How to Track Your Brand's AI Citations in 2026 (ChatGPT, Perplexity, Gemini)

The Dark Traffic Problem

In 2022, "dark traffic" — website visits with no referrer data — was a minor analytics nuisance. In 2026, it's a measurement crisis. SparkToro and Datos research found that 70.6% of AI referral traffic arrives as Direct in Google Analytics 4. When someone asks ChatGPT about your brand, gets a response that mentions your site, and clicks through — GA4 sees that as Direct traffic. No source. No medium. No way to know it came from an AI.

The problem compounds because different AI platforms handle referrers differently. Perplexity passes referrer data fairly consistently (you'll see perplexity.ai in your referrals). ChatGPT links from within the app strip the referrer entirely. Gemini varies by surface. Claude usually strips it. The result is a fragmented picture that makes AI-sourced traffic nearly invisible in standard analytics setups.

Setting Up GA4 Custom Channel Groups for AI Traffic

The first step is capturing every referral you cansee. GA4's default channel grouping lumps many AI referrers into "Unassigned." A custom channel group fixes that.

In GA4, go to Admin → Data Settings → Channel Groups → Create new channel group. Add a channel called "AI Search" and define it with the following source/medium conditions:

Source matches regex:
perplexity\.ai|chat\.openai\.com|chatgpt\.com|
gemini\.google\.com|bard\.google\.com|
copilot\.microsoft\.com|bing\.com\/chat|
claude\.ai|you\.com|phind\.com|
poe\.com|character\.ai

Once configured, you'll see an "AI Search" channel in your acquisition reports. It won't capture the dark traffic, but it will correctly attribute the clicks that do pass referrer data — primarily Perplexity and some Copilot traffic.

UTM Parameters: Your Best Defense Against Dark Traffic

For traffic you control — links in your own AI-visible content — UTM parameters let you tag and trace even when referrers are stripped. The key is being consistent.

Platform-Specific Tracking Approaches

Perplexity

Perplexity is the most transparent AI platform for tracking. Their crawler (PerplexityBot) hits your pages regularly, and referral traffic from perplexity.ai shows up in GA4 with referrer intact. Monitor your Referrals report for perplexity.aiand create a dedicated segment for it. You can also sign up for Perplexity's Publisher Program, which gives you verified citation data and some revenue sharing on clicks.

ChatGPT

ChatGPT is harder. Links within ChatGPT conversations strip referrers, so most ChatGPT-driven traffic appears as Direct. The most reliable signal is a spike in branded search + direct traffic together after a period when your brand appears in popular ChatGPT outputs. You can also use tools like Profound or Semrush's AI product to directly query ChatGPT with your brand-relevant questions and see whether you're cited.

Google Gemini / AI Overviews

Google Search Console now labels some AI Overview traffic separately in its reports. Monitor your GSC performance data filtered by "AI Overviews" appearance type. For Gemini.google.com as a direct interface, it functions similarly to other AI chat — referrer data is inconsistent. The bigger opportunity is Google's AI Overviews in search results, which you can track through GSC.

AI Share of Voice: The Metric That Actually Matters

Clicks and referrals only tell you about the users who clicked. AI SOV tells you how often you appear in AI-generated answers regardless of click behavior. This is the metric that correlates most strongly with brand consideration and long-term organic traffic growth.

Track AI SOV weekly against competitors. If your SOV drops across all platforms simultaneously, it usually means a competitor has published stronger, more-cited content in that topic area. If it drops on one platform only, there may be a technical issue (e.g., your robots.txt blocking that platform's crawler).

Dedicated AI Monitoring Tools

Manual tracking is a start, but dedicated tools automate the heavy lifting. Here are the main options as of 2026:

Building Your AI Citation Tracking Dashboard

Whether you use a dedicated tool or build your own, your tracking dashboard should answer five questions every week:

Benchmarks: What Good Actually Looks Like

Without context, AI SOV numbers are meaningless. Here are rough benchmarks based on aggregated data from B2B SaaS and marketing companies as of mid-2026:

Most brands starting AI tracking today will find themselves in the 0–5% range. The brands that started GEO work in 2024–2025 are now in the 20–40% range. The compounding effect is real — the longer you've been building authoritative, citable content, the higher your AI SOV climbs.

Quick-Start: Your First AI Tracking Week