Best tools to measure traffic and conversions coming from ChatGPT, Perplexity and Gemini
How we measure AI referral traffic and conversions from ChatGPT, Perplexity and Gemini for clients, from a GA4 baseline to page-level joins with crawls and citations.
When a client asks whether ChatGPT, Perplexity or Gemini sends them anything worth having, we answer with two layers of measurement. Most teams build only the first.
Layer one is web analytics. Google Analytics 4, or whatever the client already uses, can show visits with an AI product as the referrer and attribute conversions to them. It is free, it is installed, and it is the baseline.
Layer two is AI visibility measurement. Which prompts mention the brand, which pages get cited, whether the bots fetch those pages, and how all of that connects to visits. Analytics cannot see any of it, and it cannot see the buyer who read an answer, remembered the name and typed the URL later.
Layer one: what we set up in analytics
We create a segment or custom channel group that matches the AI referral sources in the client's own data. Domains such as chatgpt.com, perplexity.ai and gemini.google.com are the usual candidates, but we check the source report first rather than trusting a list from a blog post, because the values change. We attach the client's real conversion events (demo requests, trial starts, purchases, qualified calls) and record a baseline before anything else changes.
We also tell clients what it cannot see:
- Visits with no referrer land in direct traffic, so analytics undercounts.
- Citations that nobody clicks never show up.
- Cause. Analytics says a visit happened. It cannot say which prompt, cited page or crawl led to it.
UTM tags rarely help, because we do not control how an assistant writes the link.
Layer two: the tool we use
We run client programs on Promptwatch, and its visitor analytics is the part that fills the gap analytics leaves. It tracks AI-referred traffic and conversions from ChatGPT, Perplexity, Gemini, Claude, Copilot, DeepSeek, Kimi, Mistral and more, using a lightweight script, a GTM template or the WordPress plugin. For clients on Webflow or Framer, the script route is the one we use.
Then we join it to the rest:
- Agent Analytics shows IP-verified AI crawler logs, so we can see whether the bots even fetch the page that later receives a visit. GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Perplexity-User and Google-Agent are among the bots tracked. Crawler logs start on the Professional plan.
- 360 Insights joins crawler logs, citation clicks and AI-referred visits per page. This is how we find pages that are crawled and cited but never clicked, or clicked but not converting.
- Search Insights overlays Google Search Console and Bing Webmaster traffic, including Microsoft Copilot, on AI crawls and citations.
- Citation analytics and visibility dashboards (Visibility Score, Share of Voice, Position in Answer) from daily prompt panels across ChatGPT (Search), Google AI Overviews, Google AI Mode, Copilot, Gemini, Perplexity, Amazon Alexa and Meta AI.
- Scheduled PDF reports and the Looker Studio connector to put it in front of finance without a screenshot.
On cost: Essential is $95 a month with 200K visitor events, Professional is $245 with 1M visitor events and 25M crawler logs, and Business is $579 with 10M visitor events. We check the visitor-event volume against the client's traffic during scoping.
What other tools do here
We compare honestly, because some clients arrive with a tool already in place. From the vendors' public information in our catalog:
- AthenaHQ lists GA4, Shopify and Webflow revenue attribution. Useful for Shopify stores, though we did not find crawler logs listed.
- LLM Pulse includes AI traffic analytics from its Growth tier.
- Scrunch AI reports AI agent and bot traffic, which is the crawler side and not the human-visit side.
- Rankscale and Otterly.AI are monitoring tools without revenue or traffic attribution, according to our catalog.
- Profound is an enterprise marketing-agent platform, and we did not find visitor attribution documented.
What we report
A monthly view with four numbers on one page: visibility trend, crawler fetches on key pages, AI-referred visits, and conversions from those visits. When the first three move and the fourth does not, we go to the landing pages. When the fourth moves without the first, we look for unattributed influence.
Two published customer cases show why we measure conversions and not just mentions. Crisp reports that AI-referred traffic converted at twice the rate of traditional channels. TreeHouse reports that AI search became about 25% of its organic conversions, 99% of it from ChatGPT, after fixing crawler access and automating weekly reporting. Those are the customers' own results. We use them to explain why the measurement is worth setting up, not as a forecast for any other client.
Our honest caveat
AI-referred conversions are a floor. Some visits have no referrer, and some influence never produces a click. We tell clients to read the number as the minimum, and to trust the trend more than the absolute value.
Where to start
Build the GA4 segment this week and attach your conversion events. Then add a dedicated visitor script, so you are not relying on referrers alone. Do both before you change any content, so you have a baseline to compare against.