AI agents are beginning to blur the line between legitimate customers and automated traffic, creating a new security challenge for enterprises. Cequence Security says traffic matching Meta’s Muse personal AI agent appeared at more than half of the customers it studied within two weeks of Muse’s launch.
The analysis, covering customers across financial services, retail, travel, software and other sectors between September 1 and 24, found that Muse traffic grew nearly sixfold at the median customer. The agent runs a real Chrome browser in the cloud and routes traffic through a consumer VPN, while presenting itself to websites like a conventional browser rather than identifying as an AI agent.
“Every CISO is about to hear the same question from the business, which is whether the company can let AI agents in. Saying no is no longer the safe answer, because customers are choosing agents and will take their spending wherever those agents are welcome. Our job is to make yes the safe answer.”
— Ameya Talwalkar, Co-founder and CEO, Cequence Security
That creates a significant gap for conventional bot and security controls. Cequence found Muse successfully completing multi-factor authentication at financial institutions, while also browsing products, comparing options, filling carts and progressing through checkout.
The implications extend beyond Muse itself. If an agent can legitimately act for a customer, blocking the entire agent risks blocking valuable transactions. Conversely, allowing every request creates opportunities for compromised credentials or malicious agents to abuse trusted access.
Cequence’s new Agent Trust capability addresses this distinction by combining agent identity verification with behavioural analysis. It provides an inventory of detected agents, an issuer registry, activity logging and transaction-level controls that can block, rate-limit or challenge individual actions.
“AI agents like Meta’s Muse are bots that act on behalf of real customers. Many businesses can’t see them, and those that can are tempted to block them. Agent Trust lets businesses block the action, not the agent.”
— Hari Nair, VP of Product Management, Cequence Security
The approach shifts the security question from whether an organisation should allow AI agents to whether each action is legitimate.
For enterprises, that distinction could become increasingly important as customers delegate browsing, authentication and purchasing to AI agents. Cequence argues that the future of agentic commerce will require businesses to make controlled access possible rather than simply defaulting to blocking automated actors.
