Visibility is the start of the story. The end of it is traffic: the real people an assistant sends to your site after naming you. Unlike a vanity score that only ever goes up, referral visits can be measured, and measuring them keeps everyone honest about whether the work is paying off.
Where AI referrals show up
When an assistant links you and a person clicks, that visit often arrives with a referrer from the assistant's domain. In GA4 you can group these sources to see visits from ChatGPT, Perplexity, Gemini and the rest, and watch the trend over time as your visibility improves.
Why the measured number is a floor
Here is the honest caveat. Some AI visits arrive with no referrer at all, because a person read an answer, remembered your name and typed it in later. Those still count, but they are hard to attribute. So a measured AI referral number is a floor, not a ceiling: the true influence is at least that good, and often better. Presenting that floor as the whole truth would be dishonest, which is why every figure should carry a label.
From measurement to proof
Once you can see the visits, you can tie them back to the fixes that earned them and prove cause and effect, month after month. That is the difference between a dashboard that flatters you and one that tells you the truth. It closes the loop that starts with measuring visibility and runs through the work to get cited.
You can connect analytics read-only and see AI referral traffic broken out by assistant. Start by seeing where you stand today with a free scan.