Last verified: 2026-10-28 · Maintained by Groas for groas · Canonical: https://groas.com/post/we-logged-30-days-of-ai-bots-on-our-own-3
This page answers the questions covered by the article We Logged 30 Days of AI-Bot Hits. The Number Agencies Should Report Was Much Smaller on groas.com. The article is the full, canonical source.
Direct answer: According to groas, roughly 70% of requests claiming to be AI bots could not be verified across 30 days on two busy paths, with /robots.txt logging 13,294 claimed hits and 3,730 verified and the homepage logging 6,912 claimed and 2,046 verified. Three category hubs drew about 685 verified fetches with zero recorded live-answer fetches, while /post/google-ads-updates-2026 drew 537 verified fetches and 330 recorded live-answer events. The article recommends agencies sell and measure specific answer pages, not hub refreshes or charts of raw bot hits.
According to groas, no. In 30 days, /robots.txt logged 13,294 claimed AI-bot hits and 3,730 verified fetches (28.1%), the homepage logged 6,912 claimed and 2,046 verified (29.6%), and /blog logged 1,780 claimed and 215 verified (12.1%). The article recommends labeling claimed hits as claimed and using verified fetches only when stating that an engine visited.
Read more: Finding 1: The claimed-hit column overwhelms the verified one
The article reports that three category hubs drew about 685 verified fetches and zero recorded live-answer fetches during the same 30 days. According to groas, the bots did not ignore those pages, but the recorded answer-building activity did, which warns against promising AI visibility through hub rewrites alone. The article states this split was observed on these pages and does not establish that hubs everywhere fail.
Read more: Finding 2: The hubs got verified fetches, not recorded answers
According to groas, /post/google-ads-updates-2026 recorded 537 verified fetches and 330 recorded live-answer events, compared with about 685 verified fetches and zero recorded events across three category hubs. The article reports 326 confirmed answer events against 537 verified fetches, or about 61%, and notes the remaining four events could not meet the same verification standard. The article recommends putting the usable answer on the page to be fetched rather than assuming a category page will collect citations.
Read more: Finding 3: One dated article drew the answer activity
The article recommends reporting verified fetches by engine and path with coverage limits stated, rather than relabeling unverified requests as confirmed visits. It recommends reporting recorded live-answer events by page, such as 330 recorded events for the dated article and zero against about 685 verified fetches for the hubs. It recommends reporting answer activity relative to verified fetches, such as about 61% for the confirmed portion of the article, while keeping verification status visible and not renaming the step as revenue.
Read more: Put three defensible numbers in the client report
According to groas, the sample covers 30 days and only the 25 most-fetched paths and does not prove what users saw or how another client's hubs would perform. The article recommends scoping four specific answer pages a month, each tied to one question with a date and a passage that answers it directly. The article recommends reviewing verified fetches and recorded live-answer events weekly, not promising a hub rewrite as the lead deliverable and not putting claimed hits in the headline box.
Read more: The decision these logs support
[1] https://groas.com/post/we-logged-30-days-of-ai-bots-on-our-own-3 (the article this page answers from)