The Missing Piece:
AI Traffic → Revenue

12 AI engines tracked

The pixel recognizes traffic from every major AI platform: ChatGPT (including chat.openai.com), Perplexity, Gemini (including Bard), Copilot (including Bing Chat), Grok (including X/Twitter integration), You.com, Meta AI, Poe, Phind, Mistral (including Le Chat), and more. Each engine is tracked separately with its own visit counts, conversion rates, and revenue attribution.

How does the pixel detect AI visitors?

Not all AI traffic carries a referrer header. ChatGPT added utm_source=chatgpt.com to outbound links in June 2025, but many engines still strip referrers entirely. The pixel uses two approaches:

• HTTP Referrer — primary method. Matches the document.referrer against 19 known AI domain patterns. Fastest and most reliable when available
• UTM parameters — fallback. Scans utm_source in the URL query string against both full domain names and short aliases (e.g., “chatgpt”, “perplexity”, “gemini”). Catches traffic that other analytics tools miss entirely

If neither method detects an AI source, the script exits immediately without sending any data. No tracking of non-AI visitors.

What does the pixel track?

The pixel tracks more than just pageviews. Each AI visitor gets a unique session ID and a visitor ID cookie (365-day expiry). You see total unique sessions, average pages per session, and whether each visitor is new or returning. A 30-day attribution cookie stores which AI engine brought the visitor, their landing page, and the timestamp — so when they convert days later, the revenue is attributed to the original AI source.

SPA support

Single-page applications don’t trigger new page loads. The pixel intercepts history.pushState, replaceState, and popstate events to track client-side navigation within the same session. Every page the AI visitor views is recorded, even on React, Vue, or Angular sites where navigation happens without a full page reload.

How do I measure ROI from AI traffic?

Revenue attribution goes beyond counting conversions. The dashboard shows a full funnel: AI Visits → Conversions → Revenue. For each AI engine, you see:

• Conversion rate — what percentage of visitors from this engine convert
• Revenue per engine — actual dollar amount attributed to each AI source
• Average days to convert — how long between the first AI visit and the conversion event
• Conversions by event type — which actions (purchase, signup, lead, contact, download) each engine drives
• Top converting pages — which landing pages produce the most revenue from AI traffic
• ROI calculation — your AI-attributed revenue minus your ARPAI platform spend, expressed as ROI percentage

Geographic and device breakdowns

Every AI visit is enriched with GeoIP data (country and city) and device detection (desktop, mobile, tablet). The dashboard shows traffic breakdown by country (top 10), by language (top 10), by device type, and by landing page (top 10). You see which countries send the most AI traffic, what devices those visitors use, and which pages they land on.

Pixel status indicator

The dashboard shows a live status of your pixel: green (active — data received within 24 hours), yellow (waiting for AI traffic — data within 7 days), or red (no recent data). You always know whether the pixel is working and whether AI engines are currently sending traffic to your site.