We asked ChatGPT 11 buyer questions. One tool showed up 8 times.

How we ran the test

On October 8 and 9, 2026, we asked ChatGPT 11 questions a real buyer of AI visibility tools would ask. Logged out, web search on, one run each. Nothing clever, just what a buyer types, for different budgets, team sizes and needs. Three examples:

• “What are the best tools to track my brand's visibility in ChatGPT and AI search?”
• “I'm a solo founder with no SEO experience. Which AI tool can track if ChatGPT recommends my startup, audit my site, and write the articles to get me cited?”
• “What is the cheapest good tool to track my brand in ChatGPT, Perplexity and Google AI Overviews?”

Who ChatGPT recommended

Google agreed. In 8 similar searches, the AI Overview named Otterly 5 times, Semrush 5, Peec AI 4, Rankscale 4 and Profound 3. Two engines, one short list.

The tools that do the most got recommended the least

Sight AI, Omnia and Leapd don't just track AI visibility. They also help write and fix the content. On paper, they're the more complete products. They appeared zero times, even when a buyer asked outright for a tool that tracks and fixes. ChatGPT answered with tools that mostly track.

Where ChatGPT's answers really come from

Look at what ChatGPT cited: CiteCycle, TechnologyAdvice, CitationDesk, Zapier, Cybernews and the Semrush blog. Third-party “best tools” lists. Google leaned on the same kind of page, and Zapier's list came first more than once.

So the mechanism is simple. AI doesn't test the tools. It reads what others wrote about them, then repeats the names that come up most. The tool on 20 lists beats the better tool on 2.

A bigger dataset points the same way. Surfer studied 289,105 URLs cited by AI engines across about 1,000 prompts. Blog posts were only 28.7% of the citations, but brand mentions inside blog posts had the strongest link to how strongly AI recommended a brand. That's a correlation, and Surfer sells content tools, but it matches what we saw.

Change the question, and the winner changes

Three of our questions asked for a tool that does the work: write the content, publish it, do the outreach. Those answers looked nothing like the rest. Nimt, AirOps, Focala and AutoPublish came up, and none of the trackers did. So ChatGPT has no single “best AEO tool”. It has a best tool per question.

Location changes it too. Logged out, ChatGPT used our location in India, and two budget answers came back in rupees, with Indian tools first.

What ChatGPT checks when buyers ask about you by name

We also asked about 9 of these tools by name: “What is [tool]? Is it good, what do people and reviews say about it, and who should use it?” Every answer leaned on two sources: G2 and the company's own website. Tools with fewer G2 reviews got a more careful tone.

What to do if AI never names you

1. Ask ChatGPT your buyers' questions, then ask for its sources. Those pages are your map.
2. Get onto the lists it cites. Pitch the writers and offer free access so they can test you.
3. Write your own honest comparison page, pros and cons included. AI reads those too.
4. Build reviews on G2or the review site your category uses. That's where AI checks you by name.
5. Describe your product the same way on every site, so AI can be confident about what you do.

We're re-running these 11 questions monthly, 3 to 5 runs each. Selka is an AEO tool too, and the same rules apply to us. The AI tracker runs this kind of test for your own buyer questions, and our Reddit guide covers the community threads AI reads.