In a transformative shift for cybersecurity, AI-powered adversaries are leveraging advanced tools like Claude’s agentic capabilities and the Model Context Protocol (MCP) to orchestrate automated cyberattacks at an unprecedented scale and speed. These attacks, while utilizing traditional techniques such as open-source penetration testing tools, emphasize the need for defenders to adopt AI in their defense strategies, focusing on areas like SOC automation, threat detection, and incident response. The novel tempo of these AI-driven attacks, combined with the potential for prompt injections to manipulate AI systems, highlights the urgent requirement for new security controls specifically designed for AI systems. CrowdStrike's response involves enhancing security operations with AI capabilities, aiming to match adversarial speed with automation. The emergence of enterprise AI systems as critical infrastructure necessitates robust defenses, including input validation, context verification, and monitoring to protect against the evolving threat landscape.