Abstract
The increasing adoption of cloud computing, remote work environments, Internet of Things (IoT) devices, and encrypted communication has significantly expanded the attack surface of modern enterprise networks. Traditional rule-based and signature-driven firewall systems are no longer sufficient to defend against advanced cyber threats such as zero-day attacks, lateral movement, and stealthy intrusion attempts. These conventional approaches lack adaptability, generate high false-positive rates, and fail to provide continuous trust evaluation required in dynamic network environments. To address these limitations, this paper proposes an AI-driven Next-generation firewall (NGFW) architecture designed to support dynamic threat detection and Zero Trust implementation. The proposed framework integrates network traffic monitoring, behavioral flow analysis, AI-based threat detection, and dynamic policy enforcement into a unified security system. By analyzing traffic patterns at the flow level, the system continuously evaluates risk and enforces least-privilege access decisions without relying on static rules or predefined signatures.