
How to Tell If AI Search Is Sending You Leads
Measure it three ways: referral traffic from assistant domains in your analytics, a tracked panel of buying questions you re-ask on a schedule to see whether you are named, and softer signals like branded searches and enquiries that arrive already knowing what you do.
This is the most common reason businesses dismiss AI search: there is no ranking report, so it feels unmeasurable. It is measurable, just not with the tools built for the previous era.
Do AI assistants show up in analytics at all?
Sometimes, yes. When an assistant links a source and the user clicks, it arrives as ordinary referral traffic from that assistant's domain, so you can segment it in your analytics the same way you would any other referrer.
The catch is that this only counts the visits where someone clicked. A large share of AI-influenced buyers never click at all, so referral numbers understate the effect, usually badly.
What is a prompt panel and why does it matter?
A prompt panel is a fixed list of the questions your buyers actually ask, re-run on a schedule, with a record of whether you were named each time. It is the closest equivalent to rank tracking that exists for answer engines.
- Write questions the way a buyer would type them, not the way you would
- Keep the wording frozen, because changing it resets the trend
- Record a rate, such as named in six of ten runs, rather than a yes or no
- Re-run on a consistent cadence and never read a trend off one or two runs
- Track competitor mentions in the same answers for context
What indirect signals should you watch?
Watch branded search volume, direct traffic, and the quality of inbound enquiries. When AI visibility improves, the usual first sign is that more people arrive already knowing your name and what you do, having never clicked a link from you.
That is also why attribution feels worse than reality here. The influence happened in a conversation you cannot see.
Should you ask leads where they heard about you?
Yes, and make it a free text field rather than a dropdown. Dropdowns constrain people to the channels you already believed in, and the answers naming an assistant are exactly the ones a prewritten list will miss.
How long before any of this moves?
Longer than paid ads and shorter than traditional rankings once the foundations are in place. Assistants have to encounter and trust corroborating sources, which is a matter of months rather than days.
Judge it on a quarterly view. Reading a prompt panel week to week produces noise and bad decisions.
What is the honest limitation?
You cannot get a complete, deterministic picture, and anyone selling you one is overstating what is possible. The realistic goal is a directional, consistent measurement that tells you whether you are becoming more nameable over time. See also whether SEO still matters.
Engage Ads builds paid advertising, AI search visibility and lead follow-up as one system for high-ticket service businesses. If you want to see where yours is leaking, start with a no-cost growth audit.
Related: the AI Visibility Scan, 0 to 100.
Once you can see whether AI search is sending anyone, the next question is how to get named more often. That is covered step by step in how to get your business recommended by ChatGPT, Gemini and other AI assistants.