Ai Traffic Tracking
Separate ChatGPT, Perplexity, Claude, and Copilot referrals from blended GA4 referral traffic so you can measure GEO impact and conversion quality.
Overview
AI Traffic Tracking is an agent skill most often used in Grow (also Launch) that guides solo builders to isolate AI search and assistant referral traffic in GA4 and Google Search Console.
Install
npx skills add https://github.com/kostja94/marketing-skills --skill ai-traffic-trackingWhat is this skill?
- Step-by-step GA4 Free form exploration using Session source / medium to isolate AI referrers
- Explains why AI visitors often land as Referral, Organic, or Direct without custom reporting
- Covers Google AI Overviews separately from classic organic so you can judge summary-box traffic
- Maps AI-driven search sources (ChatGPT, Perplexity, Gemini, Claude, Copilot) for comparable session metrics
- Pairs with generative-engine-optimization when you need strategy, not just dashboards
- Recommends GA4 Free form exploration as the primary Option 1 reporting path
- Names six AI-driven search surfaces including ChatGPT, Perplexity, Gemini, Claude, and Copilot
Adoption & trust: 775 installs on skills.sh; 586 GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your GA4 property shows referral or direct spikes but you cannot tell which visits came from ChatGPT, Perplexity, or AI Overviews.
Who is it for?
Founders with live GA4 and Search Console who are investing in GEO and need clean source-level reporting without hiring an analytics consultant.
Skip if: Teams with no site traffic yet, no analytics access, or people who only want content strategy without touching GA4 explorations.
When should I use this skill?
User wants AI traffic in GA4 or GSC, mentions ChatGPT/Perplexity/Claude referrals, AI Overviews, or how to track AI search traffic.
What do I get? / Deliverables
You get repeatable GA4/GSC views that segment AI-driven sessions and engagement so GEO experiments have measurable baselines.
- GA4 exploration or comparable report isolating AI session sources
- Documented filter or dimension approach for AI Overviews versus organic
- Baseline metrics table: sessions, engagement rate, key events by AI source
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Grow because the skill’s job is ongoing measurement in GA4 and Search Console, not drafting GEO copy or distribution tactics. Analytics is the right subphase: custom dimensions, exploration reports, and GSC filters are instrumentation work, not lifecycle email or support ops.
Where it fits
After publishing AI-optimized FAQ pages, you stand up a GA4 exploration to compare assistant referrals against organic search.
Each Monday you refresh session-source cuts to see whether Perplexity or ChatGPT traffic converts on signup better than Google organic.
You launch a waitlist landing page and need to prove early AI-referred visitors engage before scaling paid distribution.
Post-incident you verify whether an outage dip showed up in AI referral channels versus direct return traffic.
How it compares
Measurement workflow for analytics consoles—not a generative-engine-optimization playbook or an MCP server that pulls metrics automatically.
Common Questions / FAQ
Who is ai-traffic-tracking for?
Solo and indie builders running marketing sites or SaaS who use Claude Code or similar agents to configure GA4 and GSC and care about AI referral quality, not just total sessions.
When should I use ai-traffic-tracking?
During Launch when you start GEO campaigns and need baselines; in Grow for weekly analytics reviews; after shipping landing pages when Perplexity or ChatGPT traffic appears in vague referral buckets; before major pricing or positioning changes when you want AI-sourced conversion
Is ai-traffic-tracking safe to install?
Treat it as documentation-style procedural knowledge—review the Security Audits panel on this Prism page and confirm the skill source before letting an agent act on live analytics admin settings.
Workflow Chain
Then invoke: generative engine optimization
SKILL.md
READMESKILL.md - Ai Traffic Tracking
# Analytics: AI Traffic Guides tracking of AI-driven search traffic in Google Analytics 4 and Google Search Console. **When invoking**: On **first use**, if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On **subsequent use** or when the user asks to skip, go directly to the main output. ## Scope - **AI search traffic**: Track in GA4 and GSC; separate AI sources from organic referral - **Google AI Overviews**: AI summary box in Google search (formerly SGE) - **AI-driven search**: Traffic from ChatGPT, Perplexity, Gemini, Claude, Copilot, etc. ## Why Track Separately - AI traffic is growing but GA4 often groups it as Referral, Organic, or Direct - AI visitors may have stronger intent and higher conversion - Separating AI Overviews from organic helps assess AI impact ## GA4: AI-Driven Search ### Option 1: Exploration Report (Recommended) 1. **Explore** →**Free form** 2. **Dimensions**: `Session source` (or `Session source / medium`) 3. **Metrics**: Sessions, Engagement rate, Event count, etc. 4. **Filters**: Add filter →`Session source` **Matches regex** →use regex below 5. Configure table, save report **Regex (common AI sources):** ``` chatgpt\.com|openai\.com|openai|perplexity\.ai|perplexity|doubao\.com|chat\.qwen\.ai|copilot\.microsoft\.com|copilot\.com|(business\.)?gemini\.google|chat\.deepseek\.com|deepseek\.com|poe\.com|anthropic\.com|claude\.ai|bard\.google\.com|edgeservices\.bing\.com ``` ### Option 2: Custom Channel Group 1. **Admin** →**Data Display** →**Channel Groups** 2. Copy default group, name e.g. "Default and AI Chatbots" 3. Add channel "AI Chatbots": `Source` **Matches regex** (same regex) 4. **Important**: Place "AI Chatbots" above "Referral" so it matches first 5. Save and use in Traffic Acquisition ### Option 3: Custom Report 1. **Reports** →**Library** →Create Detail Report 2. Use Traffic Acquisition template 3. Add filter: `Session source` **Matches regex** (same regex) 4. Save and add to menu ## Common AI Source Domains | Platform | GA4 Source examples | |----------|---------------------| | ChatGPT | chatgpt.com, openai | | Perplexity | perplexity.ai, perplexity | | Copilot | copilot.com, copilot.microsoft.com | | Gemini | business.gemini.google, gemini.google | | Claude | claude.ai, anthropic.com | | Bing Chat | edgeservices.bing.com | ## Google AI Overviews - **GA4 + URL fragment**: Some AI Overview clicks add URL fragments; can use GTM (partial coverage) - **GSC**: For AI Overviews analysis in GSC (filter, limitations), see **google-search-console** ## Checklist - [ ] AI sources identified in GA4 (Session Source) - [ ] AI traffic Exploration report created - [ ] Channel group updated with AI above Referral (if used) - [ ] Custom report added to Library (optional) - [ ] GTM + URL fragment for AI Overviews (optional) - [ ] GSC AI-oriented query filter (optional; see **google-search-console**) ## Output Format - **GA4 setup**: Exploration, channel group, or custom report - **Regex**: Adapted to user's observed sources ## Related Skills - **generative-engine-optimization**: GEO strategy; AI traffic tracking measures GEO impact - **traffic-analysis**: Traffic sources, attribution, UTM - **analytics-tracking**: GA4 events and conversions - **google-search-console**: GSC AI traffic analysis - **robots-txt**: AI crawler allow/block strategy