Published 09.11.2026
On November 4, Cloudflare started showing AI agents in its analytics and added a rule for each one: allow, block on pages with ads, or block. If you ever turned on "Block AI bots", agents are most likely already blocked on your ad pages.
On November 4, 2026, Cloudflare added a separate traffic category for AI agents to its dashboard, along with rules for each of them. Cloudflare calls an agent a bot that acts for a person in real time: for example, an assistant opening a page at a user's request, or an AI browser moving through a site on its own.
What changed
- Every agent is visible. Bot analytics now has a "Personal Assistants" category and a list of agents with their requests, pages and response codes.
- A rule for each agent. The setting is now called "AI Crawler and Agent access policies". For each supported agent you can choose allow, block on pages with ads, or block. By default an agent inherits the overall agent policy.
- A question about ads. When you add a site, Cloudflare asks whether it shows ads, and the recommended settings depend on the answer.
Cloudflare doesn't say exactly which agents are covered or which plans include this.
Who this may affect
- New sites with ads. Since September 15, Cloudflare recommends blocking agents on pages with ads and disallowing training for them. Search stays open.
- Sites that once turned on "Block AI bots". When they moved to the new settings, agents were set to block on pages with ads automatically.
- Other existing sites. Cloudflare says nothing changed for them.
In practice, a site can still appear in AI answers through search crawlers, yet an assistant asked to open a specific page with ads may be refused.
What to check
- The agent setting. Open "AI Crawler and Agent access policies" in Cloudflare and see what's selected for agents overall and for each one.
- Analytics. The "Personal Assistants" category shows which agents your site refuses.
- Whether you need the block at all. If your site earns from leads and sales rather than ads, there's usually no reason to block agents: real people are behind them.
Why Cloudflare is doing this
On November 5, Cloudflare published an analysis: when an agent acts for a person, the site sees only the agent, while the site pays for disputed purchases, account abuse and reading without ad views. Cloudflare proposes signing agent requests (Web Bot Auth) and is forming a Trusted Agent Delegation group to take the issue to the IETF. Keep in mind that Cloudflare sells bot protection itself.
Check your site
The free aiseen.cloud check shows how your site responds to bots from ChatGPT, Claude, Perplexity and other AI services. Our requests don't come from OpenAI or Anthropic addresses, so it's worth checking the agent settings in the Cloudflare dashboard as well.
Sources: Cloudflare, November 4, Cloudflare, November 5, Cloudflare, September 15.