic in that table isn't an artifact of our method. It's what the format does.

2. Funding does not buy AI visibility

This is the finding I'd tattoo on the category. Profound has raised $155M and cannot get an AI engine to name it when asked directly about its own category. Meanwhile Ahrefs — which is bootstrapped and does not primarily sell AI visibility — gets recommended for it eight times.

Why? Because Ahrefs has been written about, linked to, reviewed, compared, argued over and listed in "best tools" roundups for fifteen years. The models have read a decade of the internet talking about Ahrefs. They have read almost nothing about Profound, because Profound is two years old.

AI engines recommend what the web has already corroborated. They are not evaluating your product. They are aggregating what other people have said about it. This is the single mechanic that explains why your competitor shows up in ChatGPT and you don't, and it's why you cannot buy your way in with a bigger content budget on your own domain.

3. New categories are invisible by default — including yours

If you are building something genuinely new, the models have no corroboration to draw on, so they fall back to the nearest established category. Ask for an AI-visibility tool and you get an SEO tool, because "SEO tool" is a concept the model has seen ten million times and "AI visibility tool" is one it has seen a few thousand.

That's not a bug in the model. It's a description of where your category actually sits in the world's collective writing. And the fix isn't clever prompting — it's getting other people to write about you.

How we ran this

No trickery, and you can reproduce it.

• 25 buyer questions, written as a person would actually type them. "What is the best SEO tool for a small agency?", "What is the best alternative to Ahrefs?", "Best rank tracker for tracking hundreds of keywords?"
• Three engines: ChatGPT, Perplexity and Gemini, through their APIs. We left out Google AI Overviews because it bills per search against a hard cap and wouldn't have changed the shape of the result.
• 74 non-empty answers collected. We counted a brand as "named" if it appeared anywhere in the answer text, and "named first" if it was the first brand to appear.
• We did not tell the engines to rank anything, mention anyone, or produce a list. We asked the question and read what came back.

The obvious caveats: LLM outputs are non-deterministic, so a rerun will shift the individual numbers. Seventy-four answers is enough to see a 78%-versus-3% gap; it is not enough to tell an 8% brand from a 6% one. Trust the top of the table, be skeptical of the middle.

What to do next

1. Find out where you actually stand. Not where you think you stand. Run a free report on your domain — it takes about 30 seconds, no signup, and it will scan you across all four engines including AI Overviews. If you're choosing a tool to do this continuously, compare every AI visibility tracker side by side or read the full buyer's guide.
2. Check whether you're even crawlable. A surprising number of brands are blocking GPTBot at the CDN without knowing it. Ours was. Check yourdomain.com/robots.txt from the public internet, not from your server.
3. Stop optimising your own domain and go get corroborated. Reviews on G2 and Capterra, inclusion in "best X tools" roundups, genuine mentions on Reddit and in industry press. That is the raw material the models are actually reading.
4. Measure the right number. Mentions are vanity if you're always mentioned third. Track Share of Voice and position, not just presence.
5. Be patient, and be honest with yourself about the timeline. Corroboration accretes over months. Anyone who tells you they can get you recommended by ChatGPT next week is selling something.

We'll rerun this study each quarter and publish the movement. If a challenger ever breaks Semrush and Ahrefs' grip, that will be worth knowing — and worth understanding how they did it.