LLM Referral Tracking

What is LLM referral tracking?

LLM referral tracking identifies human visitors who arrive at your site after clicking a link in an AI platform's response. This is fundamentally different from bot tracking, which monitors AI crawlers. Referral tracking captures actual people: potential customers who asked an AI assistant a question, received a recommendation that included your business, and then clicked through to your site.

Traditional analytics tools like Google Analytics do not break out AI platform referrals as a distinct traffic source. CitedPro fills that gap.

Tracked platforms

CitedPro detects referrals from 20+ AI platforms, including:

Additional platforms detected include Poe, Mistral, Pi, Cohere, Character.ai, and Groq. The platform database is automatically updated from api.srworks.co with a daily cache, and falls back to a bundled list if the API is unavailable.

How detection works

CitedPro uses a multi-layered approach to identify AI platform referrals:

Server-side detection

On every page load, CitedPro checks the HTTP referrer header and the utm_source query parameter against its database of known AI platform domains. If either matches a known platform, the visit is recorded immediately on the server side.

Client-side fallback

CitedPro also loads a lightweight JavaScript file (cited.js) that checks document.referrer and URL parameters in the browser. If a match is found, it sends an AJAX request back to WordPress to record the visit. This catches referrals that may be missed server-side due to caching or CDN configurations. No cookies are used.

Rate limiting and deduplication

To prevent abuse, the client-side tracking endpoint is rate-limited to 30 requests per 60 seconds per IP address. Additionally, a 60-second transient is set per IP+URL combination to prevent duplicate recordings of the same visit.

Data storage

LLM referral visits are stored in the same wp_cited_bot_visits table as bot tracking data, distinguished by the is_human=1 flag. This unified storage makes it easy to compare bot activity and human referrals side by side in the Analytics tab.

Each referral record includes the source platform, landing page URL, timestamp, and referrer string.

Viewing referral data

1. Go to CitedPro → Analytics
2. Look for the LLM Referrals section
3. View daily trends with platform attribution
4. See which pages receive the most AI referral traffic

The Dashboard also shows a quick summary: total LLM referral count and the top referring platform.

Why this matters

Traffic from AI platforms is a new and growing source of visitors that most analytics tools do not track automatically. When someone asks ChatGPT for a recommendation and clicks through to your site, that visit typically shows up as "direct" or "referral" in Google Analytics with no indication it came from an AI assistant.

CitedPro makes this traffic visible so you can:

• Measure AI search ROI: Quantify how much traffic AI assistants actually send to your site
• Track platform trends: See which AI platforms mention you most and where your visibility is growing
• Understand visitor intent: Users arriving from AI platforms have already received a recommendation and are actively researching, making them high-intent visitors
• Justify optimization efforts: Use concrete referral numbers to demonstrate the value of AI search optimization

Limitations

• Direct links: If a user copies a URL from an AI response and pastes it into their browser, there is no referrer header to detect
• Privacy browsers: Some browsers and privacy extensions strip referrer headers entirely
• Mobile apps: AI platform mobile apps may not pass referrer data when opening links in external browsers
• Full-page caching: If your caching setup serves fully static pages, server-side detection may be bypassed (the client-side fallback handles this case)