AI chatbots are sending traffic to your website right now — but GA4 shows it as "direct." Here's how to identify, attribute, and measure revenue from ChatGPT, Perplexity, and AI Overviews.
Check Your AI Visibility Score →Something has been quietly happening to your analytics: a growing chunk of visitors is arriving at your site from ChatGPT, Perplexity, Google's AI Overviews, and Copilot — and GA4 is misclassifying most of it as direct traffic.
This isn't a small rounding error. ChatGPT alone accounts for more than 90% of all LLM-referred web traffic, according to NeilPatel's 2026 analysis. Google's AI Mode rolled out globally in March 2026 across 200+ countries. The traffic is real — but marketers can't see it.
The consequence: teams are making budget decisions based on SEO and paid channels while ignoring a channel that's growing faster than anything they've seen since social media emerged.
Before you can attribute revenue, you need to know you're even in the game — that you're actually appearing in AI responses when someone asks a question in your category. That's where an AI visibility score becomes your starting point.
LLM-driven visitors have a distinctive behavioral signature in your analytics. Researchers at SearchAtlas, alhena.ai, and NeilPatel have documented these patterns — once you know what to look for, they're surprisingly consistent.
AI chatbot users have already decided they trust your content before clicking. They arrive with intent — session depth is often 1 page, but time on page is high.
Instead of landing on your homepage, LLM visitors land deep on specific informational pages — exactly the ones that answer the question the AI cited you for.
Some AI crawlers and embedded browsers execute minimal JS. You may see visitors with unusual browser fingerprints or lower-than-average client-side event counts.
If your brand gets cited in a popular ChatGPT conversation, direct traffic spikes with no corresponding increase in paid, email, or organic search channels.
Create a GA4 exploration segment using these conditions to isolate likely LLM-attributed sessions:
Here's the practical setup to start capturing and attributing LLM traffic in your analytics stack.
Where you have control (llms.txt, schema markup, landing pages you submit to directories), add UTM parameters: ?utm_source=chatgpt&utm_medium=ai_referral&utm_campaign=ai_visibility. When ChatGPT surfaces a page with this URL in context, some users will click the exact URL — and GA4 will capture it correctly.
Go to Admin → Data Display → Channel Groups → Create custom group. Add rules to catch known AI referrers: source contains "chatgpt", "perplexity", "copilot", "gemini". GA4 is slowly adding native detection but custom groups give you control today.
Create an automated insight alert for when direct traffic to deep informational pages spikes by >30% week-over-week. These spikes are your early warning that you're being cited somewhere in the AI ecosystem.
For purchases, signups, or lead forms: add a hidden field that captures the current UTM values (from sessionStorage). When a user arrives from an AI citation and converts later in the same session, you'll capture the full attribution chain even if they browsed multiple pages.
Use a tool like QuicklyTools AEO Checker to track when your brand appears in AI responses. When visibility scores improve or you get cited for new topics, check GA4 for corresponding direct traffic spikes 24-72 hours later.
Revenue attribution from AI traffic follows the same logic as any dark-funnel attribution — you're looking for correlation, not direct click-to-sale tracking in most cases.
Given the limitations of referrer tracking from AI sources, the most reliable attribution approach is a correlation + time-window model:
Perfect attribution from AI sources isn't achievable today — the same way social dark traffic from WhatsApp and Slack groups can't be perfectly attributed. What you can measure:
The teams winning in AI search in 2026 are treating it like early SEO in 2010: imperfect data, high-direction signals, directionally correct investments. Don't wait for perfect attribution before optimizing your AI visibility.
Attribution only matters if you're actually appearing in AI results. Here are the key tools in the stack.
The fastest way to understand your AI search visibility. Score 0-100 across Technical AEO, Content AEO, and AI Visibility. Instant results, no setup, free to use. If your score is low, you're not appearing in AI results — attribution becomes irrelevant.
Check your AEO score free → quicklytools.dev/tools/aeo-checker/Your primary data source for attribution. The segments and custom channel groups described in this guide are built directly in GA4. Free for most teams.
While GSC focuses on traditional search, AI Overviews are increasingly linked from search results. Monitor impressions and clicks for your AI-optimized pages as a secondary signal.
Not a tracking tool, but foundational. Adding llms.txt and schema markup to your site helps AI systems properly attribute your content — which makes UTM tracking via AI citations more viable.
You can't attribute traffic from AI results you're not appearing in. Check your AEO score in 60 seconds — free, no account required.
Run Your Free AEO Check →