AI bots are automated programs dispatched by AI services or search services. They read public website content and may be used to train models, answer user questions, create AI search results, or for general search indexing.
The "AI Bots" tab in "Analytics" allows you to see which known AI bots have visited your public website and whether they were allowed or blocked. This article explains how to read the records; if you want to change which bots can access your site, read "Manage AI Bot Access to Your Website".
Regular human visitors do not appear on this tab. AI bots do not execute frontend JavaScript tracking for typical websites, so they are counted independently from server-side logs.
Purpose classification is based on publicly available information about known bots and does not guarantee the exact final use for every visit.
The system compares the source IP and the operator's published network ranges to confirm that the request credibly originates from the official bot.
The request's name resembles a known AI bot, but the source IP does not match the official ranges. This could be a third-party impersonation, so do not treat it as a genuine visit from the official AI service.
Some services lack public IP information for comparison, or the verification could not be completed for a specific request. Unverified does not mean malicious; it simply means the system cannot confirm its authenticity.
"Allowed" in the records means the request was permitted to continue accessing the public website. "Blocked" means it matched your blocking rules in "Console → Search, AI & Tracking → AI Bot Blocking", and the public website responded with a denial of access.
Blocked records are still valuable. They confirm that your rules are working and which bots are continuously attempting access.
Page details list the URLs the bots have accessed. If an important page has not been read by any known bots, it doesn't mean it won't appear in search results; crawl frequency is determined by each service, and it takes time, especially for newly published or sparse content.
Not necessarily. A visit simply means the bot has read your public website. Whether it's used for training, inclusion, citation, or presentation in responses is determined by each service.
Not necessarily. A newly published website, low traffic, bots not yet scheduled for a visit, or bots not on the current recognition list can all result in a temporary lack of data in the report.
First, check if the request is causing abnormal traffic. If it's just a few records, immediate action is usually not required. If you notice persistent anomalies, please note the date, take screenshots, and contact customer support with the URLs.
