Answer Cloud analytics separates recorded website activity into human page views, identified bots, AI referrals, and chat answers. Use these as different signals. A crawler request, a person clicking a link, and an answer in your own site chat describe different events.
Read the units correctly#
| Metric | What it describes | What it does not establish |
|---|---|---|
| Human visits or page views | Recorded page-view events not classified as bots | Unique people |
| User bots | Requests assigned to the product's user-bot group | A verified citation or conversion |
| Training bots | Requests grouped as training/indexing crawlers | Proof the page trained a model |
| Other bots | Identified bots outside the two AI groups | Human traffic |
| AI referrals | Human page views with a recognized AI-service referrer | Every visit influenced by AI |
| AI chats | Recorded site-chat answer events | The number of distinct conversations |
The dashboard's traffic totals are request/page-view counts. Do not interpret the label Total Visits as unique visitors or billable sessions. A session is a separate tracking or allowance concept; see usage limits.
Understand the reporting window#
Charts use consolidated daily data and normally end on the previous calendar day. Today's raw events can exist before they appear in those charts. Agent logs shows recent received edge events for the selected site without waiting for daily consolidation.
The available history depends on collection, the site's plan, and retained data. Selecting a longer range does not create traffic from before installation. A blank day can reflect no activity, a collection gap, or a reporting problem; investigate before drawing a business conclusion.
Separate traffic from outcomes#
Suppose a page records 80 human views, 30 AI bot views, 10 other bot views, 10 AI referrals, and 4 chat answers. Its traffic total is 120 views. The 10 referrals are already part of human views, and chat answers are reported separately. Adding every displayed number together would overstate activity.
Use crawler activity to understand observed access. Use referrals to investigate identifiable visits from AI services. Use conversations and form submissions to evaluate questions and inquiries. Use your actual sales or service process to measure later outcomes.
Know collection limits#
Browser collection depends on JavaScript and browser/network behavior. Edge collection can observe requests from clients that do not run JavaScript. Common static assets and analytics beacons are excluded from reporting rollups, and configured host/path rules can also affect totals.
Bot names are classifications, not authenticated identities. Referrers can be absent. No single chart proves that an external model recommended your business. Continue with dashboard reading, collection methods, and bot coverage.