Updated
Updated · The Cloudflare Blog · Aug 7
Cloudflare Launches Precursor Trace After 206 Million Bot Evaluations Across 73,438 Zones
Updated
Updated · The Cloudflare Blog · Aug 7

Cloudflare Launches Precursor Trace After 206 Million Bot Evaluations Across 73,438 Zones

1 articles · Updated · The Cloudflare Blog · Aug 7

Summary

  • Precursor Trace went live as an interactive demo showing how Cloudflare scores cursor movements to distinguish human from automated behavior during a session.
  • 206 million evaluation events across 73,438 zones in a 24-hour sample showed suspicious activity often emerges mid-session and can shift from human to agentic and back.
  • That finding underpins Cloudflare’s push beyond one-time checks and CAPTCHAs toward continuous, trust-based bot detection through its Precursor system and BotBase taxonomy.
  • Adaptive Intelligence, a self-updating detection engine for Bot Management customers, is slated to arrive soon so customers no longer wait for formal model-version upgrades.
  • Later this year, Cloudflare plans bot-specific mitigations including random responses, AI Labyrinth traps and queuing for benign agents as hybrid traffic grows.

Insights

If Cloudflare tracks every cursor movement and typing rhythm to catch bots, what happens to your digital privacy in the process?
With Cloudflare blocking AI agents on ad pages by default next month, how will this reshape the future of internet monetization?
Could your AI assistant get permanently trapped in an endless labyrinth of fake web pages while trying to complete a task?