Measuring AI Visibility

• Correlation is not causation: how to honestly measure AI visibility work

  "Our work got you mentioned" requires isolating one cause in a system that changes weekly for its own reasons. Nobody can fully do that, including us. Here is what honest looks like instead.

• What a good AI visibility report looks like (and what a junk one looks like)

  The five-second tell is whether every claim has an answer attached. A real report shows you what the AI said. A junk one shows you a scary score and an invoice.

• How we track AI mentions: our methodology in full

  The full method behind our numbers. What the free scan checks, what continuous tracking measures, how findings get graded, and what we refuse to conclude.

• What we log and why: the case for keeping receipts on AI answers

  Every answer stored word for word, dated, with the question and the work performed alongside. Receipts are what make honest before-and-after possible in a field that mostly asserts it.

• Mention rate, citation rate, and share of voice: the three numbers that matter

  Mention rate is how often AI names you. Citation rate is how often it credits a source about you. Share of voice is your slice of all names given. Each is useful, each is blind somewhere.

• Why one screenshot of ChatGPT proves nothing

  One answer, one phrasing, one day, one location. Change any of the four and the answer can change. A single screenshot is a data point, not a verdict.

• Prompt sampling: why we test ten prompts and not one

  One question is an anecdote. Ten designed questions are a sample. Here is how the ten get chosen, what a sample that size can support, and why not a hundred.

• Tracked prompts: the questions behind your mention rate

  A tracked prompt is a question customers ask AI assistants, tracked so you can see whether your business gets mentioned in the answers. Here is why questions, not keywords, are the unit.

• What competitor mentions tell you (and what they don't)

  When AI names a competitor instead of you, it just showed you which sources it trusts in your market. That's a diagnostic, not espionage, and it's the fastest map to what you're missing.

• When the free monthly check stops being enough

  A manual monthly check of what AI says about you is genuinely good, and for some businesses it's all you need. Here is the honest math on where it stops, in owner-hours and blind spots.