AI crawler analytics

Steps

1. get_ai_bot_traffic view: "summary", days: 30. If tracking is not installed, explain it is a one-line setup in SEOsorted (Bot traffic → settings) and stop.
2. Present visits by vendor (OpenAI, Anthropic, Perplexity, Google, Microsoft, others), trend, and share that are AI crawlers vs search crawlers.
3. view: "vendor_pages" with the top vendor(s) to see which pages they read most — these are the pages most likely to feed AI answers.
4. view: "missing_pages": URLs bots request that return 404. Each is either a page worth creating (bots expect it — often from old links or hallucinated URLs) or a redirect to add. track_ai_search_visibility operation: "crawler_misses" splits them into moved pages with a suggested redirect target and expected pages worth creating (free).

Recommendations to make

• An AI vendor absent entirely → audit_any_page operation: "robots" with the site url shows whether each AI crawler may fetch it (free); then check CDN bot blocking.
• Important pages never visited → strengthen internal links to them, add them to the sitemap, consider an llms.txt (llms-txt-generator).
• Missing pages with many hits → create or redirect.