AI Tracking: How to Detect AI Bots and Referrals on Your Website

AI answer engines — ChatGPT, Perplexity, Claude, Gemini — send bots that crawl your website to feed their answers. But do you know which bots visit your site, which pages they index, and how much traffic these AIs generate for you?

That's exactly what AI tracking provides: a detection layer installed on your server that automatically identifies AI bots and measures referrals from conversational AI platforms.

Why Track AI Bots?

1. Classic Analytics Are Blind

Google Analytics, Matomo and traditional analytics tools work via browser-side JavaScript. However, AI bots like GPTBot or ClaudeBot don't execute JavaScript. They make direct HTTP requests, completely invisible to your analytics.

Result: you don't know if AIs are indexing your content. It's like having a store without security cameras — important visitors pass through without you knowing.

2. AI Referral Traffic Is Growing Fast

According to recent studies, traffic from AI answer engines grows 30 to 50% per quarter. When a user asks Perplexity a question and clicks on a cited source, that's qualified traffic with a conversion rate often higher than classic SEO.

3. Optimize Your Content Strategy

By knowing which pages are crawled by AI bots, you can:

• Identify content that interests LLMs
• Adapt your schema markup to improve understanding
• Check that your robots.txt isn't blocking bots you want to allow
• Measure the impact of your AEO optimizations

How Does AI Tracking Work?

User-Agent Detection

Each AI bot sends a specific User-Agent in its HTTP requests. The tracker maintains a database of known AI bot signatures and compares them in real-time:

AI Referral Detection

When a user visits your site after seeing an AI response, the browser sends an HTTP Referer header. The tracker analyzes this referrer to identify AI platform domains:

• chat.openai.com → ChatGPT
• perplexity.ai → Perplexity
• claude.ai → Claude
• gemini.google.com → Gemini

HMAC Signature for Security

Tracking events are signed with an HMAC secret to ensure data isn't tampered with. Each payload is hashed with SHA-256 and the signature is verified server-side before acceptance.

The 4 Required Credentials

To set up AI tracking on your site, you need 4 credentials obtained from the API & Integrations page of your AI Labs Audit account:

1. API Key (trk_live_...) — Unique key to authenticate tracking requests
2. HMAC Secret — Secret key to sign payloads (never expose client-side)
3. Client ID — Identifier linking data to your client account
4. API URL — API endpoint: https://ailabsaudit.com/api/v1

6 Installation Methods

The AI Labs Audit tracker adapts to all technical environments. See our complete installation tutorial for detailed instructions:

1. WordPress Plugin — Easiest method. Download, activate, enter credentials.
2. PHP — Middleware for Laravel, Symfony or custom PHP sites.
3. Node.js — Async middleware for Express, Fastify and any Node framework.
4. Python — Native integration with Django and Flask.
5. Cloudflare Worker — Edge-level detection, no code changes needed.
6. JavaScript Pixel — Fallback for referral detection when server access isn't available.

What You Can Measure

Once tracking is installed, your metrics dashboard shows:

• Crawls per bot — Pages crawled by each AI bot with trends
• Most crawled pages — URLs that interest AIs the most
• Referrals by source — Incoming traffic from ChatGPT, Perplexity, Claude, etc.
• Per-page analysis — For each URL: which bots visited, when, and how often
• Trends — Crawl and referral evolution over time

GDPR Compliance

The AI Labs Audit tracker is designed to be 100% GDPR compliant: no personal data, no cookies, no IP tracking, no fingerprinting. Only AI bot User-Agents and referrer domains are analyzed.

Get Started Now

1. Go to API & Integrations
2. Create a Tracking IA key
3. Follow the installation tutorial
4. Check the first events on your Tracking dashboard