Tuesday, February 10, 2026
Radware, a global leader in application security and multi-cloud delivery solutions, has launched its new Agentic AI Protection Solution to help enterprises defend against emerging AI-driven threats. With organizations rapidly adopting autonomous AI agents to automate workflows and improve decision-making, security risks have grown significantly. Consequently, Radware has expanded its platform to address the evolving challenges of the agentic AI era.
In January, Gartner projected that global AI spending would reach $2.5 trillion by 2026, with $51.3 billion specifically allocated to AI security. Highlighting this shift, Radware’s leadership emphasized the urgent need for stronger safeguards. According to Roy Zisapel, president and CEO of Radware, “We have seen the AI security wave coming for many months, and knew we needed to support our customers in providing the security surrounding it.” He further explained that while businesses recognize AI’s value, they also understand its risks. Therefore, Radware is leveraging decades of expertise in advanced algorithms to secure complex digital environments.
Moreover, traditional AI security models primarily depend on static guardrails designed for governance rather than real-time protection. As agentic AI systems scale, these approaches often fall short. In response, Radware’s solution uses behavioral analysis to detect suspicious activities dynamically, ensuring stronger protection against misuse and manipulation.
Additionally, the new platform tackles a wide range of risks, including prompt injection attacks, tool abuse, unauthorized access, and human–agent trust exploitation. To strengthen its capabilities, Radware has built the solution around four core pillars. First, it provides real-time visibility into all AI agents within an organization. Second, it applies intent-based security through advanced runtime monitoring. Third, it integrates seamlessly with third-party platforms such as Microsoft 365 Copilot and AWS Bedrock. Finally, it maintains a continuous AI security posture using a dynamic Risk Graph Map that highlights vulnerabilities.
Furthermore, Radware aligns its solution with the OWASP Top 10 for Agentic AI and utilizes the AI Vulnerability Scoring System (AIVSS) to assess risks effectively. This alignment helps enterprises adopt AI agents safely while maintaining control over sensitive data.
Recently, Radware also uncovered a critical vulnerability called ZombieAgent, which enables attackers to implant malicious instructions into AI agents without user interaction. These hidden commands can lead to silent data theft while bypassing traditional security systems. As Constance Stack, chief growth officer at Radware, warned, “Vulnerabilities like ZombieAgent illustrate a critical structural weakness in today’s agentic AI platforms.”
Overall, Radware’s Agentic AI Protection Solution builds on its earlier LLM Firewall launch, reinforcing its commitment to securing AI-driven enterprises in an increasingly complex digital landscape.
To join our expert panel discussions, reach out to info@intentamplify.com
Recommended News
Tags:
LeadDesk strengthens its AI-driven CCaaS platform by acquiring Fluentic, expanding multilingual customer service capabilities across European contact centres.
Bright Pattern partners with Xceed to deliver AI-powered, cloud-based omnichannel contact center solutions that enhance automation, security, and measurable CX performance across MEA.
