Research
Original studies on how AI assistants pick the companies they name: which sources they lean on, how answers differ between engines, and how much of it moves month to month. Every study publishes its question set, its sample size, the systems and dates it ran on, and the method behind each figure, so the numbers can be argued with.
One rule, and it is why this page is short: a quantitative claim appears here with its source and its sample, or it does not appear. We do not repeat an industry statistic we cannot trace, and we do not put a figure on a chart that came from a single run.
The first study is still being run
We are not going to fill this page with a summary of somebody else’s survey. The first benchmark is in progress: a fixed prompt set across the assistants we test, repeated over time, measuring how much the recommended shortlist actually moves in a B2B software category.
In the meantime, the method it will use is already published, and the sample report shows the shape of the output.
Commission the same research on your category
A written AI visibility audit of your market: where you are cited today and by which systems, who is named instead of you, the sources deciding it, and a twelve-month plan. From $1,500, credited against a Scale or Agency subscription. It is a measurement and a plan, not a guarantee that any AI system will name you.
What an audit covers
- Where you are cited today, by source and by AI system
- The strengths worth protecting
- The gaps that need attention first
- Which competitors win, and on which questions
- A twelve-month plan, sequenced
The AI Search Brief
How AI assistants answer buying questions, which sources they lean on, and what that means for tech brands. Sent when we have something worth sending.