Key takeaways
- Cloudflare now shows personal AI agents as a separate traffic category, with details of which pages each agent visits and whether it hits errors.
- Sites on Cloudflare can allow, block, or block on ad-supported pages each supported personal agent, separately from search and training crawlers.
- There is no deadline, but stores and publishers that block agents broadly may want to check whether those agents are acting for customers.
What Cloudflare launched
Personal agents are AI tools that act on a person’s instructions, such as comparing products, finding support articles or completing a purchase for them. On 4 November 2026, Cloudflare announced two changes for sites that use it.
The first is visibility. The bot chart in its Security Analytics now has a separate Personal Assistants category, and a new widget lists which agents visit a site and how many requests each sends. Filtering on one agent shows the pages it requests, the responses it gets and the networks it comes from, so you can see whether it is browsing products, logging in, reaching checkout or hitting errors. Sites using Account Abuse Protection, which is in early access, can also see which customer accounts each agent acts for.
The second is control. The setting that already covered AI bots for search, training and agents is renamed AI Crawler and Agent access policies. You can set an overall policy for agents, then a different one for each supported personal agent: allow, block on pages with ads, or block. Your current search and training choices are not changed.
Who this affects
This matters if your site runs through Cloudflare and you sell online, publish content funded by ads, or manage customer accounts. Until now, agent visits looked like any other automated traffic, which made a blanket block tempting. Cloudflare’s argument is that a blocked agent may be a customer trying to buy from you.
For online stores, the analytics show whether agents reach product pages and checkout, or fail on pages they cannot use. For publishers, the ads option lets an agent read some pages while keeping it off those that ads pay for. For fraud teams, a suspicious agent on one account can be blocked for that account alone, without blocking it for everyone else.
If your site does not use Cloudflare, nothing changes for you. The announcement does not list which agents count as supported or which plans include each feature.
What to check now
There is no deadline. These are new options, and your search and training settings are untouched.
- Open the Personal Assistants view in Security Analytics to see which agents visit and how often.
- Check where they go, and whether they hit errors on product, account or checkout pages.
- In the AI Crawler and Agent access policies section of Security Settings, review the overall agents setting before choosing per-agent policies.
- If you blocked agents broadly, decide whether that still fits, given they may be acting for customers.
Agent access is set separately from the crawlers behind search and AI answers, so a change to one policy does not alter the others. If you want help reading this traffic and deciding what to allow, Cognivise’s SEO and GEO services cover how AI tools reach and use your site.
Source
- blog.cloudflare.com: See the agents, set the terms: new visibility and control for personal agents
Written by Cognivise from the source above with the help of AI, and checked against that source before it was published.