How to Tell Whether AI Search Is Sending You Customers

Your boss, your board or your business partner has started asking the same question every month: "Are we showing up in AI?" You open Google Analytics, squint, and say something confident-sounding.

Here is the honest answer. You can measure more than the people selling "AI visibility scores" let on, and less than you would like. No single dashboard tells you how often ChatGPT or Google's AI Overviews recommend you. But four free sources, read together, tell you whether AI search is sending you people and whether those people turn into customers. That last part is the one your boss cares about.

What does Google Search Console show for AI Overviews and AI Mode?

Google says sites appearing in AI features like AI Overviews and AI Mode are included in the overall search traffic and reported in the Performance report, within the "Web" search type. So it is already in your numbers. You just cannot pull it out as its own line.

Google's Performance report help page spells out how it is counted:

• Clicks: clicking a link to an external page in an AI Overview or in AI Mode counts as a click.
• Impressions: standard impression rules apply.
• Position: an AI Overview occupies a single position in the results, and every link inside it shares that position. Position in AI Mode follows the same method as a normal results page.
• Follow-ups: a follow-up question in AI Mode is treated as a new query.

In June 2026 Google added Search Generative AI performance reports to Search Console. They show impressions (how often your URLs appeared in generative AI features in Search and Discover), broken down by pages, countries, devices and dates. Google's announcement lists no click metric, and the same data stays in the overall Performance report. A note on that post says the reports reached all websites worldwide as of August 31, 2026.

What Search Console still will not tell you

• How many clicks came from AI features specifically, as opposed to regular blue links.
• Which queries triggered your AI appearances. Google's announcement of the new report doesn't describe a query breakdown.
• Anything at all about ChatGPT, Perplexity, Claude, Copilot or Gemini outside Google Search.

What to do with it

1. Open the generative AI report and note which pages earn AI impressions.
2. Compare that list against your money pages (services, quote requests, contact). If AI impressions all land on a 2019 blog post, fix that.
3. Record the monthly impression total in your reporting sheet. The trend matters more than the number.

How do you find ChatGPT and Perplexity traffic in GA4?

AI assistant visits mostly arrive as referrals. Here is the step by step, using Google's own Traffic acquisition report:

1. In GA4, open Reports, then select Acquisition, then Traffic acquisition.
2. The table defaults to Session default channel group. Click the dropdown on the first column header and switch it to Session source / medium.
3. Use the search bar above the table. Type chatgpt and press Enter. You should see rows like chatgpt.com / referral.
4. Repeat for perplexity, copilot, gemini and claude. Write down sessions and key events (form fills, calls, bookings) for each.
5. Set the date range to the last full month and compare it with the month before.
6. Save it as a comparison or exploration so next month takes two minutes.

ChatGPT often adds utm_source=chatgpt.com to the links it shows, which helps it show up cleanly as a source.

Why the GA4 number is a floor, not a total

Some AI visits will never say where they came from. A person might copy your URL out of a chat app, or the app might not pass a referrer at all, and GA4 files the visit under Direct. Others read the AI answer, then search your brand name on Google and arrive as organic traffic. So the AI referral count is the minimum AI search is doing for you, not the whole picture.

Skip comparing yourself with somebody's "average AI referral share." There is no reliable benchmark for a business like yours, and the people quoting one usually have a tool to sell. Compare yourself with yourself, month over month.

How do you check whether AI assistants actually cite you?

Do it by hand, monthly, with the same prompts every time. It is boring. It is also honest, because you see the actual answer a customer would see.

Pick 10 to 15 prompts a real customer would type. Mix three kinds:

• Category prompts: "best commercial roofer in Tampa," "accountant for small restaurants near me."
• Problem prompts: "why is my AC running but not cooling," "how do I switch accountants mid-year."
• Brand prompts: "what is [your company] known for," "is [your company] any good."

Run each prompt in the assistants your customers use. Use a logged-out or fresh session where you can, because personalization changes answers. Then log it in a sheet like this one (the rows are made-up examples):

The "which source" column is where the real insight lives. If assistants keep citing a directory, a review site or a competitor's comparison page instead of you, that tells you exactly where to show up next. If they cite you but quote something outdated, you know which page to fix.

Answers shift by wording, location and day. One month of "no" is noise. Six months of "no" while a competitor gets named every time is a signal.

Why should you ask "how did you hear about us"?

Because the customer knows something your analytics never will.

Add a required "How did you hear about us?" field to your contact form, booking page and intake call script. Include these options alongside the usual ones:

• ChatGPT
• Google AI Overview or AI Mode
• Perplexity, Copilot, Gemini or another AI assistant
• Google search (regular results)
• Google Maps
• Referral from a friend or colleague
• Other (please tell us)

Then tag the answer in your CRM so it follows the lead all the way to a closed deal. A month later you can say "four leads named an AI assistant, and one became a customer" instead of "we got some AI referrals, probably." Self-reported data is messy. It is still the only source that connects AI search to revenue in the customer's own words. (If your CRM and analytics do not talk to each other yet, start with using analytics to drive marketing decisions.)

Why is branded search growth a useful proxy?

A lot of AI influence is invisible. Someone asks ChatGPT for a recommendation, sees your name, and then googles you instead of clicking the link. That visit shows up as a branded search, not an AI referral.

Search Console now makes this easier to watch. Its branded queries filter splits Performance data into branded and non-branded queries, and an Insights card shows the click breakdown between the two. Google says the classification is done by an AI-assisted system that includes your brand name in all languages and typos, warns that some queries may be misidentified, and limits the filter to top level properties with enough query volume.

If branded searches climb while your other marketing stayed flat, something is putting your name in front of people. AI assistants are one plausible cause. It is not proof, so treat it as a supporting signal next to your referral data and form answers, not a replacement for them.

What does a useful monthly AI search report look like?

One page. Five lines:

1. Search Console: AI feature impressions this month versus last, and the top pages earning them.
2. GA4: sessions and key events from each AI assistant referrer.
3. Prompt check: how many of your prompts cited you, in which assistants, and which sources got cited instead.
4. Leads: how many inquiries named an AI assistant, and how many became customers.
5. Branded search: branded clicks and impressions versus last month.

That report will not give your boss a magic number. It gives an honest answer with evidence. And when the numbers point to a gap, the fix is usually ordinary SEO work, because AEO is just SEO done properly.

What should you do next?

Set up the GA4 view and the "how did you hear about us" field this week. Neither needs a new tool. Run your first prompt check before the month ends so you have a baseline.

If you want help finding why AI assistants cite your competitors instead of you, that is part of our SEO audit, and the ongoing work lives under SEO, AEO and GEO. Or book a free call and we will walk through your numbers with you.