In short
Connect three sources: a separate AI channel in GA4 that collects referral traffic from chatgpt.com, perplexity.ai, gemini.google.com, claude.ai and copilot.microsoft.com, a “How did you hear about us?” field on your quote request form, and the source field in your CRM. None of them is complete on its own, because many buyers arrive after an AI recommendation by searching for your company name or coming in directly, but together they show a reliable trend.
If you put effort into AI visibility, sooner or later the question comes: does it bring in leads? To answer it, it is not enough to look at how often AI mentions you. You also need to see how many prospects actually arrive from an AI recommendation, and how much business they turn into.
This article walks you through the three settings you need for measurement: the GA4 AI channel, the form field and the CRM source.
Why is it hard to measure leads coming from AI?
Because after an AI recommendation, the buyer often does not click a link. Instead they search for your company name later, or type in your web address directly, so the visit does not show up as AI traffic.
Three typical situations:
- A click from the AI answer: in this case GA4 usually sees the referring site, for example chatgpt.com. This is the easy-to-measure part, although for clicks coming from some apps the referrer data can be lost.
- A brand-name search or direct visit: the buyer remembers your name and comes back later through Google or directly. This looks like organic or direct traffic.
- Google AI Overviews: GA4 currently tends to show clicks from here within regular Google organic traffic, without separating them.
That is why measurement rests on three sources: GA4, the quote request form and the CRM. Each one is incomplete on its own, but together they give a good estimate.
How do you set up an AI channel in GA4?
In a custom channel group, create an “AI assistants” channel that uses a regular expression to collect traffic from AI platforms, and place it above the Referral channel.
The steps:
- In GA4, open Admin, and under the data display settings find Channel groups.
- Copy the default channel group, so the other channels stay unchanged.
- Add a new channel named “AI assistants”, with the condition that Source matches the regular expression below.
- Move the channel above the Referral channel. GA4 evaluates channels in order and assigns the visit to the first one that matches, so if it stays at the bottom, this traffic keeps showing up as referral.
- Save, and select the new channel group in your acquisition reports.
Regular expression for the source: chatgpt\.com|chat\.openai\.com|perplexity\.ai|gemini\.google\.com|claude\.ai|copilot\.microsoft\.com
When a new AI platform appears among your referring sites, add it.
If you already track quote request form submissions as a key event, from now on you will also see how many quote requests came from the AI channel. If you don't track them yet, start there, because without it you are only counting visitors, not prospects.
What should you ask on the quote request form?
Add a “How did you hear about us?” field to the form, with “AI assistant (ChatGPT, Gemini, Perplexity, Claude, Copilot)” as a separate option.
This field catches the people AI recommended you to who later arrived by another route. A few tips:
- A dropdown, not just free text: that way the answers can be aggregated. Add an “Other” option with a comment field.
- Make AI a separate option: if you only offer “Internet search”, AI recommendations get lost inside it.
- Keep the form short: one extra field rarely puts a prospect off, but a long questionnaire does.
The answers are not perfect, because people don't always remember exactly where they heard of you. But the trend is clearly visible month by month.
How do you record the AI source in your CRM?
Your CRM should have a source field with “AI recommendation” as one of its values, filled in for every new prospect by whoever runs the first conversation.
The form answer can also flow into the CRM automatically, but the first conversation is the best moment to confirm it. One simple question is enough: “How did you find us?” If the answer is that they asked an AI, it is worth noting what question they asked. Over time this shows which buyer questions bring in real business.
The big advantage of the CRM source field is that it ties not only the inquiry but also the later proposal and contract to the AI source. So you see not only how many prospects came in, but also how much business came out of them.
Curious who AI recommends in your industry right now?
After a short conversation, we build your industry's AI ranking free of charge: who AI recommends for typical buyer questions, and where you rank against your competitors.
Get my industry's AI ranking →How do you connect the three data sources?
Each month, keep a simple table with three columns side by side: quote requests from the GA4 AI channel, the number of people who chose AI on the form, and the opportunities recorded as AI-sourced in the CRM.
The three numbers will never be identical, and they don't need to be. What matters is the trend: if all three move together, that is a real change. It is worth adding the AI share, the appearance rate and the recommendation position from your AI answer measurements as well. If your position in AI answers improves and the three lead numbers then grow too, that is a strong sign the work is reaching your buyers.
In our AI visibility service, we set up this measurement at the very start of the work, so that the follow-up measurement shows not just mentions but leads as well.
What not to expect from the measurement
An exact, one-to-one matching number. Measuring leads from AI recommendations is an estimate: good for following the trend, not for bookkeeping.
Change in AI answers also takes time: the content and reputation effect typically shows up within 2-6 months. That is why it is worth setting up measurement now, even if you are not yet working on AI visibility deliberately: you will have a baseline number to compare against later.
If you also want to know what AI says about you right now, request the free AI visibility assessment. And if the terms are still unclear, we summarized them in our article on the differences between GEO, AEO and AI SEO.
Frequently asked questions
Can GA4 see traffic coming from ChatGPT?
Partly. If the visitor clicks a link in the AI answer, GA4 usually records it as a referral source, for example as chatgpt.com. But if they later search for your company name or come in directly, the role of AI is not visible in GA4.
Why does the AI channel have to sit above the Referral channel?
Because GA4 evaluates the channel group rules in order and assigns the visit to the first channel that matches. If the AI channel is at the bottom, these visits keep showing up as referral traffic.
Is it enough to watch GA4 alone?
No. GA4 only sees direct clicks. The “How did you hear about us?” field and the CRM source field catch the people who arrived by another route after an AI recommendation.
Can you measure traffic from Google AI Overviews?
Not easily as a separate channel at the moment, because in GA4 these clicks typically appear within regular Google organic traffic. The form field and the CRM source help fill that gap.