Your AI visibility report looks good. But who’s finding you?

Sounds promising. Until you check how many questions already included your company’s name.

Research published by Momentic on 19 October 2026 shows how much that changes the result. Across 129,652 AI answers covering 21 brands, the combined mention rate was 39%. For questions without a brand name, it was 11.3%. When the question included the brand’s name, it reached 85.8%. (Read the research)

The same research found another gap worth paying attention to. Across its category questions, brands were mentioned in 11.3% of answers but recommended in only 3.6%. These findings describe that sample. They are not benchmarks for every business.

There are several different outcomes hiding inside the phrase “AI visibility.” A useful report needs to make them clear.

What are you actually trying to measure?

Consider a recruitment agency in Zurich.

“What services does [Agency name] offer?” checks whether AI describes it correctly. “How does [Agency name] compare with [Competitor]?” checks how it performs when a buyer is already considering it. “Who can help us recruit a German-speaking finance director in Zurich?” checks whether it enters the shortlist without being named.

All three questions have business value. They represent different stages of a buying decision. Combining them into one score makes it difficult to see where the company needs work.

The same applies to what happens inside the answer. A citation to your salary report shows your content appearing as a source. Your agency appearing in a list puts your name in front of the reader. A recommendation explaining why you suit the assignment gives the buyer a reason to consider you.

A report that counts all of those as the same success leaves you guessing.

Start with inquiries that could actually become business

Take your last ten qualified inquiries. Write down the problem each customer brought, the requirements that affected their choice and the reasons they considered you.

Turn those into questions without inserting your company’s name.

For the recruitment agency, that could include a confidential replacement, a difficult language requirement or experience hiring for a particular industry. These distinctions matter because they affect which provider fits the job.

Keep questions about your own company in a separate group. You still need to know whether AI gets your services, locations and experience right.

Make the result lead to a decision

For each question, record whether your company appears, whether it is recommended and what reason the answer gives. Save the sources and the date alongside the result.

Then investigate the recurring gaps.

If your agency appears but is described as serving the wrong market, check where that description comes from. If competitors are recommended for a specialization you share, inspect the evidence supporting them. If your website is cited while other agencies are recommended, examine what the cited page establishes about your business.

Keep the questions and test conditions consistent between reviews. If you change them, show that separately. Otherwise you can make a score improve simply by asking easier questions.

You should be able to read a visibility report and know which business decision follows from it. If all it gives you is a percentage and a few flattering screenshots, you still don’t know whether your company appears when someone needs what you offer.

Every Thursday, What AI Sees explores what AI finds about companies and how it describes them.

Want to see what it finds about yours? Send us your website URL and highlight the pages you rely on to win customers.