Proofpoint announced two new agentic cybersecurity systems aimed at helping organizations secure sensitive data, workplace communications, and autonomous AI agents as AI adoption accelerates across enterprises. The systems are the Proofpoint Agentic Data and AI Security System and the Proofpoint Agentic Collaboration Security System, each addressing a different layer of AI-related risk.
According to Proofpoint, most existing security tools are built to observe either AI behavior or data risk, but not both at once, leaving gaps as AI agents increasingly access and act on enterprise data. The company’s 2026 AI and Human Risk Landscape report found that 87% of organizations have moved AI assistants beyond the pilot stage, yet 52% lack confidence that their current controls could detect a compromise.
Proofpoint Agentic Data And AI Security System
The Proofpoint Agentic Data and AI Security System is built on the Proofpoint Knowledge Graph, which connects AI activity with data sensitivity, identity, access, behavior, and intent.
It is powered by three autonomous agents: Zero-Touch Detection, which surfaces high-priority actions instead of flooding teams with anomalies; Instant Investigation, which automatically reconstructs incidents across data, identity, and behavior; and Protection Optimization, which carries out access remediation and refines data loss prevention (DLP) policies with human oversight.
The system also introduces Semantic Business Policies, which translate plain-language governance rules into enforceable runtime controls, and Agentic Insights, which continuously scans AI interactions to surface risks organizations have not yet formally defined.
Proofpoint Agentic Collaboration Security System
The second system, Proofpoint Agentic Collaboration Security System, applies multi-stage AI reasoning across email, collaboration tools, and the browser to counter attacks that mimic legitimate business activity, such as compromised supplier accounts or fraudulent payment requests that match known relationships. It is powered by the Nexus Intent-Based Detection Model, which analyses context to distinguish malicious intent from ordinary workflows both before a message is delivered and after it reaches the inbox.
Beyond detection, the system includes Autonomous Threat Investigation for reconstructing attacks and assessing blast radius, and Adaptive User Protection, which identifies high-risk users and adjusts defenses accordingly, supported by Blue Team and Red Team agents that test protections for weaknesses. Proofpoint has also introduced Advanced Browser Protection, developed with Push Security, to extend coverage to post-click phishing, credential theft, session hijacking, and malicious browser extensions.
Availability
Capabilities within the Agentic Data and AI Security System, including Semantic Business Policies and Agentic Insights, are expected to be available by year-end 2026. The Agentic Collaboration Security System will roll out in the first quarter of 2027 as an update to Proofpoint’s existing collaboration security solution, with availability varying by country depending on data residency requirements.
Pokdepinion: AI agents running wild in company systems sounds like a headache waiting to happen, so tools like this feel overdue.
