San Jose, California, USA, October 6th / PRNewswire / -- AI Native cybersecurity company PRE Security today announced the launch of PRE AgentGuard, which is a capability that embeds PRE Security Platform with ADPR (proactive detection, prediction, and response) functions. This initiative aims to help enterprises monitor and control autonomous AI agents operating within their environments.

Today, AI agents can write code, execute commands, use credentials, query databases, and call API, and often do so at machine speed with almost no human supervision. This presents new security challenges: what happens when agents are manipulated, compromised, malfunction, or begin to perform actions that were not their original intent?

AgentGuard treats each agent as a persistent identity with task objectives, behavioral history, and expected activity patterns. It does not merely check isolated keywords or actions, but continuously assesses whether the agent is still performing the assigned tasks, what the next steps might be, and whether such actions should be permitted.

Behavior analysis tailored for AI agents

The core of AgentGuard is AIBA (which includes AI Agent Behavioral Analytics and AI proxy behavior analysis). According to PRE, this capability extends the UEBA model, which the security team already relies on, to the AI proxy level. It evaluates the identity, tasks, tools, resources, and action sequences of each proxy in order to identify deviations and dangerous trajectories.

AgentGuard is located between the proxy and the tools it uses. It scores each action based on the proxy's tasks and history, and enhances these actions in conjunction with PRE Threat Intelligence. As a result, proxy access to known malicious domains is treated differently from calls to trusted services. During execution, AgentGuard can allow, monitor, require manual approval, or directly block such actions.

What is seen is the trajectory, not just a single action.

Taking an accounts payable agent as an example, the normal process involves reading invoices, verifying suppliers, checking purchase orders, and preparing for payments. If such an agent is manipulated, its activities may shift to accessing credential storage, enumerating sensitive data, and attempting to make external transfers. Each step may seem reasonable on its own, but together, they constitute an ongoing data breach.

PRE indicates that the company has demonstrated this scenario: AgentGuard identifies the trajectory of the proxy and prevents external transfers before any data leaves the environment. In another demonstration targeting the AI programming assistant, the same system command was allowed during normal operation, but once the proxy's previous behavior became suspicious, that command was restricted.

PRE Security, co-founder and CEO, and John Peterson stated: "The prompt word filter asks whether a prompt is dangerous. AgentGuard asks whether an agent is becoming dangerous. We don't just tell analysts that a certain agent has leaked data; instead, we can inform them that it is developing in a direction towards leakage, and we stop it before the data is released."

Extending predictive security to the digital workforce

PRE indicates that AgentGuard has extended its mission of predicting and preventing cyberattacks to the field of autonomous AI. PRE is also integrating AgentGuard with its Log2NLP ™, FAST autonomous investigations, and Agentic Surveillance technology in order to prevent, investigate, and continuously monitor suspicious proxy activities on the same platform.

For more information or to request a demonstration, please visit https :// www.presecurity.ai.

About PRE Security

PRE Security is a AI native cybersecurity company dedicated to using natural language processing, as well as advanced AI and machine learning technologies, to transform the way organizations detect and prevent cyber threats. PRE Security was built from scratch, adopting generative, predictive, and proxy AI approaches, with the aim of overcoming the inherent limitations of traditional SIEM and XDR solutions. The company is headquartered in Silicon Valley and was founded by seasoned cybersecurity professionals John, " JP ", Peterson, and Paul Jespersen.

For more information: https :// www.presecurity.ai.

Contact persons: John Peterson, PRE Security