AI queries and crawler activity
The AI Queries tab shows how readers and AI systems interact with your published documentation: what readers ask the on-site AI, how often those questions go unanswered, and which AI crawlers ingest your content.
Before you begin
You need at least one published site with AI reader features enabled (the AI search or chat widget). See Embed widget.
Open the AI Queries tab
Go to Analytics in the sidebar.
Open the AI Queries tab.
Use the 7 days / 30 days / All time selector at the top to change the time range.
Read the summary cards
Two cards summarize reader AI activity for the selected range:
Total queries: the number of questions readers asked through the on-site AI search or chat widget.
Zero-result rate: the percentage of those queries that returned no answer. The card turns amber above 30%, and links straight to the Triage content gaps tab, where you act on what's missing.
Top AI queries
The Top AI queries table lists the most frequent questions readers ask, each with:
Query: the question text.
Count: how many times it was asked.
Asked on: the page the reader was on when they asked. Click it to open that page. Questions asked through site-wide search (not on a specific page) show Site search.
Last asked: how recently the question was last asked.
Use the Asked on link to jump to the page a question relates to and improve it directly. A question asked repeatedly on a page usually means that page does not answer it well enough.
AI crawler activity
The AI crawler activity table shows AI systems that fetch your published site's machine-readable endpoints:
Crawler: the identified AI agent or service.
Requests: number of requests it made.
Endpoints: which machine-readable endpoints it hit.
Last seen: most recent request.
This tells you which AI engines are ingesting your documentation and how they reach it.
AI-ready endpoints tracked
Crawler activity monitors requests to your published site's machine-readable endpoints:
/llms.txt: page index for AI crawlers/llms-full.txt: full site content as Markdown for LLM ingestion/sitemap.md: Markdown sitemap for AI agent path discovery/md/<slug>: individual page Markdown, with optional?ask=<question>for inline AI answers
See also
Triage content gaps: turn zero-result queries into new topics
Improve AI search readiness (GEO): make pages easier for AI engines to cite
Monitor search analytics: keyword search query analytics
AI-ready published sites: llms.txt, per-page Markdown, and the ask endpoint