How to measure traffic and conversions from AI search
Visits from ChatGPT, Perplexity, and Gemini are easy to undercount in standard analytics. How to track AI-referred sessions and connect them to conversions.
Visibility in AI answers is a leading indicator. Sooner or later someone asks the fair question: what did it do for revenue? To answer it, you need to see visits from AI assistants clearly and follow them through to a conversion. Most analytics setups make that harder than it should be.
Where AI traffic comes from
When an AI answer includes a link and the user clicks it, the visit usually arrives with a referrer such as chatgpt.com, perplexity.ai, gemini.google.com, or copilot.microsoft.com. Those are the easy cases.
The harder ones:
- Missing referrers. Some apps and in-app browsers don't pass a referrer, so the visit lands in direct.
- Influence without a click. A buyer reads a recommendation in ChatGPT, then searches your brand on Google the next day. The AI did the persuading, but branded search gets the credit.
- AI Overviews. Clicks from Google's AI Overviews and AI Mode come through as Google traffic, mixed in with regular organic.
Whatever number you report will undercount AI's real influence. That's fine as long as everyone knows it and you track the measurable part the same way every month.
Step 1: separate AI referrers in analytics
In GA4 or a similar tool, create a custom channel group or segment that matches known AI referrer domains. Keep the list current, since new assistants launch and existing ones change domains. Put the channel right next to organic search in your reports so people actually look at it.
Step 2: track the conversions that matter
Pageviews from AI referrers are nice to have. Signups, demo requests, and purchases are the point. Make sure your conversion events fire on the pages AI visitors land on, which may be deep pages like comparisons, docs, or pricing rather than the homepage.
Step 3: look at landing pages
Group AI-referred sessions by landing page. This shows which of your pages the engines send people to, and the list may not match the pages you'd guess. Pages that get AI traffic but don't convert need a better next step. Pages you want cited that get no AI traffic go back on the content list.
Step 4: watch branded search alongside it
Because a lot of AI influence surfaces later as branded searches, keep an eye on branded impressions in Google Search Console next to your AI visibility trend. Don't overclaim the link between the two. But if both climb after a round of AI SEO work, that's a reasonable signal to bring to a budget meeting.
Step 5: connect it to visibility data
Traffic numbers alone can't tell you why something changed. The useful view lines up which prompts you appear in, which pages get cited, and which pages receive AI-referred visits that convert.
That's how we report for clients, using Promptwatch. Its visitor analytics tracks AI-referred traffic and conversions through a lightweight script or a Google Tag Manager template, in the same workspace as prompt tracking and citation analytics. Because Agent Analytics also logs AI crawler visits, we can follow a single page from the moment a bot fetched it, to its first appearance as a citation, to the visits and signups it produced. It connects to Google Search Console and Looker Studio as well, so the AI numbers can sit in the same dashboard as the rest of the marketing reporting.
Setting expectations
Don't judge an AI SEO program on raw session counts alone. Look at conversion rate and lead quality from AI-referred visits, and remember the measured number is a floor, not a ceiling. If those visitors convert well, that's the argument for investing more, and you'll have the data to make it.