Is any company invisible to AI engines entirely

A natural fear for any company evaluating its AI presence is total invisibility, the possibility that no engine, under any prompt in the study, ever names it. This finding checks that directly.

Across the full cohort of 30,564 companies drawn from 170,457 usable answers to 2,500 prompts, the count of companies that never appeared on any of the three engines, ChatGPT, Claude, or Gemini, under any prompt in the study, was 0. That is 0% of the cohort.

In other words, essentially every company in this study's cohort was named by at least one engine at least once. This is a meaningfully different result than finding that most companies struggle for any AI presence. It means the more common problem, documented in the companion finding on cross-engine agreement, is partial visibility, not zero visibility. Companies show up, but inconsistently across engines, rather than not showing up at all.

Why might this be true? The prompts used in this study were built around real buyer intents in specific categories, commercial-investigation, informational, navigational, and transactional, across a wide range of business categories. A company that made it into this cohort at all did so because it was named at least once by at least one engine somewhere in the dataset. That is partly a construction effect: the cohort itself is built from companies that appeared in engine answers, not from an independent registry of all companies that exist in each category, so a company with truly zero presence on any engine would not have entered the cohort to be counted in the first place, unless it was named by exactly one engine on exactly one prompt and nothing more.

This matters for interpretation. The 0% figure should be read as 'of companies that appeared in engine answers somewhere in this study, none failed to appear on any single engine when checked across the full prompt set,' rather than as a claim about every company that exists in these categories, including ones this study's prompt set never surfaced at all. A company entirely absent from the underlying web and directory sources these engines draw from would simply never enter this cohort, and this finding cannot speak to that population.

What follows for a company checking its own standing: if you are already appearing in AI answers at all, even rarely or on only one engine, the realistic next question is not 'how do I become visible' but 'how do I close the gap between the engine where I appear and the engines where I don't,' which is a narrower and more solvable problem. The companion finding on 96% of the cohort being visible somewhere but not everywhere describes the size of that more common gap in detail.