How to Measure AI Search Visibility When Your Analytics Say Zero

You measure AI search visibility by combining several signals, not trusting one number. Track the AI referrals your analytics can see, then ask every new customer how they found you. Most AI influence is invisible in analytics, so one intake question often tells you more than any dashboard.

Why does my analytics show zero leads from AI search?

Because most AI influence never shows up as a click. A homeowner reads an answer in ChatGPT or Perplexity, closes the app, then calls you or searches your company name days later. Your analytics logs that as direct or organic traffic. The AI did the work and got no credit.

Louise Linehan at Ahrefs put it plainly in her piece, How to Measure AI Search Visibility When Attribution Falls Short: "Most AI influence happens out of sight." Her example is a buyer who reads about a company in Gemini on Monday, Googles the brand the following week, and shows up in analytics as organic search.

Think about how that plays out for a pest-control company. A homeowner finds ants in the kitchen and asks an AI assistant who to call. It names three companies. She doesn't click anything. She Googles your name that evening, fills out your form, and your CRM says "Google organic."

If you only trust the dashboard, you will decide AI search does nothing for you. That's the wrong conclusion, and it's an expensive one.

What is the simplest way to track AI leads?

Add one question everywhere you capture a lead: "How did you hear about us?" Put it on your contact form, your booking page and your phone script. Record the answer in your CRM word for word. It costs nothing, takes ten seconds, and catches leads your analytics cannot see.

Ahrefs points to the same fix. Their advice is to skip stitching together imperfect datasets and just ask customers. Self-reported attribution isn't perfect, but it's about as close to the truth as attribution gets.

Here's how to set it up:

Add a "How did you hear about us?" field to your website form. Make it a free-text box, not a dropdown.
Train whoever answers the phone to ask it on every new-customer call, and to type the answer in.
Add the question to your estimate or booking confirmation, for the leads who never used the form.
Once a month, read the answers and count the ones that mention ChatGPT, Perplexity, Gemini, Claude, Copilot or "AI" in general.
Add a separate tally for "Google" and "I searched your name." Some of those are AI leads in disguise.

Free text matters. A dropdown with "Google" and "Facebook" forces people to pick the closest thing, and AI gets lost. When people write their own answer, you will see things like "asked ChatGPT for a landscaper who does drainage." Nothing else in your data will tell you that.

How big can this be? Ahrefs reports that more than one in five people who sign up for their product say they found them through ChatGPT or Claude. That puts AI ahead of YouTube as a source, and second only to Google. They also say that relying on referral traffic alone would have badly undersold it. Ahrefs sells software, not lawn care, so your number will differ. But the lesson holds: the real figure is bigger than the tracked one.

What AI traffic can my analytics actually measure?

Your analytics can measure sessions that arrive from AI assistants as referral traffic, plus any form fills those sessions produce. That is the visible slice of AI attribution. It's small, but it shows which assistants matter and which pages they send people to. Track it, just don't stop there.

Ahrefs has a guide on setting this up in GA4, and their own free analytics tool has an "AI search" filter. Whatever tool you use, you want to see three things: how many visitors come from AI, which platforms send them, and which pages they land on.

For a sense of how lopsided it can be, Ahrefs shared numbers from wordcount.com, a site they own. ChatGPT made up 92.4% of AI search visitors, Perplexity 4.6%, Gemini 2.7% and Copilot 0.2%. That's one site, not your market. But it shows why it's worth checking which assistant is actually sending you people before you chase all of them.

Conversions are the second piece. Ahrefs found that 6% of their AI search visitors sign up for a free account, which they say is 4X their organic search conversion rate. They are careful to note a free signup isn't worth anything on its own, so they look at how many become paying customers.

You can copy that thinking. Count the form fills and calls that come from AI visitors, then check how many turn into booked jobs. Multiply by your average job value. That gives you a tracked revenue figure, and it's the floor, not the ceiling.

How do I use direct traffic and brand searches as clues?

When AI recommends you, people often search your company name or type your web address afterward. So rising direct visits and rising branded searches can be signs that AI is working, even when no AI click is recorded. They are circumstantial evidence, not proof, but they are worth watching.

Ahrefs lists both as signals in its guide: watching direct traffic for "dark funnel" insights, and measuring the brand demand AI search creates. The idea is simple. If assistants keep naming you, more people will go looking for you by name.

Here's what to watch without turning it into a research project:

Direct visits to your website. Note the monthly trend, not the daily noise.
Searches for your company name. Your Google Business Profile performance data shows what people typed before finding you.
Calls and direction requests from your profile. If these climb while your ad spend stays flat, something is sending people your way.

None of these says "ChatGPT sent this one." Put them next to your intake-question answers. If both go up together, you have a reasonable story. If the intake answers mention AI and your branded searches are climbing, you have a better one.

How do I know if AI assistants actually recommend my business?

Ask them. Type the questions your customers type into ChatGPT, Perplexity and Gemini, such as "best pest control company in [your town]," and write down who gets named. Do it monthly, same questions each time. If you're never named, that is a visibility problem, not a measurement problem.

This is the part most owners skip, because it feels too simple. But it's a direct test. Your analytics can't tell you whether an assistant names you. Reading the answer can.

Run it like this:

Pick five to ten questions a homeowner would ask, including your service and your city.
Ask each one in two or three assistants.
Record which companies are named and whether you are one of them.
Repeat next month and compare.

AI answers vary from one run to the next, so look at the pattern across several questions, not one lucky or unlucky result. If you'd rather not do it by hand, our AI Visibility Report does the checking for you. And your free Local Visibility Score shows how you look in local search, which is where many of those AI answers get their information.

Ahrefs also suggests reporting web mentions as an early sign of AI visibility growth. That makes sense for local businesses. Assistants can only recommend what they find. A complete Google Business Profile, consistent listings and real reviews give them something to find.

What should I report each month?

Report a short list: AI referral visits, conversions from those visits, intake-question mentions of AI, branded searches, and whether assistants name you. Five lines is enough. The point is a trend you can read in two minutes, so you know whether to put more effort into it.

A simple monthly scorecard looks like this:

AI referral visits and which assistants sent them.
Leads and booked jobs from those visits.
Leads who said AI sent them, from your intake question.
Branded searches and direct visits, as a trend.
Assistant check results: how many of your test questions named you.

Keep the same format every month. Don't expect clean numbers. Expect a direction. If the intake answers mentioning AI grow from a couple a month to several, that's worth knowing, and it's a fair reason to invest in the profile and listings that feed those answers. If you want help with that side of it, that's what our local search service does.

The honest summary: you will never get perfect attribution from AI search. You can get good enough evidence, and it starts with asking people how they found you.

Common questions

Can I track ChatGPT leads in Google Analytics?

Partly. Visits where someone clicks a link from an AI assistant can show up as referral traffic if you set up a report for it. But many people read the answer and search your name later, so they appear as organic or direct. Pair your analytics with a "How did you hear about us?" question.

What is the best question to ask new customers?

"How did you hear about us?" in a free-text box. Skip dropdowns, because they push people into the nearest option and hide AI mentions. On phone calls, ask it the same way every time and have the answer typed into your CRM exactly as the customer said it.

Is self-reported attribution accurate enough to trust?

It isn't perfect, since people forget or simplify. Ahrefs notes that, but also says hearing directly from customers is about as close to the truth as attribution gets. Treat it as a trend, not an exact count, and compare it against your analytics and branded search numbers.

How often should I check if AI assistants mention my business?

Once a month is enough for most home-services owners. Use the same set of questions each time, with your service and city in them, and write down who gets named. AI answers change between runs, so look at the pattern across questions, not a single result.