Method · our own data
How to measure whether AI recommends you: our method and our own data
AI visibility is whether you appear in, and are cited by, the answer when a customer asks an AI. We measure it with a fixed set of questions customers really ask, asked of several AI engines at the same time each month, with every answer and source archived as it is. We started with our own site: on 26 October 2026, NextAgent appeared in 0 of 54 answers. That is the starting point; we will update this page every month.
Updated 2026-10-27 · NextAgent · Markdown version (same text)
How is AI visibility different from search rankings?
Rankings tell you where you sit when someone searches a word. AI visibility tells you whether the AI’s answer to a question includes you and cites your page.
They are related but not the same. An AI answer names only a few businesses and blends several sources; a high-ranking site is not always cited, and a cited one does not always rank first.
How can I tell whether ChatGPT recommends my business?
You can test it yourself:
- Write down 10 to 20 questions your customers really ask, in their words and in each language you serve
- Ask them in a new chat, without earlier conversations
- Record the date, the full answer and the cited links
- Ask again on the same day next month
Answers vary from one try to the next, so one answer tells you little; look at a set of questions over several months.
How do we measure it?
- Questions: 27, by the industries we serve: 7 medical aesthetics, 7 accounting, 5 education and 8 general AI-search questions; 14 in Chinese and 13 in English, drawn from real customer inquiries and search data
- Engines: Claude (with web search) and Tavily for now; ChatGPT, Perplexity and Gemini will join once connected
- Records: the time of each question, the full answer and the cited links, archived as they are
- Frequency: monthly, the same questions, compared side by side with day 0
Our own site on day 0
26 October 2026: 27 questions × 2 engines = 54 answers, NextAgent mentioned or cited: 0 times.
The most-cited sources were Zhihu, Reddit, YouTube, Wikipedia and several agency directories. When customers ask "who should I hire", AI relies heavily on places other than the businesses’ own websites.
We will update the result here every month, whether it moves or not.
Is llms.txt worth adding?
llms.txt is a proposed convention: a short introduction and link list for AI, placed at the root of a site. It costs little, and our site has one.
It does not bring recommendations by itself: our site has llms.txt and still scored 0 on day 0. Structured data (schema) is similar: it helps search engines understand a page, without guaranteeing a recommendation. What works is answering customers’ questions clearly, keeping your details consistent, and being mentioned by trusted places.
- Source: llms.txt proposal
Is publishing the same content as a web page and as Markdown cheating?
No, as long as the content is the same. What Google prohibits is cloaking: showing search engines one thing and people another. Each of these guides has a Markdown version with exactly the same text, in a format AI reads easily.
What does it cost to test my own business?
Start with the CA$79 Website Check, which includes whether you appear in AI answers. For more depth there is the deep audit (CA$990) and monthly plans, starting at CA$800/month. All prices are on the pricing page.
Want to see how AI answers your customers today?
We ask AI the questions your customers actually ask and show you the answers and sources as they are.