AI Visibility

Requires

The free Ryze MCP connector (https://connector.get-ryze.ai/mcp). Claude Code: claude mcp add ryze --transport http https://connector.get-ryze.ai/mcp. claude.ai / Desktop / Cursor: add it as a custom connector with the same URL. Setup: https://www.get-ryze.ai/how-to-connect-claude-to-google-meta-ads-mcp

Every "AI visibility tracker" samples prompts and guesses. GA4 records the actual clicks AI engines send. Measure the real thing first, then diagnose.

Workflow

1. Traffic by engine. google_analytics__getAiTrafficByEngine — sessions per AI source (ChatGPT, Perplexity, Gemini, Claude, Copilot…), current vs previous period.
2. Trend. google_analytics__getAiTrafficDaily — is AI traffic growing, and did anything spike (a spike = something started citing you; find it).
3. Cited pages. google_analytics__getAiLandingPages — which URLs AI engines actually send people to. These are your proven-citable pages.
4. Referral detail. google_analytics__getAIReferrals for source-level detail and conversions — is AI traffic converting better or worse than organic.
5. Diagnose the winners. Read the top 2–3 cited pages (fetch them). Note the pattern: direct answers high on the page, stats/definitions, clean headings, schema. That pattern is the citation recipe for this site.
6. Find the misses. Cross-reference with google_search_console__runRawSearchAnalytics: pages with strong organic queries but zero AI referrals — candidates to restructure toward the citation recipe from step 5.

Output

Verdict (AI traffic share of organic, trend, top engine), then: engine table, cited-pages table with conversions, the citation recipe observed on winning pages, and a top-5 list of pages to restructure. Keep it concrete — name the pages and the exact change.

Fallback

If the GA4 AI-traffic tools aren't available in the workspace, build the same report with google_analytics__runRawReport filtered on sessionSource matching known AI referrers (chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, copilot.microsoft.com).