As personal artificial intelligence agents rapidly evolve to act as the primary digital intermediaries between humans and the open internet, website owners face a growing operational challenge. These AI assistants—tasked by users with everything from comparing products and surfacing support documentation to completing e-commerce transactions—appear in standard web server logs simply as automated traffic. Without clear visibility into which specific agents are visiting, what pages they are scraping or interacting with, and whether their presence supports or harms business goals, site operators are often forced into a difficult binary choice: block all automated traffic for safety or risk losing valuable human customers.
Addressing this industry-wide shift in web traffic composition, Cloudflare has announced a comprehensive suite of new analytics tools and granular access policies designed to help website owners understand, evaluate, and manage personal AI agents on their own terms. By integrating data from the world’s largest verified bot directory directly into user traffic details, the platform aims to transform unexplained automated requests into actionable insights, enabling businesses to welcome authorized AI traffic safely and effectively.
See how agents access your site
Understanding the exact identity behind an incoming request is the first step toward managing automated traffic with confidence. Cloudflare’s Security Analytics dashboard now incorporates Personal Assistants as a distinct, dedicated category within its Bot Analysis chart. Sitting alongside existing classifications such as Verified Bot, Likely Human, Likely Automated, and Automated, this new category allows security teams to isolate and monitor personal agent traffic trends over time with precision.
Further expanding this visibility, the platform’s Top Statistics view now features a dedicated Personal Assistants widget. This component details which individual agents are visiting a site and quantifies the exact volume of requests generated by each one. When an administrator filters data by a specific agent, the entire dashboard dynamically updates to reflect only that entity’s footprint. This includes the precise URL paths requested, the corresponding HTTP status codes returned, and the underlying network origins. Consequently, site owners can immediately determine whether a given agent is successfully logging in, navigating product catalogs, reaching checkout pages, or continually hitting error boundaries on restricted sections of the site.
For websites utilizing Cloudflare’s Account Abuse Protection—currently available in early access—fraud teams and administrators gain even deeper visibility down to the individual user level. A comprehensive population overview page displays which personal agents are actively utilized across a site’s user base, alongside the precise number of accounts engaging them. This granular breakdown helps organizations evaluate whether personal agent traffic constitutes a meaningful, legitimate component of their operational business model.
Every login and sign-up event within protected accounts now explicitly records whether a personal agent initiated the request, identifying the specific assistant involved. Fraud analysts gain the ability to filter accounts directly based on AI assistant interaction events or the sheer volume of integrated assistants. Furthermore, individual account profiles provide a clear audit trail of the agents acting on behalf of that specific user.
If an unusual login occurs, an analyst can quickly verify whether the request originated from a trusted agent the customer has relied upon for weeks or an unfamiliar entity appearing for the first time. Should an attacker attempt to exploit a user account by masquerading through an agent, fraud teams can apply targeted mitigations—such as a Web Application Firewall rule tied to a Hashed User ID—without resorting to a blunt-force block that would disrupt the agent for other legitimate users. Together, these monitoring capabilities provide a rigorous foundation for deciding which personal agents to welcome, restrict, or turn away.

Set access policies that reflect your preferences
Building upon the initial release of AI bot policies earlier this year, which allowed site owners to establish separate access controls for search engines, model training crawlers, and general agent behaviors, Cloudflare is now expanding its configuration framework. The existing agent policy setting has been renamed to AI Crawler and Agent access policies to reflect the inclusion of robust, granular control options specifically tailored for supported personal agents.
Within the Security Settings interface, administrators can establish a baseline policy for overall agent traffic while defining distinct exceptions for individual, supported personal agents. Available control actions include allowing traffic unconditionally, blocking access specifically on pages containing advertisements, or enforcing a total block. For instance, a publisher might configure the broad agent policy to restrict access on ad-monetized pages while selectively granting full access to a preferred personal assistant that drives valuable readership. Conversely, another organization might choose a completely different posture depending on its unique business model.
These configuration choices are also integrated into the onboarding workflow for new zones added to Cloudflare. During initial setup, the platform inquires whether a site monetizes its content through advertisements, automatically tailoring recommended policies to that operational reality while leaving the door open for administrators to modify their security posture at any time.
Agent identity is only part of the story
While the newly launched analytics and access policies provide unprecedented visibility and control over personal agent traffic, industry observers note that identifying the visiting bot is only part of a much larger ecosystem transformation. As personal agents increasingly operate as deep intermediaries between humans and digital services, websites frequently lose crucial contextual information that traditionally accompanied a direct human connection. An e-commerce platform or media site may recognize the specific agent making an inbound request, but it often lacks vital data regarding whether the human behind the request is an existing customer, what precise parameters the user authorized the agent to execute, whether an active subscription is in place, or which specific compliance protections should apply.
Bridging this informational gap will require the development of privacy-aware mechanisms capable of securely carrying user context through an intermediary agent, helping target sites accurately assess reputation, and establishing clear accountability frameworks when errors or security incidents occur. Cloudflare indicates it will continue collaborating with agent operators and other stakeholders across the digital ecosystem to address these challenges, ensuring site owners can confidently grant agents broader access while preserving direct, trusted relationships with their end users.
Ultimately, the human asking an AI assistant to compare products, locate support documentation, or finalize a purchase remains a customer attempting to engage with a business. By reviewing current access patterns through newly established analytical tools and configuring appropriate policies within Security Settings, site owners can navigate the rise of personal agents on terms that align with their commercial objectives.