AI assistant traffic may be fake, agencies need forensic measurement
One site found 81.8% of AI assistant traffic was fake, and ClaudeBot outpaced Googlebot. Agencies now need log-level checks before calling AI visits real.
A new site found that 81.8% of visits labeled as coming from an AI assistant were fake, and the Googlebot number was worse, a blunt warning that dashboard labels can look authoritative while failing simple verification. ClaudeBot also outpaced Googlebot on the same new site, showing how easily modern crawlers, scrapers and assistant traffic can blur together before anyone checks the underlying request data.
The wider backdrop is already noisy. Cloudflare said on June 4 that automated bot and agent traffic had surpassed human traffic online, and NBC News put the split at 57.4% automated requests and 42.6% human-generated requests across a selection of websites it hosts. That scale helps explain why agencies are seeing more synthetic sessions, more crawl chatter and more false confidence in analytics exports that were never designed to distinguish a buyer from a bot.
The technical standard for verification is not complicated, but it is routinely ignored. Google says a request can be checked against its real crawler infrastructure by validating the IP address, running reverse DNS lookups and confirming that the hostname ends in googlebot.com or google.com. A user-agent string by itself is not proof. OpenAI now documents separate crawlers for different purposes, including GPTBot and OAI-SearchBot, while Anthropic documents ClaudeBot and says it can be blocked with robots.txt. Anthropic also warns that blocking its crawler may reduce a site’s visibility in user-directed web search.
That mix of legitimate crawlers, AI retrieval systems and spoofed agents is exactly where agency reporting can go wrong. A traffic spike tagged to an AI assistant may be synthetic, misclassified or generated by a crawler that never had a chance of becoming a lead. The safest measurement stack pairs server logs with IP validation, user-agent checks and first-party analytics before anyone claims that AI visibility is driving demand. Without that, budget decisions, content priorities and client reporting can all be built on traffic that never represented a human at all.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
