Is my company named when buyers ask an AI which tool to use?
Fifty buying prompts, twelve B2B SaaS categories, five engines: ChatGPT, Claude, Perplexity, Copilot and Google (Gemini and AI Overviews). Three runs each, on different days, nothing discarded for being unflattering.
0/50Answers that named someone else
0 named you
Research9 min read
Most agencies selling AI visibility put themselves at the top of their own best-agency lists. We read 107 of their sites and compared what they charge and promise with who AI names.
Anyone can publish a number. Almost nobody publishes the method.
Prompt set, engine list, run count and scoring rule, written down before the run rather than after it. If the method is not in the post, the number in the post means nothing.
Published at the same length as the tests that found something. A desk that only publishes what worked is a marketing department with a research budget.
No blended visibility index. Raw counts, shares and distributions, reported per engine, so you can disagree with our reading and still use the data.
Engines change retrieval behaviour without announcing it. Every post carries the date it was run and the date its sources were last checked.
The Research posts meet all four conditions and publish their data. The guides to getting cited, measurement, choosing a partner and pricing carry sources and a checked date but no method of their own, because a published rate or a buying decision is not a finding and there is nothing to re-run.
Every post here came out of the same panel we run for clients. The first one tells you where you stand across the five engines, with the nulls left in. If the answer is that you are already fine, we will say so and you will have spent nothing further.