What is AI visibility, and why it's the new front door
AI visibility is how often AI assistants name your brand when buyers ask them a question in your category. It is measured by asking the assistants the questions your buyers ask, recording which brands they name, and tracking that over time. It is not the same thing as a Google ranking, and a good position on Google does not guarantee it.
Why a Google ranking does not carry over
Ahrefs tested 15,000 long-tail queries against ChatGPT, Gemini, Copilot and Perplexity and found that only 12% of the URLs those assistants cited also ranked in Google's top 10 for the original prompt. Per engine the spread was wide: Perplexity 28.6%, Gemini 8.6%, Copilot 8.2%, and ChatGPT 8% for in-text citations and 6.1% for its reference list.
The researchers' own reading is that assistants retrieve differently, breaking a question into several sub-queries rather than reading a ranked list. The practical consequence is simple: your Google position tells you very little about whether an assistant will mention you.
What AI visibility actually measures
A usable measurement has four parts, and a number missing any of them is not worth acting on.
- The prompt set: the real questions buyers ask, not keywords rewritten as questions.
- The engine: ChatGPT, Google AI Overviews and Perplexity answer differently, so a blended score hides more than it shows.
- The sample size: how many times each prompt was asked, because assistant answers vary between runs.
- The date: answers change as models and indexes change, so an undated score is not a score.
Branded and unbranded questions are different tests
The most common mistake is measuring only prompts that contain your brand name. An assistant asked "what is Acme Tools" will almost always describe Acme Tools. That is a reading-comprehension test, not a visibility test.
The question that matters commercially is the unbranded one: "best tools for a small workshop". If you are named there, you are in the consideration set. Track the two separately and never average them into one figure, because a strong branded score will mask a discovery score of zero.
Where AI answers show up at all
Ahrefs' analysis of 55.8 million AI Overviews drawn from 590 million keywords found that AI Overviews appear on roughly 12.5% to 16.5% of searches in the countries where they have rolled out, and about 16% of US desktop keywords. Notably, 97.70% of those AI Overviews were on informational queries.
That shape matters for where to spend effort: AI answers currently dominate the research stage of a buying decision rather than the final transactional query.
How to start
Pick ten questions a buyer would genuinely type, ask each one across every assistant you care about, and write down which brands get named and how often. Do it again in a month. That is the whole method, and it is the method Vorne automates.