Analytics, rewritten: the answer happens before the visit.

Buyers now get a shortlist from an AI answer and never touch your site. The dashboard sees none of it.

What the old playbook assumed

That the journey ran through your site, so sessions, rankings and conversion rate described it. If a buyer considered you, there was a visit somewhere to prove it.

What broke

The shortlist is increasingly written by an answer engine, on a page you do not own, in a conversation you cannot see. A brand can be left out of every answer while its traffic report looks normal, because the buyers it lost never arrived to be counted.

What to measure instead

The questions your buyers actually ask, put to the engines they use, on a schedule. For each answer: were you retrieved, cited, named, recommended? Share of Recommendation is how often you are the one named when a buyer asks who to choose.

One reading is noise

The same question to the same engine gives a different answer on a different day. A single screenshot proves nothing in either direction. Read it as a trend across repeated runs, and never compare two readings taken with different instruments.

What counts as proof

A claim you can check: the question, the engine, the date, and the answer as it was captured. If an agency or a tool cannot show you those four things, you are looking at an estimate.

Everything in this guide

Questions

Can Google Analytics show AI visibility?

Only the small share of AI answers that end in a click. It cannot show the answers where you were left out, which is where the loss is.

Which AI visibility tool should we buy?

Start by running a manual panel on your own buyer questions for a week. You will learn what you need to track before a vendor tells you.

How often should we measure?

Often enough to see a trend, which means repeated runs rather than a single check. Weekly is a reasonable place to start.

Run the free audit