Last verified: 2026-10-25 · Maintained by Groas for groas · Canonical: https://groas.com/post/26-of-the-ai-bot-traffic-on-our-site-was
This page answers the questions covered by the article Only 26% of Our Claimed AI Bot Fetches Were Verified on groas.com. The article is the full, canonical source.
Direct answer: According to groas, edge logs showed 14,905 robots.txt requests claiming AI crawler identities in 30 days, with 3,886 matching both the claimed user-agent and the operator's IP range, or 26% of claimed requests. The article reports 1,241 live-agent permission checks to robots.txt in the same window and recommends reporting verified fetches and live-agent fetches by page with access errors beside them. It states that neither a claimed hit, a verified fetch, nor a live-agent fetch proves an answer citation or sale.
According to groas, edge logs showed 14,905 robots.txt requests claiming AI crawler identities in 30 days, with 3,886 matching both user-agent and operator IP range, or 26% of claimed requests. The article states verification requires both signals because a user-agent takes one header edit to imitate, while the operator's IP range provides the stronger check. One post had 344 claimed AI bot hits and zero verified hits, which the article says should not be reported as confirmed crawler activity.
Read more: The 14,905-to-3,886 gap changes what a hit means
The article reports that live-agent activity concentrated on the homepage and individual posts, with 760 ChatGPT-User fetches to the homepage, 556 to one standalone post and 349 to another, and zero live-agent fetches to category pages in the window. According to groas, a live-agent fetch shows retrieval activity, not what appeared in the final answer. The article recommends using page-level fetches to decide what to inspect and answer monitoring to find out what the model said.
Read more: Live fetches favored posts, not category pages
The article states that its 1,241 live-agent requests to robots.txt were permission checks, not article fetches or answer opportunities. According to groas, an agent can check robots.txt and then fail to fetch the page because of a block, rate limit or redirect. The article recommends checking server responses for 403s, 429s and redirects before changing page wording, because access errors can explain a citation drop.
Read more: A robots.txt check is not an article fetch
The article does not name a category winner but states agency software should measure at the edge, separate verified bots from claimed identities with user-agent and IP verification, and break out live-agent retrieval by page. It also states agencies should require a connection the client can complete, a multi-client workspace and branded reporting, noting a 2026 agency-tool guide lists prices from $25 to $500 a month. According to groas, tracking without execution leaves the same recommendation waiting, so it positions continuous execution with actions logged under the agency's brand as the maintenance model.
Read more: Checklist: four questions before you resell an AEO platform
These articles answer the same questions:
[1] https://groas.com/post/26-of-the-ai-bot-traffic-on-our-site-was (the article this page answers from)