Personal agents act on a person's instructions — comparing products, locating support articles, completing a purchase — and every one of those actions arrives at a website as automated traffic. Without a way to identify the agent behind a request, site owners face a choice between blocking automation wholesale and losing customers. Cloudflare is now shipping analytics that separate personal agent traffic from other automation, plus finer-grained access policies to act on what the analytics show.

Identifying the agent behind the request

The basis for the new views is Cloudflare's verified bot directory, which the company describes as the largest in the world. That intelligence now feeds traffic details.

  • In Security Analytics, the Bot Analysis chart gains Personal Assistants as a category alongside Verified Bot, Likely Human, Likely Automated, and Automated. Selecting it filters the chart to personal agent traffic over time.
  • In Top statistics, a Personal Assistants widget lists the agents visiting the site and their request counts. Filtering on one agent updates every other widget to that agent's traffic alone — requested paths, response status codes, and originating networks.

The practical question this answers is behavioral: whether an agent is logging in, reading product pages, reaching checkout, or failing on pages it cannot use.

Agents and user accounts

Sites using Account Abuse Protection (in early access) can attribute agent requests to the users they act for. The population overview page shows which personal agents are active across users and how many accounts use them, giving a measure of whether agent traffic matters to the business. Login and sign-up events now record whether a personal agent made the request and which one, and fraud teams can filter accounts by AI assistant events or by number of AI assistants.

Each user's account overview lists the agents acting for that account, so an analyst reviewing an unusual login can distinguish an agent the customer has relied on for weeks from one that appeared today. When an attacker appears to be using an agent on an account, the analyst can mitigate that account with a WAF rule on its Hashed User ID — without blocking the agent for everyone else.

These views inform the decision of which agents to welcome, limit, or block.

Policy controls and the renamed setting

AI bot policies, introduced this summer, let site owners make separate access decisions for AI bots used for Search, Training, and Agent behaviors. The Agents policy now expands to granular controls for supported personal agents, and the setting is renamed AI Crawler and Agent access policies. In Security Settings, an overall Agents policy can be set alongside a per-agent choice of Allow (do not block), Block on pages with ads, or Block — for instance, an overall Block on pages with ads with Allow for a particular agent you want to welcome. Search and Training policies are unchanged.

The same choices appear during onboarding when a zone is added: Cloudflare asks whether pages are monetized with ads and tailors recommended policies to the answer, and the controls remain editable later under Security Settings. Policy is site-specific — one publisher may allow an agent throughout its site so readers can find articles, while another blocks the same agent where ads fund the content.

What remains unsolved: user context

Agent identity answers only part of the question. A site may recognize the agent while lacking what it would otherwise know about the user: existing-customer status, what the agent was authorized to do, subscription state, and which protections apply.

Cloudflare frames the next goal as preserving that relationship context, which it says will require privacy-aware ways to (1) carry user context through an agent, (2) help sites assess reputation, and (3) establish clear accountability when something goes wrong. The company says it will share more as it works with agent operators and others across the ecosystem.

Site owners can review personal agent activity in the new analytics, then set the terms under Security Settings → AI Crawler and Agent access policies.