Visibility went up. Did revenue? How to close the loop on AI search

Three questions, usually answered by three different tools

Question	Answered by	What it can’t tell you
Does AI recommend us?	AI visibility tools	Whether anyone acted on it
Who visited the website?	Analytics	Which visitors became paying customers
Which jobs did we book?	Your CRM	Where those customers first heard of you

Each tool is honest about its own slice. The trouble is in the joins. A report that says “mentions up 30%” and a CRM that says “bookings flat” leave you guessing whether the work paid off.

What closing the loop looks like

Take a moving company in Austin.

1. Visibility: a weekly check asks ChatGPT the questions customers ask, such as “best long-distance movers in Austin”. It records whether the company is named, in what position, who is named instead, and which websites the answer drew on.
2. Attribution: the website records each visitor who arrives from ChatGPT, the page they landed on, and whether they sent a form or tapped the phone number. Customers who say “I found you on ChatGPT” are counted too, labelled separately.
3. Revenue: when a lead is booked, its value is recorded against the same lead.

Now the sentences that matter can be written: “ChatGPT named us for two of three questions this month. It sent 98 visitors. Six became leads, five booked, worth $10,900. We’re left out of the long-distance question, and that page is where most AI leads land.” That last sentence is a decision, not a statistic.

The honest limits

Correlation, not proof. If visibility and AI leads rise together, that’s encouraging. It isn’t a controlled experiment. Seasonality and other marketing move the same numbers.

AI answers vary. The same question can get a different answer an hour later. That’s why a weekly check and a trend are more useful than any single result.

Attribution is a minimum. Some customers ask an AI and then type your address in directly. No tool sees that visit as AI. It’s why every figure here is labelled by evidence, and why the total is a floor you can defend, not an estimate. (More: why every AI attribution tool undercounts.)

Where Which AI Sent fits

Attribution is the core of the product: AI visits through to leads, bookings and revenue. On top of that it runs a ChatGPT visibility check each week on paid plans, with a scoreboard of who is named and the sources behind the answers. It isn’t trying to replace a full visibility platform that tracks hundreds of prompts across every engine. If you have one, keep it. What it adds is the join: the What to do next card reads both sides and points at the gap worth closing.

For agencies, that join is the client conversation: not “mentions are up”, but “this work produced these leads and this revenue”. The numbers are also available through the API for your own reports.